. 3.1.» 7&1. 3%.wa Ill-.1 ’ .. II“... gilt”. n7. . . .1: Jamar“ var‘... "aura... . .1: 2:. . . . c... 1.... 11* 3.2.1.3 l. :Plvnvknvuu z.- "ufafljwztv ‘. .2. kn”. thud-tam": t‘b. .1. $“(L 1" .J 1!... mnuwflm Eva.” E, a. [I ‘i‘ig3v~ 1 {Lu vv’ptx‘ACIw .- (”u-wot! .ll... . Ia .Ifdfihlufz. !. 171...! a O; . ‘ ‘ 3.4.1 1 a, . ,f‘. '1 “aha“... lnl I. {A .«fiu ha...“ Iii . I .Hi ‘iiufidax 13.554! {0 .7 v .. 1.53.. filial! 3. it. .. nub"! I .23, a... a; I .. - . . . hit)... .{rhflo’nufi .Iildkihuuyfiv (In . “'1... 'TFWS K 3 LIBRARY [00b Michigan State University This is to certify that the dissertation entitled MODELING THE CUMULATIVE EFFECTS OF ASPEN MANAGEMENT PRACTICES ON TIMBER AND WILDLIFE AT MULTIPLE SPATIAL AND TEMPORAL SCALES presented by Alexandra B. Felix has been accepted towards fulfillment of the requirements for the Ph.D. degree in Fisheries and Wildlife , l x ( a f\ "A “W k'i—i‘J~J\/\,\ \ -J‘\'\v 3N 3‘:L.5._5 Malpr Professor’s Signature m “‘~~"L\«\ I“. zany?) Date MSU is an affinnative-action, equal-opportunity employer PLACE IN RETURN Box to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE H SEP 93/375111 061812.53 213% 5/08 K IProVAocS-PrelelRC/Dateoue indd MODELING THE CUMULATIVE EFFECTS OF ASPEN MANAGEMENT PRACTICES ON TIMBER AND WILDLIFE AT MULTIPLE SPATIAL AND TEMPORAL SCALES VOLUME I By Alexandra B. Felix A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Fisheries and Wildlife 2008 ABSTRACT MODELING THE CUMULATIVE EFFECTS OF ASPEN MANAGEMENT PRACTICES ON TIMBER AND WILDLIFE AT MULTIPLE SPATIAL AND TEMPORAL SCALES By Alexandra B. Felix To effectively manage wildlife and timber resources, managers and planners need to understand how different ecological conditions and management practices affect forest characteristics and subsequently how wildlife respond. To address this need, I refined an ecological classification system used to define habitat types, and investigated the potential Spatial distribution of aspen, ecological differences associated with aspen in 3 age classes (20—29, 50—59, > 70) and 6 habitat types, avian community associations, ruffed grouse selection of aspen for drumming habitat, and the cumulative effects of aspen management practices on wildlife and timber in a 12,300 km2 area in Michigan’s Upper Peninsula from 2004—2006. The goals of the project were to understand how forest management may affect the structure of vegetation within ecosystems, the distribution of resources within the landscape, and wildlife response to those changes; and to investigate the feasibility of different aspen management scenarios on timber sustainability and maintenance of wildlife habitat. Differences in vegetation structure and composition were observed among age classes and habitat types (p < 0.10); most notable were stem and shrub density, herbaceous cover, and conifer cover. These differences contributed to identification of 7 avian community associations and difl‘erences in grouse drumming habitat selection. Grouse selection was higher (p < 0.05) for young (0—29-year—old) and old (> 50 years) aspen and aspen on habitat types supported by mesic loamy soils. Grouse did not select drumming areas in non-aspen vegetation types and in aspen growing on xeric sandy soils. I used goal programming to develop a timber harvest schedule for 10 decades that minimized the deviation (in terms of area) from meeting the Michigan Department of Natural Resources’ goal of establishing and maintaining an even-age class distribution of aspen in 8 age classes in 9 habitat types under 3 management scenarios (i.e., harvest all aspen by 80 years, harvest all aspen > 60 years, and intensively harvest aspen by requiring 2 60% to be harvested from the 40-year age class). The spatial feasibility and spatial and temporal consequences of the harvest schedule was evaluated through simulations using the program HARVEST. The goal was feasible by decade 8 under scenario 1 (i.e., harvest by 80 years), it was attainable in 6 habitat types by decade 8 under scenario 2 (i.e., harvest aspen > 60), and it was not attainable under intensive management. Meeting the goal increased habitat availability for most avian communities, but decreased the amount of highly selected grouse drumming habitat. Aspen volume harvested was sustainable at168,000 cords/decade once the goal was met. Within many state and federal agencies, forest management goals include sustainable management to meet a variety of objectives. The cumulative effects of management decisions are what determine how well long-term forest management goals are being met. The results of this project provide quantitative answers to questions such as how much aspen to cut, where can it be cut, and what are the effects on timber production and wildlife at multiple levels of biological organization. With increasing demands on natural resources, answers to such questions are essential for agencies and other landowners to sustain timber and wildlife resources. ACKNOWLEDGEMENTS This project would not have been possible without the financial support provided by the Michigan Department of Natural Resources, Michigan State University Land Policy Program, Rocky Mountain Goats Foundation, and Safari Club International. Funding provided by the Michigan Department of Natural Resources was through the F ederal Aid in Restoration Act under Pittman-Robertson project W-147-R. I am grateful to many individuals for support and assistance throughout the duration of this project. I would like to thank Ed Dobrowolski, Renee Duffey, Craig Felix, Ryan Fry, Tim Hiller, Scott Hughes, Alan Leach, Jeff Locher, Jeff Luwicki, my dad, Jonathan McClain, Michelle Prehoda, Nancy Schwalm, Dave Unger, and Dan Walsh for assistance with field work. Gary, Su, Jim, and Jason Housmann, who rented a cabin to me, were extremely hospitable and helpful during summer and winter field seasons. Mark Mackay, Jim Ferris, Craig Albright, Rob Aho, Doug Wagner, Mike Koss, and Bob Doepker from the Department of Natural Resources were very helpful, insightful, and influential during the development and implementation of this project. Dr. Gary Roloff, Dr. Eric Gustafson, Sue Lietz, and Mark Parkinson were helpful in providing information or insights on various aspects of this project. Thanks to Karen Wallace for encouraging me and helping me through the last year’s stretch of my dissertation. Thanks to lab mates Dan Linden and Dan Walsh for the support and humor that made working fun. It was great to be a part of the ne plus ultra lab! I am grateful to my committee members Dr. Dean Beyer, Dr. Dan Hayes, Dr. Larry Leefers, and Dr. Scott Winterstein for the advice and guidance throughout the duration of this project. Dr. Beyer was very helpful providing insights and direction from iv an agency perspective. Dr. Dan Hayes helped me understand selection functions and provided much assistance with the completion of Chapter 3. There is no way I could have completed Chapter 4 without Dr. Larry Leefers. Thank you for the late-night responses to urgent requests for assistance with goal programming and harvest simulations. Thank you, also for telling me to “hang in there!” Thanks to Dr. Winterstein for assistance with principal components analysis and determining avian community associations. I would especially like to thank my advisor and friend Dr. Rique Campa. I wish I could express in words the gratitude I feel from all the opportunities, support, and encouragement you have given me throughout my graduate career and especially during 2007. Thank you for sharing your gift of teaching, research, and mentoring with me because now what is excellent in you is a part of me, too. TABLE OF CONTENTS LIST OF TABLES .............................................................................................................. x LIST OF FIGURES ....................................................................................................... xviii INTRODUCTION ............................................................................................................... 1 STUDY AREA .................................................................................................................... 7 CHAPTER 1- Quantifying composition, distribution, and temporal changes of habitat types that can support aspen in the western Upper Peninsula, Michigan .............. 10 OBJECTIVES .......................................................................................................... 13 METHODS ............................................................................................................. 14 Landscape classification ................................................................................ 14 Selection of stands and experimental design ................................................. 14 Vegetation sampling ...................................................................................... 17 Data analysis .................................................................................................. 18 RESULTS ................................................................................................................ 23 Landscape classification ................................................................................ 23 Landscape-level characteristics and distribution of aspen ............................. 25 Stand-level characteristics of aspen: Comparison among age classes ........... 35 Live vegetation structure ....................................................................... 35 Dead vegetation .................................................................................... 35 Species composition .............................................................................. 36 Diameter of overstory tree species ........................................................ 37 Basal are of overstory tree species ........................................................ 37 Stand-level characteristics of aspen: Comparison among habitat types within different age classes ........................................................... 47 Live vegetation structure ....................................................................... 47 20-year age class ........................................................................... 47 50-year age class ........................................................................... 48 70-year age class ........................................................................... 49 Dead vegetation structure ..................................................................... 57 20-year age class ........................................................................... 57 50-year age class ........................................................................... 57 70—year age class ........................................................................... 58 Comparison of species composition among habitat types and age classes....65 Timber volume ............................................................................................... 83 DISCUSSION .......................................................................................................... 85 Differences in vegetation structural attributes across habitat types ............... 87 Representation of habitat types ...................................................................... 94 Representation of age classes ......................................................................... 96 vi Implications for Michigan’s Sustainable Forestry Initiative ......................... 97 CHAPTER 2- Wildlife-habitat relationships: spatial and temporal variations in distribution of wildlife communities that use aspen ........................................................ 101 OBJECTIVES ........................................................................................................ 105 METHODS ............................................................................................................ 106 Sampling of vegetation and avian communities .......................................... 106 Data analysis ................................................................................................ 108 RESULTS .............................................................................................................. 115 Principal components analysis for vegetation variables .............................. 115 Frequency of occurrence of breeding bird species among aspen age classes ............................................................................................... 115 Frequency of occurrence of breeding bird species among habitat types ..... 1 19 Frequency of occurrence of breeding bird species among age classes and habitat types ................................................................................ 126 Frequency of occurrence of breeding bird species throughout succession..l35 Within the TM habitat type ................................................................ 135 Within the ATD habitat type .............................................................. 135 Within the AVG habitat type ............................................................. 136 Community characteristics ............................................................................ 143 Nesting preference ............................................................................. 143 Migratory status ................................................................................. 145 Sensitive species ................................................................................ 146 Habitat suitability for indicator species ........................................................ 148 Veery .................................................................................................. 148 Ovenbird ............................................................................................ 148 Yellow-rumped warbler ..................................................................... 148 American redstart ............................................................................... 149 Black-capped chickadee ..................................................................... 149 Pileated woodpecker .......................................................................... 149 Avian communities ....................................................................................... 152 Winter resident birds ..................................................................................... 161 DISCUSSION ....................................................................................................... 166 Vegetation structural differences among age classes and habitat types ...... 166 Avian species associations with aspen age classes and habitat types .......... 167 Fine-filter assessment of habitat suitability ................................................. 172 Conclusions .................................................................................................. 1 75 CHAPTER 3- Potential of aspen in different aspen age classes and habitat types to provide drumming habitat for ruffed grouse .................................................................... 179 OBJECTIVES ........................................................................................................ 1 81 METHODS ............................................................................................................ 182 vii RESULTS .............................................................................................................. l 86 Selection by age class .................................................................................. 186 Selection by habitat type .............................................................................. 187 , Selection among age classes and habitat types ............................................ 188 Spatial distribution of potential grouse drumming habitat .......................... 191 DISCUSSION ...................................................................................................... 201 Management implications ............................................................................ 204 CHAPTER 4-Simulating effects of aspen harvest strategies on wildlife habitat and timber in northern Michigan ............................................................................................ 207 OBJECTIVES ........................................................ ' ................................................ 211 METHODS ............................................................................................................ 212 Spatial data compilation ............................................................................... 212 Development of a harvest scheduling model ............................................... 215 Model inputs and structure ................................................................. 217 Simulation descriptions ................................................................................ 226 HARVEST simulations ................................................................................ 226 Landscape metrics ........................................................................................ 228 RESULTS .............................................................................................................. 231 Scenario 1: Harvest all aspen by 80 years ................................................... 231 Timber harvest ................................................................................... 231 Effects on wildlife .............................................................................. 233 Avian communities .................................................................. 233 Ruffed grouse ........................................................................... 234 Scenario 2: Harvest aspen Z 60 years .......................................................... 250 Timber harvest ................................................................................... 250 Effects on wildlife .............................................................................. 252 Avian communities .................................................................. 252 Ruffed grouse ........................................................................... 252 Scenario 3: Intensive harvest of aspen (60% of all aspen harvested from the 40-year age class ................................................................. 269 Timber harvest ................................................................................... 269 Effects on wildlife .............................................................................. 270 Avian communities .................................................................. 270 Ruffed grouse ........................................................................... 271 Landscape metrics .................................................................................... 287 DISCUSSION ........................................................................................................ 290 Feasibility of implementing harvest schedules developed from goal programming methods .............................................................. 290 Spatial and temporal differences in availability and distribution of forest resources ................................................................................... 292 Comparison of effects of harvesting strategies on wildlife ......................... 293 Adaptive management ................................................................................. 295 viii CHAPTER S-Management implications ......................................................................... 298 APPENDIX 1 ................................................................................................................... 306 APPENDIX 2 ................................................................................................................... 316 APPENDIX 3 ................................................................................................................... 322 REFERENCES ................................................................................................................ 330 ix LIST OF TABLES CHAPTER 1 Table 1.1. Percent of study area (Baraga, Dickinson, Iron, and Marrquette counties) in different of cover types in the western Upper Peninsula, Michigan circa 1999 .................. 9 Table 1.2. Methods used to measure forest stand attributes in aspen stands in Michigan’s Upper Peninsula ................................................................................................................. 21 Table 1.3. Codes for determining decay status of down woody debris, snags, and stumps. These were adapted from decay class codes used by the United States Department of Agriculture, Forest Service ................................................................................................ 21 Table 1.4. Landscape composition of habitat types in a 12,300 km2 area in the western Upper Peninsula, Michigan as determined from spatial habitat type delineation from literature and field data collected between 2004—2006. The study area consisted of Baraga, Dickinson, Iron, and Marquette counties .............................................................. 27 Table 1.5. Percent of asan on state lands in different habitat types in the western Upper Peninsula, Michigan. Habitat types were determined from spatial habitat type delineation fi'om literature and field data collected between 2004—2006. The study area consisted of Baraga, Dickinson, Iron, and Marquette counties .............................................................. 32 Table 1.6. Percent of aspen in different age classes and habitat types on state lands in the western Upper Peninsula, Michigan. Habitat types were determined from spatial habitat type delineation from literature (Coffman et al. 1980) and field data collected between 2004—2006. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccim'um; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; FI = F raxinus-Impatiens; FMC = F raxinus-Mentha—Carex; PCS = Picea-Chamadaphne—Sphagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer— Epigaea; TAM = T suga-Acer—Mitchella; TM = T suga-Maianthemum; TMC = T suga- Maianthemum-Coptis; TMV = Tsuga-Maianthemum-Vaccim'um; TTM = Tsuga-Thuja- Mitella; TTS = T suga-T huja-Sphagnum ............................................................................ 33 Table 1.7. Comparison of vegetation attribute variables among different age classes of aspen in the western Upper Peninsula, Michigan during 2004—2006 ................................ 39 Table 1.8. Comparison of characteristics of dead vegetation (i.e., down woody debris, snags, stumps) among different age classes of aspen in the western Upper Peninsula, Michigan during 2004—2006 .............................................................................................. 40 Table 1.9. Average number of stems and standard errors of overstory tree species occurring in different age classes of aspen on state land in the western Upper Peninsula, Michigan during 2004—2006 .............................................................................................. 41 Table 1.10. Average diameter at breast height and standard errors of overstory tree species occurring in different age classes of aspen on state land in the western Upper Peninsula, Michigan during 2004-2006 ............................................................................ 43 Table 1.11. Average basal area and standard errors of overstory tree species occurring in different age classes of aspen on state land in the western Upper Peninsula, Michigan during 2004—2006 .............................................................................................................. 45 Table 1.12. Averages and standard errors for variable characterizing live vegetation structure of aspen stands in the 20-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer—Quercus- Vaccinium; ATD = Acer—Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; QAE = QuercuS-Acer-Epigaea; TM = T suga— Maianthemum .................................................................................................................... 5 1 Table 1.13. Averages and standard errors for variable characterizing live vegetation structure of aspen stands in the 50-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. ATD = Acer- Tsuga— Dryopteris; AVO = Acer- Viola-Osmorhiza; QAE = Quercus-Acer-Epigaea; TM = T suga- Maianthemum .................................................................................................................... 53 Table 1.14. Averages and standard errors for variable characterizing live vegetation structure of aspen stands in the 70-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccim'um; ATD = Acer- Tsuga-Dryopteris; AVO = Acer— V iola-Osmorhiza; TM = T suga-Maianthemum ..................................................... 55 Table 1.15. Averages and standard errors for variable characterizing structure of down woody debris, snags, and stumps within aspen stands in the 20—year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer—Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; QAE = Quercus-Acer-Epigaea; TM = T suga~Maianthemum ......................................................................................................... 59 Table 1.16. Averages and standard errors for variable characterizing structure of down woody debris, snags, and stumps within aspen stands in the 50-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. ATD = Acer-Tsuga-Dryopteris; AVO = Acer— Viola-Osmorhiza; TM = T suga- Maianthemum .................................................................................................................... 61 Table 1.17. Averages and standard errors for variable characterizing structure of down woody debris, snags, and stumps within aspen stands in the 70-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer—Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga—Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum ............. 63 xi Table 1.18. Average number of stems/ha and standard errors for overstory tree species within aspen stands in the 20-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; QAE = Quercus-Acer—Epigaea; TM = T suga- Maianthemum .................................................................................................................... 71 Table 1.19. Average number of stems and standard errors for overstory tree species within aspen stands in the SO-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. ATD = Acer- Tsuga—Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum ............................................ 73 Table 1.20. Average number of stems and standard errors for overstory tree species within aspen stands in the 70-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer—Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer— Viola-Osmorhiza; TM = T suga-Maianthemum ..................................................... 75 Table 1.21. Average basal area and standard errors for overstory tree species within aspen stands in the 20-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccim'um; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola- Osmorhiza; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum ........................ 77 Table 1.22. Average basal area and standard errors for overstory tree species within aspen stands in the 50-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum ........................................................ 79 Table 1.23. Average basal area and standard errors for overstory tree species within aspen stands in the 70-year age class in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer—Quercus— Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola- Osmorhiza; TM = T suga-Maianthemum ........................................................................... 81 Table 1.24. Average volume (cords/ac and m3/ha) and standard errors for overstory tree species within aspen stands in the 50— and 70-year age classes in different habitat types on state land in the western Upper Peninsula, Michigan during 2004—2006. Volume in m3/ha was calculated by converting the cords/ac values recorded in the Michigan Department of Natural Resources Operations Inventory database. One cord = 3.62 m3 and includes bark, wood, and air space in a 4-ft x 4-fi x 100-in stack of logs. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza; QAE = Quercus-Acer—Epigaea; TM = T suga-Maianthemum ......................................................................................................... 84 xii CHzfl’T ER 2 Table 2.1. Forest attributes that are important for feeding and reproductive requirements of bird species that may use various age classes of aspen to meet their life requisites. Dbh = diameter at breast height. DWD = down woody debris ....................................... 111 Table 2.2. Structure of forest attributes and habitat requirements for bird species that may use various age classes of aspen to meet their life requisites. Dbh = diameter at breast height. DWD = down woody debris ..................................................................... 113 Table 2.3. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in 3 age classes of aspen during the breeding season in the Upper Peninsula, Michigan in 2004—2005 .................................. 121 Table 2.4. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in 3 habitat types supporting aspen during the breeding season in the Upper Peninsula, Michigan in 2004—2005. TM = T suga- Maianthemum, ATD = Acer—Tsuga-Drypoteris, AVO = Acer- Viola-Osmorhiza ........... 124 Table 2.5. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 20-year age class within different habitat types during the breeding season in the Upper Peninsula, Michigan in 2004—2005. QAE = Quercus-Acer-Epigaea, AQV = Acer-Quercus— Vaccim'um, TM = T suga-Maianthemum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola- Osmorhiza ........................................................................................................................ l 29 Table 2.6. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 50-year age class within different habitat types during the breeding season in the Upper Peninsula, Michigan in 2004—2005. TM = T suga-Maianthemum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Osmorhiza .............................................................................................................. l 3 1 Table 2.7. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 70-year age class within different habitat types during the breeding season in the Upper Peninsula, Michigan in 2004—2005. AQV = Acer-Quercus- Vaccinium, TM = T suga-Maianthemum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer— Viola-Osmorhiza, AOC = Acer-Osmorhiza- Caulophyllum ................................................................................................................... 133 Table 2.8. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen within the T suga-Maianthemum habitat type during the breeding season in the Upper Peninsula, Michigan in 2004— 2005 .................................................................................................................................. 137 xiii Table 2.9. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen within the Acer— T suga-Drypoteris habitat type during the breeding season in the Upper Peninsula, Michigan in 2004— 2005 .................................................................................................................................. 139 Table 2.10. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen within the Acer- Viola—Osmorhiza habitat type during the breeding season in the Upper Peninsula, Michigan in 2004— 2005 .................................................................................................................................. 141 Table 2.11. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in 3 aspen age classes during the winter (F ebruary—March) in the Upper Peninsula, Michigan in 2005-2006 .............................. 162 Table 2.12. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 20-year age class within different habitat types during the winter (February-March) in the Upper Peninsula, Michigan in 2005—2006. AQV = Acer-Quercus- Vaccinium, TM = T suga-Maianthemum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer- V iola-Osmorhiza, AOC = Acer—Osmorhiza- Caulophyllum ................................................................................................................... l 63 Table 2.13. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 50-year age class within different habitat types during the winter (February—March) in the Upper Peninsula, Michigan in 2005—2006. TM = T suga-Maianthemum, AVO = Acer- Viola- Osmorhiza ........................................................................................................................ 164 Table 2.14. Mean absolute frequencies (percent of points at which species occurred) and standard errors (SE) of bird species surveyed in aspen in the 70-year age class within different habitat types during the winter (F ebruary—March) in the Upper Peninsula, Michigan in 2005—2006. AQV = Acer-Quercus- Vaccim'um, TM = T suga-Maz'anthemum, AVO = Acer- Viola-Osmorhiza ........................................................................................ 165 CHAPTER 3 Table 3.1. Estimation of selection indices for the occurrence of drumming ruffed grouse in old (2 50 years), medium-aged (30—49 years), and young (< 30 years) aspen, or non- aspen vegetation types in the western Upper Peninsula, Michigan during 2005 and 2006 .................................................................................................................................. 192 Table 3.2. Estimation of selection indices for the occurrence of drumming ruffed grouse in different habitat types in the western Upper Peninsula, Michigan during 2005, 2006, and combined years. AQV = Acer-Quercus- Vaccinium; TM = T suga-Maianthemum; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; AOC = Acer-Osmorhiza- Caulophyllum; FMC = Fraxinus-Mentha—Carex; TTS = T suga-Thuga-Sphagnum. AQV and TM habitat types occur on dry, sandy soils, ATD occurs on sandy-loamy soils, AVO, xiv AOC, and FMC habitat types occur on mesic, loamy soils, and TTS occurs on hydric organic soils ..................................................................................................................... 193 Table 3.3. Estimation of selection indices for the occurrence of drumming ruffed grouse in different age classes and habitat types (HT) in the western Upper Peninsula, Michigan during 2005. Age classes include aspen 2 50 years (old), medium-aged aspen (30—49 years; med), young aspen < 30 years (yng), and non-aspen (n-a) vegetation types. AQV = Acer-Quercus- Vaccinium; TM = T suga-Maianthemum; AVO = Acer- Viola—Osmorhiza; AOC = Acer-Osmorhiza-Caulophyllum; FMC = Fraxinus-Mentha—Carex; TTS = T suga- Thuga-Sphagnum. AQV and TM habitat types occur on dry, sandy soils, AVG and AOC habitat types occur on mesic, loamy soils, and TTS occurs on hydric organic soils ....... 197 Table 3.4. Estimation of selection indices for the occurrence of drumming ruffed grouse in different age classes and habitat types (HT) in the western Upper Peninsula, Michigan during 2006. Age classes include aspen Z 50 years (old), medium-aged asan (30—49 years; med), young aspen < 30 years (yng), and non-aspen (n-a) vegetation types. AQV = Acer-Quercus- Vaccinium; TM = T suga-Maianthemum; ATD = Acer— T suga—Dryopteris; AVO = Acer- Viola-Osmorhiza; AOC = Acer-Osmorhiza-Caulophyllum; F MC = Fraxinus-Mentha-Carex; TTS = T suga-Ihuga-Sphagnum. AQV and TM habitat types occur on dry, sandy soils, ATD occurs on sandy-loamy soils, AVO, AOC, and FMC habitat types occur on mesic, loamy soils, and TTS occurs on hydric organic soils ....... 198 Table 3.5. Estimation of selection indices for the occurrence of drumming ruffed grouse in different age classes and habitat types (HT) in the western Upper Peninsula, Michigan during 2005 and 2006. Age classes include aspen 2 50 years (old), medium-aged aspen (30—49 years; med), young aspen < 30 years (yng), and non-aspen (n—a) vegetation types. AQV = Acer-Quercus- Vaccinium; TM = T suga-Maianthemum; ATD = Acer—Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; AOC = Acer-Osmorhiza-Caulophyllum; FMC = F raxinus-Mentha-Carex; TTS = T suga-Ihuga-Sphagnum. AQV and TM habitat types occur on dry, sandy soils, ATD occurs on sandy-loamy soils, AVO, AOC, and F MC habitat types occur on mesic, loamy soils, and TTS occurs on hydric organic soils .................................................................................................................................. 199 CHAPTER 4 Table 4.1. Average volume (cords) per acre of aspen for different rotation ages and habitat types within the western Upper Peninsula, Michigan. AOC = Acer-Osmorhiza— Caulophyllum; AQV = Acer-Quercus— Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccim'um-spp.; QAE = Quercus-Acer—Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum- Coptis ............................................................................................................................... 220 Table 4.2. Regression equations to predict volume (cords/acre) yields for aspen on 40-, 50-, 60—, 70-, and 80-year rotations in different habitat types in the western Upper Peninsula, Michigan. Volumes were determined from 2006 Operations Inventory data from the Michigan Department of Natural Resources. AOC = Acer-Osmorhiza- XV Caulophyllum; AQV = Acer—Quercus- Vaccim'um; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola—Osmorhiza; F = F marinas-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus—Acer—Epigaea; TM = T suga-Maz'anthemum; TMC = T suga-Maianthemum- Coptis ............................................................................................................................... 220 Table 4.3. Number of acres of aspen within 8 age classes in 9 different habitat types used to initialize an aspen harvest scheduling simulation model for a 300 mi2 area in the western Upper Peninsula, Michigan. Data were obtained fi'om 2006 Operations Inventory Data from the Michigan Department of Natural Resources. AOC = Acer- Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga- Dryopteris; AVO = Acer- V iola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccim'um- spp.; QAE = Quercus-Acer—Epigaea; TM = T suga-Maianthemum; TMC = T suga- Maianthemum-Coptis ....................................................................................................... 221 Table 4.4. Selection indices (W) for ruffed grouse drumming habitat in 3 habitat types (ATD, AVO, and TM) and 3 aspen age classes in the western Upper Peninsula, Michigan. AOC = Acer—Osmorhiza-Caulophyllum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Ozmorhiza, TM = T suga-Maianthemum. Aspen age classes included young (< 30 years old), medium (30—49 years old), and old (2 50 years old). Data on ruffed grouse habitat selection were collected in the spring of 2005 and 2006 on state lands in Baraga, Dickinson, Iron, and Marquette counties. Selection indices were determined based on methods described by Manly et a1. (2002) ........................................................ 224 Table 4.5. Management unit codes for specific land ownership and habitat types (HT) used as inputs for the HARVEST simulation model. NIPF = non-industrial private forests ............................................................................................................................... 230 Table 4.6. Aspen harvest schedule (number of acres harvested) for state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest all aspen by 80 years .................................................................. 242 Table 4.7. Number of cords of aspen harvested on state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest all aspen by 80 years ........................................................................................... 245 Table 4.8. Aspen harvest schedule (number of acres harvested) for state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest all aspen 2 60 years .................................................................... 261 Table 4.9. Number of cords of aspen harvested on state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest all aspen Z 60 years ............................................................................................. 264 xvi Table 4.10. Aspen harvest schedule (number of acres harvested) for state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest 60% of all aspen on a 40 year rotation. Total aspen area did not decrease ............................................................................................................................ 279 Table 4.11. Number of cords of aspen harvested on state-owned lands within different habitat types in the western Upper Peninsula, Michigan over 10 decades to establish and maintain an even-age class distribution within habitat types. Harvesting strategy was to harvest 60% of all aspen on a 40-year rotation. Total aspen area did not decrease ........ 282 CHAPTER 5 Table 5.1. Comparison of consequences of different aspen management strategies simulated for a 300 mi2 area in the western Upper Peninsula, Michigan. Strategies included: harvest all aspen by 80 years, harvest all aspen > 60 years, and intensively harvest aspen (i.e., > 60% of volume harvested on a 40-year rotation) ........................... 305 xvii LIST OF FIGURES *Some of the figures in this document are presented in color. CHAPTER I Figure 1.1. Location of a 12,3000 km2 study area in the western Upper Peninsula of Michigan .............................................................................................................................. 8 Figure 1.2. Example of a stand (represented by a dashed ellipse) that overlaps boundaries of 2 habitat types. Samples (x) within such stands would not represent the range of conditions of one specific age class and habitat type. If stand boundaries were ignored for analysis, sampled points would not be independent and thus would not represent the ecological conditions of their respective treatments .......................................................... 22 Figure 13*. Spatial distribution of habitat types in a 12,300 km2 area in the western Upper Peninsula, Michigan as determined from literature and field data collected between 2004—2006. The study area consisted of Baraga, Dickinson, Iron, and Marquette counties. Dark lines indicate boundaries of state forest land. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccim'um; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F I = F raxinus-Impatiens; FMC = F raxinus-Mentha-Carex; PCS = Picea-Chamadaphne-Sphagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsz’a; QAE = Quercus-Acer-Epigaea; TAM = T suga-Acer—Mitchella; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis; TMV = T suga- Maianthemum- Vaccinium; TTM = T suga-Thuja-Mitella; TTS = T suga-Thuja- Sphagnum ........................................................................................................................... 26 Figure 1.4. Percent of aspen within different age classes on state land in the western Upper Peninsula, Michigan based on Operations Inventory data from the Michigan Department of Natural Resources (2006) .......................................................................... 28 Figure 1.5. Comparison among percent of study area (western Upper Peninsula, Michigan), percent aspen on state land, and percent of state land within different habitat types as determined from literature and field data collected between 2004—2006. The study area consisted of Baraga, Dickinson, Iron, and Marquette counties. AOC = Acer- Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza; FI = F marinas-Impatiens; FMC = F raxinus— Mentha-Carex; PCS = Picea—Chamadaphne-Sphagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer-Epigaea; TAM = T suga- Acer-Mitchella; TM = Tsuga-Maianthemum; TMC = T suga-Maianthemum-Coptis; TMV = T suga-Maianthemum- Vaccinium; TTM = T suga-Thuja-Mitella; TTS = T suga-Thuja- Sphagnum ........................................................................................................................... 31 Figure 1.6. Cumulative proportion of aspen within different age classes and habitat types on state land in the western Upper Peninsula, Michigan. Habitat types were determined from spatial habitat type delineation from literature (Coffman et a1. 1980) and field data xviii collected between 2004—2006. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer- Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F I = F raxinus-Impatiens; FMC = F raxinus-Mentha-Carex; PCS = Picea-Chamadaphne- Sphagnum; PVC = Pinus- Vaccim'um-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer-Epigaea; TAM = T suga-Acer-Mitchella; TM = T suga- Maianthemum; TMC = Tsuga-Maianthemum-Coptis; TMV = T suga-Maianthemum- Vaccim'um; TTM = T suga-Thuja-Mitella; TTS = T suga-Thuja-Sphagnum ...................... 34 Figure 1.7. Relative differences in overstory tree species presence and dominance across habitat types sampled on state lands in the western Upper Peninsula, Michigan during 2004—2006. The location of bars represents habitat t ypes in which tree species are present. Darker shades indicate dominance. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinr'um; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola- Osmorhiza; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum ....................... 90 Figure 1.8. Summary of vegetation structural differences among aspen stands within 3 age classes and habitat types in the western Upper Peninsula, Michigan. Vegetation was sampled in 2004—2006. ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Osmorrhiza, and TM = T suga-Maianthemum ........................................................................................ 91 Figure 1.9. Summary of differences in quaking aspen stem density among different age classes and habitat types that can support aspen. Vegetation was sampled in 2004—2006. ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Osmorrhiza, and TM = T suga- Maianthemum .................................................................................................................... 92 Figure 1.10. Summary of vegetation structural differences among aspen stands within 3 age classes and habitat types in the western Upper Peninsula, Michigan. Variables presented are important in distinguishing habitat types. As aspen stands age, more variables become important in distinguishing vegetation types. Vegetation was sampled in 2004—2006. ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Osmorrhiza, and TM = T suga-Maianthemum ...................................................................................................... 93 CHAPTER 2 Figure 2.1*. Scores for the first 3 principal components describing the variability among 25 vegetation variables measured in aspen stands within 20-year, 50-year, and 70-year age classes occurring within 3 different habitat types in the western Upper Peninsula, Michigan in 2005. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum .................................................................................................... 1 17 Figure 2.2. Number of cavity-nesting and non-cavity-nesting birds in 20-year (A), 50- year (B), and 70-year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus— Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola- Osmorhiza; TM = T suga-Maianthemum ......................................................................... 144 xix Figure 2.3. Number of year-round resident birds, short-distance migrants, and neotropical migrants in 20-year (A), 50-year (B), and 70-year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccim’um; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum ........... 147 Figure 2.4. Habitat suitability scores for bird species in 20-year (A), 50-year (B), and 70- year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer- Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer— Viola-Osmorhiza; TM = T suga-Maianthemum .................................................................................................... 151 Figure 2.5. Avian community associations with 20—50-year-old aspen within the ATD and AVO habitat types in the western Upper Peninsula, Michigan. TM = T suga- Maianthemum, ATD = Acer- Tsuga-Drypoteris, and AVO = Acer- Viola- Osmorhiza ........................................................................................................................ 154 Figure 2.6. Avian community associations with 50-year-old aspen within the ATD and AVO habitat types in the western Upper Peninsula, Michigan. TM = Tsuga- Maianthemum, ATD = Acer-Tsuga-Drypoteris, and AVO = Acer- Viola- Osmorhiza ........................................................................................................................ 155 Figure 2.7. Avian community associations with 50-year-old aspen in the western Upper Peninsula, Michigan. TM = T suga-Maianthemum, ATD = Acer-Tsuga-Drypoteris, and AVO = Acer- Viola-Osmorhiza ........................................................................................ 156 Figure 2.8. Avian community associations with 20-year-old aspen within the Tsuga- Maianthemum (TM) habitat type in the western Upper Peninsula, Michigan. ATD = Acer- T suga-Drypoteris, and AVO = Acer- Viola-Osmorhz'za .......................................... 157 Figure 2.9. Avian community associations with aspen 2 70 years old within the T suga- Maianthemum (TM) habitat type in the western Upper Peninsula, Michigan. ATD = Acer-Tsuga-Drypoteris, and AVO = Acer- Viola-Osmorhiza .......................................... 158 Figure 2.10. Avian community associations with aspen within the T suga-Maianthemum (TM) habitat type in the western Upper Peninsula, Michigan. ATD = Acer- T suga— Drypoteris, and AVO = Acer- Viola-Osmorhiza .............................................................. 159 Figure 2.11. Avian community associations with aspen within multiple age classes and habitat types habitat type in the western Upper Peninsula, Michigan. TM = T suga- Maianthemum, ATD = Acer-Tsuga-Drypoteris, and AVO = Acer- Viola- Osmorhiza ........................................................................................................................ l 60 CHAPTER 3 Figure 3.]. Cumulative percent of aspen age classes and habitat types used by drumming ruffed grouse in 2005 (A) and available (B) in the western Upper Peninsula, Michigan. Habitat types listed include those that occur within Baraga, Dickinson, Iron, and Marquette counties. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccim‘um; ATD = Acer- T suga-Dryopterz’s; AVO = Acer- Viola-Osmorhiza; FI = Fraxinus-Impatiens; FMC = Fraxinus-Mentha-Carex; PCS = Picea-Chamadaphne- Sphagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer—Epigaea; TAM = T suga-Acer-Mitchella; TM = T suga- Maianthemum; TMC = Tsuga-Maianthemum-Coptis; TMV = T suga-Maianthemum- Vaccinium; TTM = T suga-Thuja-Mitella; TTS = T suga-Thuja-Sphagnum .................... 195 Figure 3.2. Cumulative percent of aspen age classes and habitat types used by drumming ruffed grouse in 2006 (A) and available (B) in the western Upper Peninsula, Michigan. Habitat types listed include those that occur within Baraga, Dickinson, Iron, and Marquette counties. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; FI =- F raxinus-Impatiens; FMC = F raxinuS-Mentha-Carex; PCS = Picea-Chamadaphne- Sphagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer-Epigaea; TAM = T suga-Acer-Mitchella; TM = T suga- Maianthemum; TMC = Tsuga-Maianthemum-Coptis; TMV = T saga-Maianthemum- Vaccim'um; TTM = T suga-Thuja-Mitella; TTS = T suga-Thuja-Sphagnum .................... 196 Figure 3.3. Spatial distribution of habitat types ruffed grouse would potentially select for drumming in the western Upper Peninsula, Michigan. Selection indices > 1.0 indicate selection; those < 1.0 indicate avoidance. The area includes Baraga, Dickinson, Iron, and Marquette counties. This figure was based on data collected during mid-April in 2005 and 2006 ........................................................................................................................... 200 CHAPTER 4 Figure 4.1. Location of study area (approximately 300 miz) in the western Upper Peninsula, Michigan for timber harvest simulations. Black areas are public lands managed by the Michigan Department of Natural Resources, dark gray areas are industry-owned forests, and light gray areas are non-industrial private forests .............. 214 Figure 4.2. Schematic diagram of harvest scheduling model structure. Boxes in bold are model outputs ................................................................................................................... 222 Figure 4.3. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). All aspen was harvested by 80 years old. AQV = Acer—Quercus- Vaccinium; Pinus- Vaccinium-spp.; QAE = Quercus-Acer—Epigaea ............................... 235 Figure 4.4. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest xxi (n = 5 simulations). All aspen was harvested by 80 years old. ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum ...................... 236 Figure 4.5. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza- Caulophyllum; F = Fraxinus-spp.; TMC = T suga-Maianthemum-Coptis ....................... 237 Figure 4.6. Number of deficient acres of aspen needed to establish and maintain an even- age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccim'um; ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccim'um- spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga- Maianthemum-Coptis ....................................................................................................... 238 Figure 4.7. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. All aspen was harvested by 80 years old ............................................................................... 239 Figure 4.8. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccim'um-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis ............................................................................................ 240 Figure 4.9. Volume (cords) to be harvested in aspen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus— Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis ............................................................................... 241 Figure 4.10. Expected percent changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. Percents may sum to > 100 because some communities overlap age classes or habitat types. ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with 20—50-year-old aspen in the ATD and AVO habitat types, aspen Z 50 years within the ATD and AVO types, aspen Z 50 years old, asan in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type ................................................................... 248 xxii Figure 4.11. Expected changes in ruffed grouse drumming activity (number of drumming grouse) within preferred habitat types and collectively at the end of each of 10 decades following simulated aspen harvest in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. [Note: these projected trends do not consider natural cyclic fluctuations in grouse populations] AOC = Acer-Osmorhiza-Caulophyllum; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga- Maianthemum .................................................................................................................. 249 Figure 4.12. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen Z 60 was harvested. AQV = Acer-Quercus- Vaccinium; Pinus- Vaccinium-spp.; QAE = Quercus-Acer—Epigaea .................................................. 254 Figure 4.13. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen 2 60 was harvested. ATD = Acer- T suga-Dryopterz‘s; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum .......................................... 255 Figure 4.14. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the begimring of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen _>_ 60 was harvested. AOC = Acer-Osmorhiza- Caulophyllum; F = Fraxinus-spp.; TMC = T suga-Maianthemum-Coptis ....................... 256 Figure 4.15. Number of deficient acres of aspen needed to establish and maintain an even-age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. Only aspen 2 60 was harvested. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga- Maianthemum-Coptis ....................................................................................................... 257 Figure 4.16. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. Only aspen 2 60 was harvested ................................................................................................. 258 Figure 4.17. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Only aspen Z 60 was harvested. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Mar'anthemum; TMC = T suga- Maianthemum-Coptis ....................................................................................................... 259 Figure 4.18. Volume (cords) to be harvested in aspen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class xxiii distribution within habitat types in the Upper Peninsula of Michigan. Only aspen 2 60 was harvested. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-Spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis .......................................... 260 Figure 4.19. Expected changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Only aspen Z 60 years old was harvested. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with 20—50-year-old aspen in the ATD and AVO habitat types, aspen 2 50 years within the ATD and AVO types, aspen Z 50 years old, aspen in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type .......................................................................................................... 267 Figure 4.20. Expected changes in ruffed grouse drumming activity (number of grouse drumming) within preferred habitat types and collectively following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Only aspen 2 60 years old was harvested. [Note: these projected trends do not consider natural cyclic fluctuations in grouse populations] AOC = Acer-Osmorhiza-Caulophyllum; ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza ..................................................................... 268 Figure 4.21. Average proportion of asan within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. AQV = Acer-Quercus- Vaccinium; Pinus- Vaccinium-spp.; QAE = Quercus-Acer- Epigaea ............................................................................................................................ 272 Figure 4.22. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola—Osmorhiza; TM = T suga- Maianthemum .................................................................................................................. 273 Figure 4.23. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; F = Fraxinus-spp.; TMC = Tsuga-Maianthemum- Captis ............................................................................................................................... 274 Figure 4.24. Number of deficient acres of aspen needed to establish and maintain an even-age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = xxiv Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis .......................................... 275 Figure 4.25. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. Most (60%) of the aspen was harvested on a 40-year rotation ................................................. 276 Figure 4.26. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T saga-Maianthemum-Coptis .......................................... 277 Figure 4.27. Volume (cords) to be harvested in aspen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola- Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer- Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum-Coptis ................. 278 Figure 4.28. Expected changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with 20—50-year-old aspen in the ATD and AVO habitat types, aspen Z 50 years within the ATD and AVO types, aspen Z 50 years old, aspen in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type ................................................................... 285 Figure 4.29. Expected changes in ruffed grouse drumming activity (number of drumming grouse) within preferred habitat types and collectively following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. [Note: these projected trends do not consider natural cyclic fluctuations in grouse populations] AOC = Acer-Osmorhiza-Caulophyllum; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga- Maianthemum .................................................................................................................. 286 Figure 4.30. Amount of forested area classified as interior and edge after 10 decades of simulated aspen harvest in the western Upper Peninsula, Michigan under 3 management scenarios: harvest all aspen by 80 years, harvest aspen > 60 years, and a more intensive harvest where 60% of all aspen harvested was in the 40-year age class. Edge area was defined as area within 295 feet (90 m) of a forest edge ................................................... 288 XXV Figure 4.31”“. Maps of forest interior (dark shades) and edge (light shades) in a 300 mi2 area in the western Upper Peninsula, Michigan during decade 3 of a 10-decade simulated aspen harvest where all aspen > 60 years old was harvested (A; more interior), and where Z 60% of the aspen harvested was from the 40-year age class (B; less interior). Edge was considered any area within 295 ft (90 m) of a forest edge ............................................... 289 Figure 4.32. Areas (indicated in black) where aspen is expected to be lost due to succession in the next 10 decades in a 300 mi2 area in the western Upper Peninsula. White areas indicate state land, light gray areas are commercial forest land, and dark gray areas are non-industrial private forests ............................................................................ 297 xxvi INTRODUCTION Wildlife biologists and managers face many issues related to the utilization of forest resources and wildlife relationships within forest ecosystems. Historically, forest management focused primarily on organizing and managing lands to grow and harvest timber products (Kessler et a1. 1992, Nyland 2002). Since the mid 19603, however, a new philosophy of forest management has developed, which includes managing forests to produce timber and sustain viable forest ecosystems for multiple purposes (e. g., wildlife habitat, recreatiOn, timber) (Nyland 2002). Sustaining forest ecosystems to meet a diversity of forest and wildlife objectives is a challenging task for natural resource managers. To meet this challenge, managers need to enhance their understanding of the ecological factors that influence vegetation structure, composition, and distribution; how specific forest manipulations and natural disturbances affect temporal changes in forest characteristics across landscapes; and subsequently, how wildlife that depend on these dynamic systems respond. DeStefano (2002) reported that there is a national concern among biologists about the alteration of forest vegetation structure due to timber harvesting and its effects on wildlife populations and communities. To address this concern, a high research priority includes obtaining information on wildlife Species’ life requisites and habitat relationships within forests at multiple spatial (DeStefano 2002) and temporal scales (Bissonette and Storch 2007). There is still much information to be learned about forests, particularly in forests that are not well represented in the landscape, to understand their ecological contributions and functions at multiple levels (i.e., stand, community, landscape). This information especially becomes critical as agencies develop . (p; .\"" a. .t. -z' a" ‘ I 6.1“" 1‘\\ ." Cr. ~u-.- ‘ 1... ‘~-o u. .9.‘ 'I- management plans within an ecosystem management framework. That is, one which considers maintaining ecological processes and properties of systems at multiple spatial scales, and considers social and economic objectives. Kessler et a1. (1992), for example, explained that a major challenge facing natural resource scientists is how to sustain natural systems and human commodities in view of the growing human population and its associated demands on natural resources. Therefore, natural resources management and research should focus on understanding function and processes that occur within systems and how they are affected by patterns of resource use and cumulative effects of management (Kessler et a1. 1992). Development and refining of ecosystem management paradigms must incorporate methods to sustain ecosystems but not diminish the importance of products produced by those systems. Aspen (Populus spp.), for example, is a commercially valuable timber resource that is used to produce pallets, plywood, and wood pulp for paper, cardboard, and boxes. Aspen [quaking (P. tremuloides) and big-tooth (P. grandidentata)] comprises 47% of roundwood (i.e., logs, bolts, or other round sections cut from trees) harvested for pulpwood, which in turn constitutes more than half the industrial timber products harvested annually in the Lake States (Piva 2003). The aspen forest type in the western Upper Peninsula (UP) of Michigan ranks second to the beech-maple forest type in area, stem density, and net standing volume (Leatherberry 1994). As such, aspen is important in sustaining pulpwood production in the western UP, which was the top pulpwood- producing region of Michigan in 2002 (i.e., 43% of roundwood pulpwood harvested in Michigan came from this region) (Piva 2003). Additionally, 308,000 cords of aspen pulpwood is produced annually from the western UP (Piva 2003) and is worth approximately $18.5 million [at a value of $60 per cord delivered to the mill (Miller 1 998)] . Aspen is a valuable resource not only economically, but also ecologically. Several wildlife species, for example, depend on aspen to meet their life requisites (Stelfox 1995). For instance, ruffed grouse (Bonasa umbellus) utilize catkins from mature male aspen as their staple winter food source (Gullion and Svoboda 1972). White-tailed deer (Odocoileus virginianus) rely heavily on aspen leaves and new shoots for spring and summer foraging (Kohn and Mooty 1971). Parsons et al. (2003) found that bat species in British Columbia depend on the decay-causing fungus (Phellinus tremulae) in old aspen to create cavities for roosting habitat. Pojar (1995) found bird communities including neotropical migrants to be associated with specific aspen seral stages. Other wildlife such as woodcock (Scolopax minor), beaver (Castor canadensz's), woodpeckers, and snowshoe hares (Lepus amerz'canus) also obtain some life requisites from aspen stands. Although aspen may live past 100-years old, most of the aspen in Michigan show signs of deterioration over 60 to 70-years old and lose their commercial value (Graham 1963). If aspen is not harvested, shade tolerant species [e.g., red maple (Acer rubrum), American beech (Fagus grandifolia), white pine (Pinus strobus), balsam fir (Abies balsamea)] eventually will dominate the stands. Letting aspen succeed to older age classes where shade tolerant species dominate may make aspen regeneration difficult on those sites. Because aspen is commercially valuable, however, most aspen in the Great Lakes Region are harvested on 40- to 60-year-old rotations depending on site quality (Brinkman and Roe 1975, Nyland 2002). Most aspen stands in the western UP are in the 1 1 to 20-year-old (27%) and 21 to 30-year-old (22%) age classes, respectively (Doepker et al. 2002). Other age classes of aspen are not as well represented in the landscape. Approximately 8% of all aspen in Iron and Dickinson counties, for instance, is 40 to 60- years old; 20% of aspen is 60 years or older; some may be inaccessible for harvesting because it occurs in lowland habitat types. To increase the diversity of aspen age classes for future timber and wildlife values, Doepker et al. (2002) recommended increasing the amount of 40 to 60-year-old aspen to represent 40% of aspen in the landscape. Although the timber value of older ages classes depreciates, old aspen stands still are likely to provide critical wildlife habitat components (Stelfox 1995). Because 40 to 60-year-old aspen stands are typically harvested, older age classes of aspen are not as well represented in the landscape and wildlife values and ecosystem integrity associated with occurrence of these age classes, in turn, may not be represented. Subsequently, Michigan Department of Natural Resource (MDNR) managers in the western UP are developing plans to maintain a diversity of aspen age classes within the landscape to sustain timber yield, wildlife values, ecosystem integrity, and social and economic values associated with the aspen forest (R. Doepker, MDNR, personal communication). In addition to maintaining a diversity of aspen age classes in landscapes, another one of Michigan’s aspen management objectives is to maintain the old aspen component for wildlife use, biodiversity, and to meet ecosystem-management goals without compromising timber harvest yield (Doepker et a1. 2002). It would be beneficial, then, to quantify the structure and composition of old aspen (> 50 years), determine its importance to wildlife and its role in ecosystems, and model the effects of aspen harvest strategies on wildlife species and communities. This information would help managers understand the impacts of applied aspen management practices on biological processes and the wildlife communities which depend on them. This information will also assist managers in managing Michigan’s forest at a sustainable level and maintain certification through the Forest Stewardship Council and the Sustainable Forestry Initiative. The following are the overall goals of this project: 1) Quantify landscape composition, distribution and temporal changes of habitat types that could support aspen in the western UP of Michigan. 2) Determine how habitat suitability, and thus wildlife species and communities respond to changes in aspen management practices at the stand and landscape levels. 3) Quantify the effects of harvest simulations on wildlife habitat suitability, and selected wildlife species (e.g., white-tailed deer, ruffed grouse), communities (e.g., plant, avian), and timber production. The first chapter of this dissertation describes procedures used to quantify landscape composition, distribution of habitat types, and structure and composition of aspen within different age classes and habitat types. The information presented demonstrates the value of understanding spatial patterns and temporal changes of aspen forests in the western Upper Peninsula for planning ecosystem-based management obj ectives. The second and third chapters quantify differences in wildlife communities, Speeies, and timber potential within aspen in different habitat types and age classes. ThrE=se chapters relate vegetation structure and composition within various aspen stands to p0pulation-level performance measures such as ruffed grouse drumming activity; SPetties-level measures such as presence/absence and frequency of occurrence; and coImmunity-level measures such as songbird species richness. In the fourth chapter, I describe development of aspen harvest simulation models and projections of harvest strategies on wildlife communities over time and discuss the importance of these simulations for addressing Michigan’s ecosystem management objectives. In this chapter, units of measurement are expressed in English units to reflect use by forest managers who typically use English units. The representation of data in di fferent units reflects use of the research results by multiple stakeholders. STUDY AREA The study area was a 12,300 km2 area located in the western Upper Peninsula (UP) of Michigan, which includes Baraga, Dickinson, Iron, and Marquette counties (Figure 1.1). Sites selected for assessment were located within the Escanaba River State Forest and the Copper Country State Forest. This region of Michigan accumulated 76.2- 86.4 cm of precipitation annually and averaged 4.5 °C (Sommers 1977). Soils in the western UP were dominantly spodosols and bedrock rich in iron occurred throughout much of the area. Well-drained sands and loams typically supported hardwood forests dominated by maple (Acer spp.) in late-successional stages. Spruce (Picea spp.) and balsam fir (Abies balsamea) occurred frequently as components of hardwood stands growing on sandy Soils. Other sub-dominant species in hardwood stands on loamy soils included basswood (T iIz'a americana), ironwood (Carpinus caroliniana), or American elm (Ulmus amerz'cana). Dry, sandy soils supported oak (Quercus spp.), white pine (Pinus strobus), and red pine (Pinus resinosa). Poorly-drained areas were interspersed throughout the r egion and were typically dominated by northem-white cedar (Thuja occidentalis), SPFLIce, fir, and tarnarack (Larix laricina; Albert 1986). Early successional and aspen (Populus spp.)-dominated forests were also interspersed throughout much of the area (Sommers 1977). Prior to European settlement, natural disturbances in the western UP were most fr‘Xluently from wind or ice storms, which created gaps within stands. Less common Stanld-replacing disturbances such as tornadoes or other intense storm systems occurred at 1000-year intervals (Frelich and Lorimer 1991). Historical disturbance patterns resulted in presettlement forests dominated by mixed conifers interspersed with northern hardwoods, wetlands, and xeric pinelands. Changes in land use and disturbances over the past 150 years has increased the proportion of maple and aspen, and resulted in the decline of mesic conifers, beech (F agus grandifolia), and yellow birch (Betula alleghaniensis; Zhang et al. 2000). According to current land cover data (MDNR 2003), approximately 11% of the study area occurs in northern hardwood vegetation types, 13% is dominated by pines, and 20% occurs in upland mixed coniferous/deciduous vegetation (Table 1.1). This study focused on those areas in northern hardwood vegetation types. Figure 1.1. Location of a 12,3000 km2 study area in the western Upper Peninsula of Michigan. Table 1.1. Percent of study area (Baraga, Dickinson, Iron, and Marrquette counties) in different of cover types in the western Upper Peninsula, Michigan circa 1999. Cover TypeI PercentageI urban 1.95 agriculture 5.34 herbaceous openland 1.80 upland Shrub/low density trees 28.69 northern hardwood 10.56 oak association 0.54 aspen association 4.89 upland deciduous 5.48 pines 12.77 other upland conifers 3.08 upland mixed forest 20.45 water 0.14 lowland forest 3.83 floating aquatic 0.05 lowland shrub 0.42 100.00 1Information was obtained from IFMAP/GAP data, Michigan Department of Natural Resources, Forest, Mineral and Fire Division available at huMWw.mcgi.state.mi.us/mgdl/?rel=thext§gaction=thmname&cid=5&cat=Land+Cove r+2001. CHAPTER 1 QUANTIFYING COMPOSITION, DISTRIBUTION AND TEMPORAL CHANGES OF HABITAT TYPES THAT CAN SUPPORT ASPEN IN THE WESTERN UPPER PENINSULA, MICHIGAN The spatial and temporal distribution and availability of ecological resources in landscapes has important implications for wildlife and forest management. It is difficult, for instance, to understand the dynamic relationships between wildlife and their habitats without understanding the underlying regulatory mechanisms within landscapes and the processes by which habitat within landscapes changes over time. This type of information is especially critical as natural resources agencies develop management plans within an ecosystem management framework to sustain forests for multiple purposes. Current land-cover classifications and maps are used widely by natural resource managers and planners to understand wildlife-habitat relationships and plan management activities, but they do not identify vegetation structure, potential vegetation trends and successional dynamics, or vegetation types on distinctive soils that may have different Wildlife values (Felix et al. 2007). Consequently, it is difficult to use only land cover to evaluate wildlife species responses to management or to ecosystem changes because assumptions about potential vegetation and successional dynamics can lead to unrealistic predictions. Recently developed approaches using habitat type classifications allow evaluations of land-use and land-cover based on biotic and abiotic properties of ecosystems. Daubenmire (1966:296) was regarded as the first to define “habitat type”. He stated that habitat types have “equivalent climax potentialities” because they occur in areas with the same ecological, geological, and climatic attributes (Daubenmire 1966: 10 297). Habitat-type classifications group communities and their environments into categories useful for management interpretation (Kotar and Burger 2000). These classifications allow an understanding of successional trajectories and distribution of ecological communities that reflect inherent site capabilities, disturbances, and potential response to management. A useful approach for assessing spatial and temporal changes in vegetation is to use ecological classification systems (e.g., habitat-type classification) to describe the potential and current ecological conditions affecting vegetation patterns. Abiotic ecological characteristics such as climate, landforms, and soil characteristics (e.g., nutrient content, moisture, texture) influence differences in vegetation structure, composition, and successional patterns within different habitat types (Crawford 1950, Daubenmire 1966). Although the boundaries and dynamics of habitat types are not static, they define a relatively narrow range of environmental conditions (Kotar and Burger 2000) that can provide a basis for predicting vegetation change over time within natural successional pathways or as a result of certain land-use and management practices. Understanding temporal changes in vegetation distribution, composition, and structure is Critical for developing forest management models, which can be used for planning and evaluating effective practices to meet ecosystem management objectives. Because Michigan Department of Natural Resources (MDNR) managers in the western Upper Peninsula (UP) are developing plans to maintain a diversity of aspen age Classes for multiple management objectives, it would be beneficial to use habitat-type Classification systems to identify areas that can potentially support aspen, and determine how the structure and composition of aspen may change throughout time. Understanding ll the ecology of habitat types that can support aspen will help facilitate more ecologically- based management prescriptions. 12 OBJECTIVES Specific objectives for this chapter were to 1) Identify the distribution and potential successional pathways of habitat types in the western UP, Michigan. 2) Assess differences in aspen stand structure and composition within and among different age classes and habitat types. 3) Determine successional changes of aspen in different habitat types. 13 METHODS Landscape classification Areas that can potentially support aspen were identified from a habitat-type database that described ecological boundaries and successional trajectories for the western UP (Felix 2003). Habitat type boundaries were updated with spatial information ( IF MAP; Integrated Forest Monitoring and Assessment Prescription) that became available in 2003 (MDNR 2003), and validated with 70 random points ground-truthed during the 2004 summer field season. Habitat types were classified based on the Coffinan et al. (1980) guide to habitat type classification for the western UP. The distribution of current aspen stands throughout the study area was determined from IFMAP. Locations, and ages of specific aspen stands on state lands were identified from Operations Inventory datasets obtained from the MDNR. Most forest compartments have been digitally mapped within the study area. From these datasets, I used ArcView 3.2 and determined the proportion of state lands in different aspen age classes and habitat tYPes. Selection of stands and experimental design Aspen occurs as early successional vegetation in several habitat types in the Western UP. For example, habitat types characterized by poorly-drained soils that Support cedar (Ihuja canadensis), white Spruce (Picea glauca), black spruce (Picea I’Iclrr'ana), and balsam fir (Abies balsamea) in mid- to late-successional stages support aspen in early stages (Coffman et al. 1980). Aspen also represents early stages of habitat 14 types characterized by well-drained soils that support northern hardwoods (Coffman et al. 1 980). Aspen stands growing in lowland habitat types in the western UP likely will not undergo any forest manipulation because it is not feasible to operate logging equipment in these poorly-drained soils (R. Doepker, MDNR, personal communication). Therefore, aspen stands were selected in upland habitat types that have been logged in the past or may be logged in the future. To characterize the range of conditions of different age classes of aspen within different habitat types across the western region of the UP, I selected replicated stands of different age classes within distinct upland habitat types. The selected habitat types were named with a genus of a tree species that Showed the strongest tendency to dominate a Community on that site in the absence of disturbance, and the genus of characteristic understory species (Coffman et a1. 1980). Tsuga-Maianthemum (TM; hemlock-wild lily- Of-the-valley) occured on well-developed sands to loam textured mesic soils, Acer- Tsuga-Dryopteris (ATD; maple-hemlock-fem spp.) occured on very fine podizolized mesic to well-drained sand, and Acer- Viola Osmorhz'za (AVO; maple-violet-sweet Cicely) oCcured on loam to silt loam mesic to well-drained soils (Coffman et al. 1980) (Table 2). Within each habitat type, I investigated 3 age classes of aspen (20 to 29-years, 50 to 59- yeélrs, >70-years), with at least 3 replicate stands in each age class. The 20-year age class Was selected to represent conditions prior to typical harvest age. The 50-year age class Was selected to represent conditions at typical harvest age, and older aspen was selected to represent late-successional aspen. Specific stands for the study were selected based on similar management Objectives and dominant overstory type as indicated from the Operations Inventory data, 15 b I‘ . . '.._ ‘Cl ‘ I ._.. ”- a and similar interiorzedge ratios. In all selected stands, aspen was the management objective and dominant overstory type. I used FRAGSTATS (McGarigal et al. 2002) to calculate interiorzedge ratios for all stands in the Operations Inventory database. Potential stands to sample in the 2004—2005 and 2005—2006 field seasons were selected based on similar interiorzedge ratios to minimize differences deviating from edge effects. Potential stands were visited prior to final stand selection to assess accessibility and validate habitat type and age class assignment. Occasionally, a stand was eliminated as a candidate for analysis because I determined that it obviously occurred in a habitat type other than the 3 initially selected for this study, it was incorrectly classified as aspen, or it had been harvested. Different stands were selected for the 2004—2005 and 2005—2006 field seasons. In the initial study design (described above), I selected aspen stands within 3 upland habitat types that were the most predominant in the western UP landscape (as determined from the spatial habitat type database developed by Felix 2003). Following field data collection that quantified vegetation composition and structure, availability of IFMAP data, and initial data analysis, I determined that selected stands actually l‘epresented 6 distinct habitat types instead of only the original 3 (TM, ATD, AVO). Additional samples occurred in Acer-Osmorhiza-Caulophyllum (AOC; maple-sweet CiCely-blue cohosh), Acer-Quercus- Vaccinium (AQV; maple-oak-blueberry spp.), and Quercus-Acer-Epigaea (QAE; oak-maple-trailing arbutus) habitat types. The ADC habitat type occurred on loam to silt-loam mesic soils similar to those of the AVO habitat %e. The AQV and QAE habitat types occurred on more xeric, well-drained sandy soils. l6 ._—‘ Vegetation Sampling Several vegetation variables were selected to measure the range of ecological conditions among different aspen age classes and habitat types. Variables quantifying live vegetation structure included basal area (mZ/ha) and Stem density (#/ha) of overstory coniferous and deciduous trees, diameter and height of overstory trees, shrub height, shrub density, and percent vertical cover of coniferous and deciduous overstory trees, vegetation in the shrub layer (2—5m), and herbaceous vegetation. Variables quantifying structure of dead vegetation included decay, density, and volume (m3/ha) of down woody debris (DWD); decay, density, diameter (cm), and height (m) of snags; and decay, density, diameter (cm), and height (cm) of stumps. Dead vegetation was considered DWD if it was > 10 cm in diameter and > 30.5 cm long. Variables quantifying species Composition included stem density and basal area of each overstory tree species within each sampled age class and habitat type. Three points were randomly selected within each stand at which all sampling occurred. Points were > 100 m away from the edge and were considered the center of each sampling plot. I used a plot design and line intercepts to measure forest attributes. Density of snags and tree stems was measured in 10 x 50-m plots (Table 1.2). Shrub Stem density, density of stumps and down woody debris, and tree diameter was measured Within 10 x 25-m plots. I used a clinometer to measure height of trees within 10 x 25-m plots; shrub height was measured with a meter stick. Canopy cover was measured using the line-intercept method (Canfield 1941), with 25-m intercepts. Diameter at breast height (DBH) was measured with a Biltrnore stick. Basal area (mZ/ha) was determined using a 10-factor prism and also was calculated from DBH (diameter at breast height) and stem density data for comparison (Table 1.2). Decay status of snags, DWD, and stumps was determined following procedures from United States Department of Agriculture, Forest Service (Table 1.3). Decay status ranged from 1—5; an assignment of 1 indicated that bark was still intact and twigs were present, 2 indicated that twigs were absent and texture of wood was partially soft, 3 indicated little bark present and wood had a soft texture, 4 indicated that no bark was present and wood texture was small, blocky pieces, and 5 indicated that the wood was soft and powdery. To compare timber volume between and among age classes and habitat types, I j oined the Operations Inventory database with the habitat type database and determined volume (cords) within different stands. I calculated average volume (m3/ha, as calculated from number of cords/acre in the database; 1 cord = 3.63 m3 of bark, wood, and air Space) and standard errors for different age classes and habitat types. I used the Operations Inventory datasets to increase the sample size for analysis, and to use volume data that was collected by DNR personnel, because those data will continue to be Collected by DNR personnel for the Operations Inventory database. Stands in the 20-year age classéwere excluded from this analysis because they typically are not harvested that young. Data analysis In the initial study design, each stand represented the evaluation unit. It was asSumeclvthat each stand would represent characteristicsjof a specific habitat type and Stand boundaries would not overlap > 1 habitat type. After the first field season, hOwever, I observed that some stands overlapped 2 habitat types or vegetation types (Figure 1.2). For example, 2 of 3 sampling points within a stand may have represented SO-year-old aspen in a specific habitat type. The third point however, may have represented a different habitat type or vegetation type. Two primary problems occurred upon observing that some stands overlapped > 1 vegetation type or habitat type. First, averaging values of 3 samples taken within each stand would not accurately reflect the range of conditions of vegetation structure and composition within a specific age class and habitat type. Second, if I used data from only the 2 points that most represented characteristics of a specific age class and habitat type, I would be throwing out data and would not have adequate sample size (_>_ 3) for statistical analyses. Therefore, to address these problems, in 2005 I decided to make age classes of specific habitat types (versus stands) the evaluation unit, and the pre-established points Within different aspen age classes were considered the samples. A problem that emerged by making the evaluation unit age class of specific habitat types was that of spatial autocorrelation. Spatial autocorrelation essentially means that samples occurring near each other are more related than more distant samples. Because samples were initially selected with stands, they were “clustered” together Within a specific habitat type and age class (Figure 1.2). Clustering (or autocorrelation) may mean that all samples were not truly independent because they occurred in stands With potentially different management. Having non-independent samples violates the assumptions of independence for assessing differences using analysis of variance (ANOVA). However, because I initially selected stands with the same management 0bjective (i.e., aspen) and management history (i.e., clearcuts), I assumed points were Still independent. 19 I used the Shapiro-Wilk procedure to test for normalin of vegetation attribute data. Many of the variables tested were non-normally distributed, so I used appropriate nonparametric tests (i.e., the Kruskal-Wallis one-way ANOVA) to compare vegetation structure and composition among age classes and habitat types. Mean values and standard errors of vegetation attributes were calculated and presented for each age class and habitat type to make comparisons to data in the literature regarding wildlife-habitat relationships that may be occurring across the study region. 20 Table 1.2. Methods used to measure forest stand attributes in aspen stands in Michigan’s Upper Peninsula. Forest Attribute Method tree stem density (#/ha) 10 x 50 m plots shrub stem density (#/ha) 10 x 25 m plots canopy height (>5 m) Haga altimeter in 10 x 25 m plots shrub height (<5 m) meter stick in 10 x 25 m plots snag density (#/ha) 10 x 50 m plots % canopy cover (>5 m) 25 m line intercepts % shrub cover (<5m) 25 In line intercepts % herbaceous cover (<1 m) 20 m line intercepts tree dbh ( cm) 10 x 25 m plots density of stumps and DWD 10 x 25 m plots basal area (mZ/ha) calculated from stem density, dbh Table 1.3. Codes for determining decay status of down woody debris, snags, and stumps. These were adapted from decay class codes used by the United States Department of Agriculture, Forest Service. Code Description 1 Bark intact, twigs present. Texture is intact. Wood is original color. Bark intact, twigs absent. Texture is intact to partially soft. Wood is original 2 color. 3 Trace of bark. Twigs are absent and texture is hard large pieces. Color of wood is original to faded 4 Bark and twigs are absent. Texture of wood is small, soft, blocky pieces. Wood is light brown to faded brown or yellowish. 5 Bark and twigs are absent. Texture of wood is soft and powdery. Color of wood \ is faded to light yellow or gray. 21 —-- o s‘ Habitat type B Habitat type A - . ...... ........... Figure 1.2. Example of a stand (represented by a dashed ellipse) that overlaps boundaries 0f 2 habitat types. Samples (x) within such stands would not represent the range of conditions of one specific age class and habitat type. If stand boundaries were ignored for analysis, sampled points would not be independent and thus would not represent the ecological conditions of their respective treatments. 22 RESULTS Landscape classification Sixteen habitat types were identified in the study area from the Coffman et a1. (1980) classification guide (Figure 1.3, Table 1.4). Habitat types characterized by xeric, excessively-drained sandy soils included: Pinus- Vaccinium-Deschampsia (PVD; pine- blueberry—hair grass), Pinus- Vaccinium-Carex (PVC; pine-blueberry-sedge spp.), Quercus—Acer-Epigaea (QAE; oak-maple-trailing arbutus), Acer-Quercus- Vaccinium (AQV; maple-oak-blueberry), and T suga-Maianthemum- Vaccinium (TMV; hemlock-wild lily-of-the-valley-blueberry) habitat types. The PVD and PVC habitat types were dominated by jack pine and/or red pine, and may have contained some white pine as a minor component of the overstory. Sedges, blueberry species, sweet fern, bracken fern, and Wintergreen frequently occurred in the understory of PVD and PVC habitat types. Red oak and red maple dominated later successional stages of QAE habitat types, and white Spruce and pine species were present as minor components. Bracken fem, Wintergreen, blueberry, and grasses frequently occurred in the understory. The AQV habitat type had a similar composition as the QAE habitat type, but balsam fir and white spruce could comprise more of the understory in AQV than in QAE. In the TMV habitat type, hemlock and red maple dominated overstory composition of climax stages. Sugar maple, white spruce, balsam fir, and red oak were minor components of stands in the TMV habitat type. Understory vegetation was dominated by bracken fern. Mesic habitat types characterized by well-drained soils included: T suga- Maianthemum (TM; hemlock-wild lily-of-the-valley), Acer-Tsuga-Drypoteris (ATD; Ina!)le-hemlock-spinulose shield fern), Acer-Viola-Osmorhiza (AVO; maple-violet spp.- 23 sweet cicely), Acer-Osmorhiza-Caulophyllum (AOC; maple-sweet cicely-blue cohosh), and Tsuga-Acer-Mitchella (TAM; hemlock-maple-partridgeberry) habitat types. All of the mesic habitat types supported typical northern hardwood species in late successional stages including: red maple, sugar maple, beech, basswood, white spruce, and balsam fir. The TM habitat type, however, supported more spruce and fir than other hardwood habitat types. The ATD, AVO, and AOC habitat types supported more basswood, black cherry, American elm, and white ash than TM habitat types. Because AVO and AOC habitat types occurred on more loamy soils, the understory of those habitat types tended to be thick. The TAM habitat type occurred on clay soils and was not common in the study area. Wild lily-of-the-valley was a characteristic understory plant of the TM habitat type. Spinulose shield fern was characteristic of the ATD habitat type, and sweet cicely and violet species were characteristic of AVO and AOC habitat types. Habitat types on poorly-drained soils included: T suga-Maianthemum-Coptis (TMC; hemlock-wild lily-of—the-valley-goldthread) , F raxinus-Mentha-Carex (FMC; ash-mint-sedge), Fraxinus-Impatiens (FI; ash-jewelweed) , T suga-Thuja-Mitella (TTM; hemlOCk-cedar-miterwort) , T suga-Thuja-Sphagnum (TTS; hemlock-cedar-sphagnum moss), and Picea-Chamadaphne-Sphagnum (PCS; spruce-leatherleaf-sphagnum moss). The TMC habitat type supported various amounts of balsam fir, white spruce, northern White cedar, black spruce, red maple, and sugar maple. The FMC and F1 habitat types We? e characterized by lowland deciduous species such as white ash, black ash, and red maPle. The TTM, TTS, and PCS habitat types occurred on poorly-drained silty or Organic soils and supported lowland conifers such as tamarack, cedar, black spruce, Spruce, and balsam fir. 24 Most (20%) of the study area was in the TM habitat type (Table 1.4, Figure 1.3). The ATD and AVO habitat types comprised 15% and 9% of the study area, respectively. Other sampled habitat types, AOC, AQV, and QAE each comprised < 5% of the western UP landscape. Landscape-level characteristics and distribution of aspen Most (54.24%) of the aspen on state lands in the study area was 3 30 years old (Figure 1.4). Approximately 15% of aspen was in the 30-year age class and 7.6% was in the 40-year age class. Approximately 23% of all aspen was considered “old” (i.e., 2 50 years old; Figure 1.4). 25 Figure 1-3- Spatial distribution of habitat types in a 12,300 km2 area in the western Upper Peninsula, Michigan as determined from literature and field data collected between 2004—2006. The study area consisted of Baraga, Dickinson, Iron, and Marquette counties. Dark lines indicate boundaries of state forest land. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer—Quercus-Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = ACer-Viola-Osmorhiza; FI = F raxinus-Impatiens; FMC = F raxinus-Mentha-Carex; PCS = Picea-P' J . L "r L o , PVC = Pinus-Vaccinium-Carex; PVD = Pinus- Vacc1nium-DeSChampsia; QAE = Quercus-Acer-Epigaea; TAM = Tsuga-Acer-Mitchella; M , = TSuga-Maianthemum; TMC = Tsuga-Maianthemum-Coptis; TMV = Tsuga- alan‘hemw-Vaccinium; TTM = Tsuga-Thuja-Mitella; TTS = Tsuga-Thuja-Sphagnum. 26 code ondofig 38,—. SE 3.2% amiss see as? Ed 58 mos s...smssamssfissssefissa 3.: 3.38 wt assessmesfisws: game e: Seen :5 Beefesismsé 8.2 3.32 use ESPssssssssasmé 8825 SN EMS. USE secesssassee e6 EN one? E esaasaswssaei .53 one»: one 8% Efi S..S§-§§m5 ea 2a «use. 00.... seasoneosafssaoisv. $2-25 o? 8.2: o>< safesaosestscw .53 E: mama: o5. Sequswéraw seesaw 8.2 8.2% 2e sssss§§sefi 2% as: m: New: EH gasses-ssssses2sm5 m3. New >9V assesseaessmtaw ta Seem mad Sweetseaessm a2 So: oi secessessaei a3 8.3 a: unassfisoésseas-ssa. 28 Ex Ge Essa mace 82 8355;... SE a s agave 25 seem: sow game: 8:58 0:852 Ea .55 SEES .ewmamm mo eoaflmeoo mos 33m 2: .eoomlvoom 5233 38230 3% 20c 98 2383.: Bob aocmoazob 25 5.53 Hanan 89% BEE—Bee we Swag—z £33an code: Bonus, 2: E «be «as. scam" a E 390 552 he eoeaomfioo venomous 42 uses 27 3.88 60-69 yrs 3.11 7.60 i Figure 1-4. Percent of aspen within different age classes on state land in the western gpper Peninsula, Michigan based on Operations Inventory data from the Michigan epartmmt of Natural Resources (2006). 28 Approximately half (50.35%) of the asan on state lands occurred on mesic soils on northern hardwood habitat types (i.e., TM, ATD, AVO, AOC) (Table 1.4). Of the northern hardwood habitat types, most (25.45%) aspen occurred in the AVO type, followed by AOC (10.27%) and TM (9.56%). Although the ATD habitat type comprised nearly 15% of the study area (Table 1.4), only 5.07% of aspen occurred in it. Nearly 25% of aspen occurred in habitat types characterized by xeric soils. The QAE and AQV habitat types each supported 11.88% and 11.26% of aspen, respectively. The remaining 25% of aspen occurred on hydric soils with TTS and TMC habitat types supporting the majority of aspen (Table 1.4). Relatively little aspen occurred in xeric PVD, PVC, and TMC habitat types, mesic TAM, or hydric TTM and PCS habitat types (Table 1.4). The proportion of aspen within each habitat type was not the same as the proportion of each habitat type in the study area or as the proportion of state land within each habitat type (Figure 1.5). The proportion of aspen in QAE, AQV, AOC, and TMC habitat types. for instance, was higher than the relative amount of each of those habitat types in the study area. The pr0portion of aspen in TM, ATD, F MC, and TTS habitat types was less than the relative amount of those habitat types in the study area. The higheSt pr0portion (27.37%) of state land occurred in the poorly-drained TTS habitat type (Figure 1.5), which only supported 11.39% of the aspen (Table 1.5). A greater Proportion of aspen occurred in the AVO, QAE, and AQV habitat types than the relative amount of those habitat types on state land. For example, 16.45%, 8.52%, and 4.68% of state land occurred in AVO, QAE, and AQV habitat types, respectively; however, approximately 25% of the aspen occurred in the AVO habitat type, 11.88% occurred in QAE” and 11.29% occurred in AQV (Figure 1.5). 29 Of all habitat types, most of the aspen < 50 years old occurred in the AVO habitat type (Table 1.6, Figure 1.6). The QAE and TM habitat types contained the second and third highest proportion of aspen < 50 years old (Table 1.6). The AVO habitat type also contained the highest proportion (3.11%) of aspen 2 70 years old. The AQV and TTS habitat types had the second (2.68%) and third (2.52%) highest amounts of aspen 2 70 years old (Table 1.6). All habitat types on state land except TAM contained some aspen within each age class, but some age classes were not well represented within some habitat types. For instance, the ATD habitat type contained very little (0.04% aspen in the 50- year age class (Table 1.6, Figure 1.6). 30 .Ezewcsaméxzerméwzfi H m: mezmegéxzsrmémah n 2: €338.33 -E:Es£:§§<éwzfi n >29 MmNNQoU-E:Essaeemegémgs u 03:. ”SefioeachSaémEH U SE. ”cameoeseimofemak H 25. Homemmamimovrwaoaszm u mm mastebéfifiooe; Anti n U>m MSeaweeamrszeaceefieebesoi M mum mxsxeb -efiesslzewxexk n USE ”REBQETSEEQ n E ”SEES-0-39: «Sow n O>< meatmaaonEM‘afiisow H QH< ”535923 eschewing. u >O< MSe:§&o~=eb-e~.£ao5m0{soV H 00¢. .3958 032832 one do: £8:an .ewesm mo eowmmmaoo «one 33m 2:. .eoomleoom 5953 380:8 See 22m use 0588: 8on eoEEoooe me 890. “853 Bahamas 55; e9: 88m mo 38th was 682 88% so comma 8080a .AqmeomE 63360; 6&3 anmogv «one bosom mo amoeba macaw :OmEmQEoU .mg Paw?— * cowl. 99m .6 E089". .0 codes .6 Ended ._u mean scan co. Enema .L 8583a... 1. V O o M A w w w cod -l l l 13:31-.- 8s r -l .l- . ill- ll- lit--3 8.0—. d 5 l l l -l, :l l.-- . 8.9 - m. | I I ll Illll | I I1 lll Ill II. II l-iII l; ill-.l lllillllil . ill I ..|l .|. !|. l-l 8.0N - l l- l J- - .3. ll 13.2.8.3 lIIlIIlIIIliIl-I -l l rill l nUo.om 31 £28238 25 “SB“: SSE—on 9 mom: 8mm 33QO E fiche 3 26 box: 33 5me mm noEmmm—o 82m 5 E085 :o%<— 8.2: 8.03.? .88- w m .o 33 $23 .533 v2 8. E mom Eamafiméfianasafiépfl o2 _ $.83. mt Eamaiwésfiémé 235 ”no 233 ES azaiésfiémé an $.82 05 §Sés§§§x§é 8825 m: ~22. a Emuuqstssxui Eu m3 398 2):. Emeoéasmsasaéémzfi $3 03;: cod _5 25 nuggets-k8 mamas Eu 8.2 mm. a? 09.. s§é§§u駧6£8v. $2-5m 3mm N332 o>< SEBficéBgsuw :33 Ba 8.: a Q? Eamobasmaséuw Eng-$3 o3 $.83 2e §§§§§§5 282 one 932 a): 5338”:-szswsaasxémé m2 _ 3.8: >o< §£8§§§®¢8w a: ~ 3.8% m3 gag-38 wéemam o; 8.» _ m oi §8-§.E.§§-§.£ 5o ”Ema Ea £E§§§Q駧§$5 Raw 8% Q; E85 96 8% 8§>2£< A32 .3 s gaseoov 25 :33“: mom 23%: .3988 88:ng mam .3: £85on dwmamm .«o 856:8 “as 22% EB .ooomivoom 5953 880:8 8% 2% Ba 2583.: 89% nowmmfimw mg “3.53 33% 8% REMMMV 203 89: ESE 59:22 .2358; 8%: 8283 2: 5 3% 5.53 828% S meN 93% g 8%.me “283% m ~ 32 8.2: $2 £5. 25 8s «3; 3S 3? m: 309 w; :5 So 86 85 So mod 85 So $.23 583 «2 86 8d :3 25 So :.o 86 8o mom 3: mg mg as 2: ”2 SN 23 and mi 2590 ”no cod 8.0 85 So 3o 25 3o 85 3E 3m :5 MS 25 2.0 Ed 2: a: Mac Us; 5 m: :5 So 5o 86 So one 33 Ed E .36 as a: Rd Ed ”do as NE 52 ad 025 $3 26% 3o 85 85 86 8o 85 cod cod cod :5 Eu 8.2 «2 one 33 O? V: a: 3: So 09.. 3.8 :.m ”3 :3 3% N3. m8 :3 w: 02.. :33 :2 NE 85 3o :6 mg as 92 $5 93.. nae-$5 e? _2 N3 85 23 2: :4 NS 8.0 2b 232 go So So 8d 86 2d 8d 3o 85 EH 8.: ”3 mg NS N; w? EN SN 85 >3. 3.: of one 35 :8 and 3a m2 m2 B5 o; 3o 23 So 3o :6 ad 25 3o oi Ed 3o 86 5o :3 8.0 mg Ed 8o 93 yam 23x 3.: E A $8 a? 01:. $3. amém 2-2 2 v 25 min: zom 23.0.62 mmflo ow< .Ezcwgmméxzfiéwznh n m: 8-2§<$§§ 633.5 M 2.5. ”£3583; -E:Em€§.~§<§m§& u >33. mnNEmD-E=Em§§.8V<éM=Q n USE- msafimfi=53<6§m~ .l. 3:. ”magma: {muwémsfi H ER SmBMEMimuvrwzuxmzm n mm mxmme-E:.E.Bo§ 932i M U>m mE:=M§Qh$§§§m£§bé8§ n mvm mamxubéfirmgézfixui n USE mazmnuufifiuzfixcxk .I. E SnESEmQéBS {mow .I. O>< “utmfichéwzflimuV n 9.5. ”SEEP-EA inaugmzmimuv. u >O< mESEEQEGDéNESEuQi8V u OO< .eoomlgom 50253 38:8 9% 2% 28 A32 .3 s sun—tog 8383: 83 5.535% 25 5.53 3.5% E3 Eggnog 89$ 8% “as“: damage/H Eggnog 5&3 5883 05 5 £53 8% go 88o 5.52 Pa 83% ”mm Efimfifi S 5%wa “28.8% ma 033“ 33 30.00 25.00 T Cumulative percent of area 5.00 - 20.00 4» 15.00 — 10.00 - - 0.00 J " Figure 1.6. ,, _-Ae ears "6707' _ ,960—69; lose-59‘ Sue-49. inso-aei i.20-29!l ‘ immej ‘CI<10 . AVO I—exm. AQV I-Wfi " Habitat types Cumulative proportion of aspen within different age classes and habitat types on State land in the western Upper Peninsula, Michigan. Habitat types were determined from spatial habitat type delineation from literature (Coffman et al. 1980) and field data collected between 2004-2006. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer- \ uercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; FI \ Fraxinus-Impatiens; FMC = F raxinus-Mentha-Carex; PCS = Picea-Chamadaphne- S'phagnum; PVC = Pinus- Vaccinium-Carex; PVD = Pinus- Vaccinium-Deschampsia; QAE = Quercus-Acer-Epigaea; TAM = Tsuga-Acer-Mitchella; TM = Tsuga- Maianthemum; TMC = Tsuga-Maianthemum-Coptis; TMV = Tsuga-Maianthemum- Vaccinium; TTM = Tsuga-Thuja-Mitella; TTS = T suga-Thuja-Sphagnum. 34 Stand-level characteristics of aspen .' Comparison among age classes Ljye vegetation structure The structure of live vegetation differed between aspen age classes (p < 0.10) among all habitat types (Table 1.7). Aspen stands in the 20-year age class had less basal area (p < 0.001), a higher deciduous stem density (p = 0.001), shorter overstory height (p < 0.001), less percent conifer cover (p = 0.071), and greater percent cover of herbaceous plants (p = 0.010) than asan in the 50— or 70-year age classes. Conifer stem density differed only between 20- and 50—year age classes (p = 0.055). Diameter of overstory trees was different among all aspen age classes (p < 0.001), with diameter increasing with age (Table 1.7). Mean stand basal area, diameter of overstory trees, overstory height, shrub stem density, and percent conifer cover increased with age, whereas deciduous stem density aJld herbaceous plant cover decreased with age (Table 1.7). Coniferous stem density, Sl'll'ub height, percent cover of deciduous trees, and percent shrub cover were highest in the 50-year age class. \D ead vegetation Characteristics of dead vegetation (i.e., down woody debris [DWD], snags, and StuInps) also differed between aspen age classes (Table 1.8). Aspen stands in the 50- and 70~year age classes did not differ in the decay of DWD, average volume per piece of DWD, and total volume of DWD, but values for those variables were lower in aspen Within the 20-year age class (p = 0.002, p = 0.015, p = 0.001, respectively). Decay, density, and total volume of DWD were highest in the 50-year age class. Density of 35 DWD did not differ between 20— and 50-year-old aspen stands, but was lowest in the 70- year age class (Table 1.8). Snag decay, density, and diameter did not differ between 50- and 70-year-old stands, but were different from 20-year-old stands (p = 0.020, p < 0.001, p < 0.001, respectively) (Table 1.8). Snag decay and diameter were lowest in the 20-year age class and increased with age. Stands in the 20-year age class had the highest snag density (35 = h 549. 1 9/ha, S.E. = 47.31). Snag height was lowest (2 = 4.30 m, S.E. = 0.44) in 50-year- old age classes, and was different between height of snags in 20- and 70-year age classes (p = 0.034). Density of stumps was highest (i = 106.32/ha, S.E. = 70.43) in 70—year—old aspen Stands (Table 1.8). Stump density did not differ between 20- and 50-year-old stands, or between 50- and 70-year-old stands; but stump density was different between 20- and 70- y efir-old age classes (p = 0.036). Stump decay, diameter, and height was similar between aspen stands within all age classes (Table 1.8). \RS ecies composition Stern density of all overstory trees differed between age classes (p = 0.059) (Table 1 ‘9). Aspen stands within all age classes had similar densities of most observed tree Species; however significant differences in densities of American basswood, balsam fir, paDer birch, quaking aspen, red maple, white pine, and white spruce were evident be“ween age classes. Density of American basswood was highest in 70-year-old stands, but was not different between basswood density in 50-year—old stands. Balsam fir density Was highest in SO—year—old aspen stands and was different from 20- and 70-year-old age 36 classes (p = 0.024). Paper birch, red maple, and white spruce densities generally increased with age, but were not different between 50- and 7 O-year-old stands. Quaking aspen density decreased with age and differed among all age classes (p < 0.001). Density of white pine also differed among all age classes (p = 0.002) (Table 1.9) with 70-year-old stands having the greatest density. Diameter 0f OVCI‘StOl’Y tree SDCClES Diameter of all overstory trees increased with age (p < 0.001) (Table 1.10). However, only diameters of bigtooth aspen, quaking aspen, red maple, sugar maple, and White pine were different among age classes. Several species did not occur frequently enough to run statistical comparisons between age classes. Diameters of bi gtooth and QUaking aspen increased with age and were different among all age classes (bigtooth aspen, p = 0.002; quaking aspen, p < 0.001). Red maple and sugar maple diameters increased with age, but were not different between 50- and 70-year age classes. Diameter of SUgar maple and red maple only differed between 20- and 70-year-old stands. White Dine was not found in any 50-year-old sampled stands, but diameters differed between 20‘ and 70—year age classes. B\asal area of overstory tree species Basal area of all overstory species was greater in older stands, but generally did not differ between 50- and 70-year age classes (p< 0.001) (Table 1.11). Significant differences in basal area of basswood, paper birch, quaking aspen, red maple, white pine, and white spruce occurred among different age classes (p < 0.01). Basal area of 37 American basswood, paper birch, red maple, and white spruce was low (< 4.2 mz/ha), but basal area of each species generally was not different between 50- and 70-year-old stands. White pine was not found in 50-year—old sampled stands and, therefore, only differed between SO-year-old stands and other age classes. Quaking aspen had the highest basal area of all species in all sampled age classes. Quaking aspen basal area was highest (p < 0.001) in the 50-year age class (i = 13.63 mz/ha, S.E = 1.33) and did not differ between 20- and 70-year-old stands (20—year-old, i = 12.98 mZ/ha, S.E = 4.62; 70- year-old, if = 11.18 mz/ha, S.E = 1.70) (Table 1.11). 38 .232 5:880 was _mwomma uemefim cemEQEoo oat—=8 m=_a3-_8_mav_v 8:0 A 3 EobbE 8: 0.8 hos...— oEmm 05 5:5 2:— 05mm 05 :o 8:73... .932 5:280 e5 _owommv oofitg mo max—wen $3-25 m___a3-_§mEv_ 05 55.302330 203 acetone.— m_o>o_ bzfimoenr 83 new 2.: a; :2... 3m Sea a: Ta is seam o :3 a? 33.3 3% :53 Ga .33 Ease gauges .83 SN 3% 3m 8.5m as awmm was 38289 ES 2 .m amt: a; $3. 2: ham .320 £68 Assv uo>oo Hawks was 3.82 8a.: 8.38 $.85 em. as 2.33 2583.. spam 2% e2 52 Re a: 25 mg 95 Emma seam m... 83 9% stem 23 33.3 mg .33 .95 Ease 83 3o amne— mvd s32 . as 3:: Lee sagas— Sod mime £3350 ma. 2. :91va emfiw «flame: 2.938on msosemoofl mmod Nada €5.63. Embo— nwhSNe omdm amuému amémfioam maceomao0 b00330 coed wed amefim 3.; .353 No.m «3.2 #2580 «63 33m 22:: .m.m M .m.m M .m.w M oBmiSr £535 3 R n a on R u a on med—o ow< in / lllllllll N mike :«wEomE .ESmcmcom hogan.» 5233 05 E comma mo momma—o own Enact—u macs 606N125 . mofimcg B: @500 N .~ 2an . . stun 5.5503.» mo 28.5w k l .u \ l - l . . Illl. llllllIJ.l -llllill. .wao: 8:280 :8 .mwflfl 058% 538.350 29:2: m:_83-_8xmavc 2: .o A 3 2.65006 8: 08 .532 2:8... 2: 53> 0:: 2:3 2: :o 828?. A33 8:880 :8 _owflmv 88:? mo mix—8:8 83-28 m:_w>»-_8_m:._v_ 05 5:5 3830.8 203 :8..an m_o>o_ M23305. 40 Mo 8883 own “SEE: macaw 69.3% £me 85% 383 :3 22d 3m and. d? 5.3. 2d $8 Ed 29.: 08d .8 $3 Rd 8.8 3.. mad. Ea 3.25.5 dmdd S.E .332 on: envdm 3.2 .88 LE3 €35 .de 3d d... ddd d... 8d 3. 38.5 383m xdd dmd .dfi 3d .3... Rd .mm.m .55 2%: dddd d2 .33— N2 .32 mod .2.» .350 .82....5 dddd d _ .: .332 88 .dmdm. :5. 33% .353 3.25 odd d _ d .3 dmd .3 d _ d .3 .388 mwmcm add 3... 3.3. a... 2.3 mg Ed. 2.5.80 o§_o> Edd _dd .32. _dd ...dd 8d .Ndd 2.80 25.8 omega dddd 3.8 .33... add .328 8.9... .3.de .330 3.59 mddd 2d .5“ ddd .3 m _ d .8 .388 £520 383 ESQ :96— .m.m M .m.m M .m.m M 0388> £528.. 3 3 3 m8? 0? . / |.|II|||I.III‘ ooomlvoom magi 535:2 633.30% Buns 58mg» 05 a... comma on .D.mv cemwsowox) GNU—V 'Ho WOwHWtUHUNHNQU MO GOWEQEOD .%.~ D~h~flrh Eufimnfiuo QNER 88m no :33 mo momma—o own thfiE E wagooo 35on 8b b 8&96 mo Echo mwmd co; 2: o e o c .880 833 50552 3.53 azuxuzmu ommd mm.m 35c. o a 8.». and *8 v8 805.52 3252238 $53.80 wood 3.: madm S.E mung cow Qua 803:0: fianmhcgo swarm mwmd o _ .m 34” o c o c x0280: Sufism 3533. “SEEK N36 3.: 3.: mm.w mmdN mmd “man Ego xoflm EQEmBmEEw ”Samoa nofio modm wmém wwd NN.NN Eng cedvu 59mm 508me EmbESNE afizmck wavd we. : mcém wim 31V 8.9 3.3 anion Sufism 3E§~3 ,3an vmod coda «gamma 3.2: accdbm 3.? «$43 *5 Ewflmm 38:35 .355 ocmd fl fim fl fiN E; NN.N— NM: afim 9:0 anaconda. azuutmfia 3.55 mvod 36v .123 Veda aNN.Nm o a: *wooBmmmn going :32 .m.m m .m.m M .m.m M A25»: umxzzfiumv 3:52on as n a E All?“ “Ia ow. film u .5 cm. 830% 8; mam—o 03‘ l / ||I\I| @8738 was t 53:92 .m 358% Rehab E0303 0:.“ E USN— vaafim USN mhmxmwo gonna: om§o>< .Q — Bank 41 .A—wwa— aux—oummU USN Ewe—ma ofimfimum :Omthn—EOO Emma—fie m=_m3u_mxma¥v Amy—.6 A a: “cohotmfi «CC 0.5 hofig OEGm Us: r375 0:: DEMm or: :0 m03-N>* Amwa— 5:880 new 3939 8.8:? mo $9.98 33.20 m:_m3-_§ma¥ 05 53> 9333.8 803 corona £32 £532“? Sod 8.8 :33: 3.8 kg»: 32: £3.32 BEESEOQQEZ mmmtmmausxmgc 333% 2g o a 3m ewm 2% :w :23 323 833% 38: ~23 3.2 :33 8.: ES min «3:». *ooeamozg “Sasha “35% 83. 3.2 05.3 o a: 9% am; 3% 333 BSQQNKNEQ “35.52% 32 :.N :.~ o a o c @323 5‘23“?th LNUV :3 2.3 3:: 3mm 3% $2 2% mamaswsm 3.005th “.3th 9&0 mom :6 o a _2 3a EEBM Ezkfla.‘ quV 83 No.2 .322 3% ammwfi 3.2 «2.9 3&8 Em mmEQBEmh §3§K 83 3% 98.2: 3.? =2.an $3 «S.E. 2&3 mafia Embhaam 55mm coed NS: 39.3 $4: ._:.3 2: .34 £253“; :92 .m.m m .m.m m .m.m m can: ispsa £33on 3 Gmlulafl 3 860% 8: mmflo ow< A608 .3 22¢ 42 333333.383 333$ . o 3.3 o c o c .833 833 505.82 333x mauxmzm wvmd 942 3.3 o c 3.3 vmém xwo 38 805.52 333.33.33.33 333.3339 mnwd N. fl ._ wmfi mwd b; on; NM.» 300338— 3333333333 3M3Q . o S.E o c o e #0280: 883m 33.33.83 .33sz _Nod mu; 23.x Nod mmfi $6 36 E30 xoflm S333§3=3§w w3~3m3n~ mood omd Luann wag shNfi— m M A «was .233 :88me 33.38333 .3336K m _ m6 wmd 8.3 mud 3.2 mm.“ 3.2 [Hm—mom Sam—3m 3839333 $an Smd $3.— a”: £6 ma.» 3m; mad 5 533m 33332353 .353 30¢ 35 mfi: w fl .3 and" OWN mmé 8% gotog 33392353 33m mmwd him 223 mm; 3.3 o c 38333 going :25. .m.m m .m.m m .m.m m 35% u§§§ 2:338; as. n a on 73.". JI: om fling... V 8.. 830% 8:. mmflo ow< .ooomeoom waist 53:22 amamnfiom Saab 5833 2: E ES 837. so 533 no momma—o own Eouowmw E mambsooo 86on out E0383 mo Echo 383333 333 Emma: 3.305 as 58836 owfio>< .2.— Each 43 .232 5:830 98 _mwflfl 2355 53888 29:38 $3348.35: 8:. A 3 East? 8: 8a 332 oEmm 05 2:3 0:: 08mm of co 82.3... Amwa— 52—380 can 69:9 3.83; «o mix—SS $3-28 £33-3me 05 53> noun—=23 22$ cotoae 226— b25395. coed $0.0 omméfi mmvd aaa.- mmvd «3.3 *wofinfioo 860%.? flunmmSEMmmc 53mm Sod o «v.3 m fl .N. 3.: end has :85 323% 333% 38.: mmmd wo.~ 5.: omd MN»: 36 3.: 83% SE? @395. 53E Rod «3 AF: o a 3m .33 .23 323 3:38me: .3szde n _ m.o o flaw o e o 3am in 833 53:383. .SuV wood mo; gnaéfi mod award 92 «mus *oEmE Smsm 3623.? SEE vomd cmdfi maém o c 3; 5.3” 0:3 BM 2355 LSV mad a: :8: SN asfi o? :5 $39: 8m $38352» 3239* oood mfim 936m mo; al.: omd awn..." *5me wcflgd Embbaam 55mm vmmd 2: 5.3 N fl .m Rama o Ed :83 39mm :22 .m.m m .m.m m .m.m m GEE 9&3»th £535 3 flmlnlslow 3 830% 8; mmflo ow< Gaoov .34 2an 44 p.- .v::— :J-..I :au wrap—1: ,uau 1-111.;5. U...— -:..¥..J_.-:u :_ ..&-.F.,....4-:v 1.1.421?! -JLL. .fl~;-aa....a.aau.‘pv 2:75.... ~v.-...u.::-.!. u-.-.. ....vL- ~.........:~ .drw:.~.u.vl\ s ~ .~ .J~.~...~ 332%38 3:35 mam: E : 3.: : : o : .830 853 2.55.82 6.33 mauxmamu mmm: 0:: :~.: : : :o: :mé Mao :2 E0552 $233298.» nazmnbub :3: 0:: :m: N : : Na: 8: ::.: @8250: fiatmhgg away”: mwm: : _ : :~.: : : : : V3280: 533m 32:33. gaagm :Nm: co: 3.: mm: E: o:.: 3.: .0530 v.85 SE§E§=EM .433on 3:: am; 3.: mm: mm: 8.: SA :33 Seawwmm cxmbsaflc: “Snack :2 93 who So 8... Ed 3.... 333 633m 32:53 3.5V :3: av: .:m.~ _w: a:v.v we; «Nb: ta Sufism gautmfiu “.355 gm: 2: 3.: B: 9.: S: 3.: E? 93:22. Sautwfia 32R 3:: mm; as: :5: saw; : a: $8383 53:022. :33 .m.m m .m.m m .m.m m 32% figs: £33on j n a on j u : om 2. u s cm 360% 8:. mmflo ow< :oomlvoom mafia: gwfioaz agmficum 3&3 3833 2.: E :5: 38m so :33 mo momma—o own 23.5:th E wagooo 86on mob @8203 :o £95 @8988 28 «an :83 oweo>< ._ 2 2an 45 .230. 5:880 can _mwo_m_ 05:me coflaanU oat—=5 $3343.35: 8. .o A 3 E9356 .0: 2a .832 2:5. 2.: .23 0:: 2:3 05 co mos—m?” Amwa— 52.230 28 _owo_mv 85:? mo max—mam €3-25 £255-3me 05 53, Bum—:28 825 Eton». m_o>o_ 5:538} 83 Sam 5.3.2 22 :33 SM .32 Bassoomeoomm =< flmzmmtcpxwmza 55mm 22 3o 8... mod 2:. 3o 2:. :23 aozfi 333% 33% 23 one an} 33 .33 85 .3: *BBQWBEB 333:. 35% Sod mg a? o a: o; .84 mi 323 358.2»:8 “SEHEK 33 So :2. o a :3 3... £803; EELSKUUQM kmbvx ES N3 2: ad a: :.o S... 2&5 swam thtmmmk husk $3 86 an... o e :d 3... Ba Ba Ezkflak kmuV wood o: a? w; :33 as “we... *2ng Ba 3.33353» “Sufism wood o: .3: MS :32 N3 .32 *5? $2.30 Eghauu. 35mm 805 S; .23 2: :22 adv .237 £25ng :32 .m.m m .m.m m .m.m m can: ismg £38on 3 g 3 860% 8:. 9.20 %< 288 ._ 2 2an 46 Stand-level characteristics of aspen: Comparison among habitat types within different age classes Live vegetation structure 20-Year age cla_s_s. Among habitat types in the 20-year age class, there were differences in basal area (p = 0.006), conifer stem density (p = 0.018), deciduous stem density (p = 0.003), diameter of overstory trees (p = 0.010), height of overstory (p = 0.014), shrub density (p = 0.068), and percent vertical cover of coniferous (p = 0.006) and deciduous trees (p < 0.001), and percent cover of herbaceous plants (p = 0.001) (Table 1.12). Average basal area was highest in TM habitat types, but was not different from basal area of AOC and QAE habitat types. Basal area was also similar among AQV, ATD, and AVO habitat types. Habitat types occurring on sandy soils (i.e., TM and QAE) had higher conifer stem densities than other habitat types. Similarly, AQV, QAE, and TM habitat types also had higher deciduous stem densities and percent deciduous cover than habitat types occurring on sandy loam or loamy soils. Diameter of overstory trees differed among habitat types in the 20—year age class (p = 0.010), but there was much overlap in confidence intervals between several pairs of habitat types (Table 1.12). For instance, differences in tree diameter were found between groups of habitat types, but not within groups. Groups are ordered from largest to smallest diameter: AOC and QAE; AQV and QAE; ATD, AVO and QAE; ATD, AVO and TM. Similarly, although differences in overstory height were evident among habitat types (p = 0.014), differences within several groups of habitat types were not significant (AOC, ATD, AVO, and QAE; AQV and ATD; ATD and TM) (Table 1.12). 47 Shrub height ranged from 2.00 m (S.E. = 0.020) in the ATD habitat type to 3.00 m (S.E. = 0.74) in the AOC habitat type, but was not significantly different among habitat types (Table 1.12). Shrub stem density was highest (3? = 8495.00/ha, S.E. = 4480.83) in the AOC habitat type, but was not different from shrub densities in AQV, AVO, or TM habitat types. Stands within the AOC habitat type, however, were extremely thick in the understory, and the fact that shrub stem density was not statistically different from other habitat types was likely due to small sample sizes (n = 4). Shrub stem density was lowest in the ATD habitat type (i = 470.00/ha, S.E. = 315.54; Table 1.12). The TM habitat type appeared to have higher percent conifer cover in the overstory ( i = 7.83%, S.E. = 2.48) than other habitat types due to the presence of balsam fir; however, percent conifer cover was not statistically different from that of stands in ATD and AQV habitat types. All habitat types had similar percent herbaceous cover (74—100%) with the exception of the QAE habitat type, which had sparse herbaceous cover ( i = 6.33%, S.E. = 4.33). There were not differences in percent shrub cover among habitat types in the 20-year age class (Table 1.12). 50-vea_r_age class. There were differences in basal area of overstory trees (p = 0.027), conifer stem density (p < 0.001), deciduous stem density (p = 0.018), shrub stem density (p = 0.002), overstory conifer cover (p = 0.012), percent cover of herbaceous plants (p = 0.005), and percent cover in the shrub layer (2—5 m) (p = 0.041; Table 1.13). Basal area of overstory trees in the AVO habitat type (i = 25.04 mz/ha, S.E. = 2.04) was different fi’om that of the TM habitat type (i = 33.84, S.E. = 2.11), but basal area of AVO or TM habitat types were different from that of the ATD habitat type (i = 29.14, S.E. = 2.66). Conifer stem density was highest in the TM habitat type (i = 1172.31, S.E. 48 = 161.43). The TM habitat type also had the highest percent conifer cover (i = 11.00, S.E. = 1.78). Conifer stem density and conifer cover did not differ between ATD and AVO habitat types. Shrub stem density was highest in the AVO habitat type (i = 12,571.11, S.E = 3626.74) and was significantly different from shrub densities in ATD or TM habitat types. Percent cover of herbaceous plants was highest in the AVO habitat type (i = 83.06, S.E. = 6.46) and lowest in the TM habitat type (i = 46.62, S.E. = 8.33). Neither were different from percent herbaceous plants in the ATD habitat type (i = 67.00, S.E. = 13.66). The AVO habitat type also had the highest percent cover within the shrub layer (x = 82.44, S.E. = 7.58), but was not different from that of the TM habitat type (i = 34.08, S.E. = 7.59; Table 1.13). 70-vear age class. All variables except basal area, diameter of overstory trees, and shrub height were significant (p < 0.001) among habitat types within the 70-year class (Table 1.14). Conifer stem density (a: = 1007.50/ha, S.E. = 151.84, p < 0.001) and conifer cover ( x = 26.44%, S.E = 5.04, p < 0.001) was significantly higher in the TM habitat type than in other habitat types. Deciduous stem density was highest in the AVO habitat type (i = 1573.33/ha, S.E. = 127.19) and lowest in the TM habitat type (i = 402.50/ha, S.E. = 42.34). Deciduous stem density ranged from 745—880/ha and was not different among AOC, AQV, and ATD habitat types. Overstory trees in the TM habitat type were shortest (if = 10.11 m, S.E. = 1.68) but were not different from height of overstory trees in the AOC habitat type (if = 10.66 m, S.E. = 2.24). The AVO habitat type supported the tallest overstory (x = 16.56, S.E. = 0.57), but height was not different from the overstory of AQV or ATD habitat types (Table 1.14). 49 Shrub stem density was highest in the AOC habitat type (i = 35,850/ha, S.E. = 12,545), but was not statistically different from shrub densities in AQV, ATD, or AVO habitat types (Table 1.14). Shrub stem density was lowest in the TM habitat type (SE = 485.00/ha, S.E. = 128.20). Deciduous cover was highest in the AQV habitat type (i = 68.75/ha, S.E. = 2.27) and lowest in the TM habitat type ( x = 39.69/ha, S.E. = 3.19), but several groups of habitat types were not different (i.e., ADC and TM; AOC, AVO, and ATD; AVO, AQV, and ATD). Herbaceous cover was significantly lower in the AOC habitat type (i = 14.01%, S.E. = 3.13) than in other habitat types. Herbaceous cover was similar among other habitat types. The TM habitat type had the lowest percent cover in the shrub layer (x = 12.19%, S.E. = 3.83), although it was not different from percent cover in the ATD habitat type. Other habitat types were similar in the amount of cover within the shrub layer (Table 1.14). 50 A888 5.88 3.2 _8 $3 3 8 6m. 8m 8 3.838 3.8 838 .883 98.88 28.88.. 28.8.. 2.8.88 2.8.88 8.88% 85 3.8 3.8 3.8 888 A88 38 3.8 _88 8.~ 88 m3 8.~ mm.~ 88 38 £88 8% 3.8 3.. 8 8.8 5.8 3.8 5.8 3.8 38.8 .88 .88 .88 3.8.8 28.8 .28 .88 38...: 3.8 8.8 8. 8 8.8 3.8 3.8 3.8 o 8.8 8.8.8 3.8.2 .288 38.8 :3.» .88 LE8 55.55 3.88 3.88 3.88 88 33:8 388 3.8 m 88.8 28.88 .882 28.88 08.88 €8.83 .888 games... 25888 8.3 8 $4.88 83.8 838 838 8.88 3.8 w 8.8 .888 38.88 .88 .88 2.8.8. A28.8 888% 880880 b08830 :28 8.8 3.8 3.8 3. 8 3.8 3.8 88.8 :88. .88 28s 28..» A.88 .83: .8358 8:. 83m 22,2 878 88 $8 3.3 one ~o>< $8 88... a T8 9% 88 So... 28.58 .3253on $20 own 893K 5 8288882 390. “958$ 8358525836388 H ER SmQMEMfimoV -wzoxmzm u m<0 ”333586-388 {mow H O>< ”wtmion-8MEHi8V n QH< ”£35888 -m3oxmzmimow n >O< m53-§a§§b -QNEgeEuQLmuV n UO< .ooomlvoom mat—6 :dwEomE 63.0.8888 5an 58mm? 05 E 85: Qflm co womb “SEE 2858.50 E 320 own 808-3 05 E $888 58.8 «o oéogm nosfiomo». 3: 3818885 038.3, 88 88.8 8883a and mowfio>< .2.“ 2an .232 5:880 8:8 Ewflfl 0:888 car—89:8 28:2: m=_83-_§mavc 8.8 A 3 8588.86 8: 28 $82 863 05 £3, 0:: 2:8 05 no 838?. Amwa— §=3m80 van 3358 35%? mo mix—8:8 83-28 3:83.88me 2: 53, 882328 0.83 82.882 £052 £3388— 888 3.8 588 9888 .88 8888 3.8 .88 8.8 $8... 3.8 8.8 8.8 8.2 a. T8 .82 8...... 68.8 93.8 88.8 88.8 828 88.8 3.8 888 £888 .88 .888 .888 .88. .888 .353 .8835: 3.8 88.8 .88 38.8 68.8 a... 8 3.8 888 .38 28.8 .88 8.8.. .888 .838 .82. .8865 A88 5.. 8 8 A88 .88 8 3.8 888 .88 :88 .8 £88 38.— .8 .88. .888 A888 3.60 Boub> :32 .88 8% fins 33 Qua 8>< 818 88.. 8T8 86.. 2.8 80... .z.§> hum—38088 mmflo own Sufism 5 8853.82 393 .8583 3808 8.. as... 2 5 Gags $3ch A538 5me Sod 3.8: $25.2 «3.2:. 238% seam came 82: :38 $9 83 N3 3." =3 95 2%: seam so: an: 93.8 3.9 SS 2.3 N? 3S @5 Ewe: $98 5.8 $9: 4me m _ 2 8.2 «~42 2.2 95V £255 east 6?: 3 ads: 3.6 w :3 .363 afismm £2.32 *Emeom 3828a 52¢: $.83 633: 3.9 Sod £36.: .853 .253 *2}an geomaou boafigo 2 _d Ed 693 3.9 Red 3.8 .3ch 3.2% Laamé 8:. Ram :32 Eng 25 Qua mo>< fins 3? 0585» b25395 320 own 393% E cowcomoume womb Sufism .Ezfimfizfiugémfim u EH ”cmuwfimimow.§u§:® n m< mmtmion-uM§&imuV u QH< .ooomlwoom waist :wwanz .flsmficom 5&3 8883 05 E 98— Sam so $93 “852 “Gustav 5 $20 own 836m 05 E 2:83 comma mo 2:8an 838%? u>= wENtflofimno oEmtg 8m fiche 335% US» mowfio>< .2; £an 53 .232 5:830 98 _awflfl osmsfim comEano 222:8 £=~343mavc Q: .o A 3 E20b€ 8: 08 5:2 2:3 2: 5:5 0:: 2:8 2: so mos—m?— Amwfl 5:830 98 _owofiv 859; no $92.28 $3.30 £=mBA§mEM on... 5:5 cam—3.8 29$ Eton—2 £05. gingham— §5 :35 SE S.E God .33 Exam £32 :8 TS is 3% 5:8 6:3 63: ~me 83 .38 $5.3 28.5 38% 90835: A83 SE 3.0.3 3.9 .33 E8 3?. S.E 3% 38205 on. c and God 3% mad ‘3: .3; .8." *moetaaou AR; .358 3093/ .35 Eng WE 3T5 mo>< fins 8,? 2§3> 52530.5 330 own 893% 5 333.0%“ 893 “8553 38V .2; 2e? 4 5 82.3 23.5 338: 628$ 83am: $9 8.3 .33. £5.33 2?.de 23%: 2.8.8.3 .2382. 32m 5.8 awe at: 3:: ewe 3.9 33 22 2..” ma." 8.” =3 25 2%: £55 30.: fine awe Gag id 5% o8... .3: .83 .fifi :52 :82: 2.5 .53. G... s an : sod a. . c c 3 8.9 ass 32 .2: 3.2 F: 2.3 3.3 6.08.5 9&3 833 25.8: 6.3: aw: : 3% 255 25.8.. .mmfifl .85: .852 .33. ..£.&... 382.5 $33 32: 82¢: 93: 5.3 £9 83 .823 .82 .533 .33 .225 .32... .8538 bowflgo 2.6 $5 sod 23V 83¢ 5.3 8.3 2...... 3...... mg" 3.9.. 3.3 35.5 .2. 33m :22 675 EH Qua RS... Anna 8,2 Qua 50... ans 80.. .325, >533on awn—o 0% 393K 5 3808a»: 890. 5553 .Ezsmfitumfig -QMEH u 3:. €u§38w®6~¢§ i8... n O>< ”wtfimonéma.“ .33.. H OH... ”85:33; -mzuxmzmimov. n >O< ”$3333.30 -ufifeEmQLmuV n 09.. .coomlvoom waist sawfiog .2355; “can: E883 05 5 was Bfim no 3%» ”SEE EobbE 5 $20 09. a??? 05 5 $53 5%.. mo vacuum cosfiowg o>= wENtBoEEo 0538., 8% $8.5 caugm 98 8383.. .v: 033. 55 :wwi 5:830 can Bwflfl osmrsm coming—=8 oat—=8 m___m3-_ameMv 8— .o A 3 EchotE “o: Pa 3:2 083 of 5:5 0:: 2:3 2: co 82.3... A33 5:880 98 _owommv 85%? mo mum—mam 33-28 m___m3-_8_m:.0_ 05 53> 38328 203 corona. £32 bzfimnokr $3 938 :33 $2: 4363 ame w 33. .32 8.5.3 2%.? .3me an; a: «:3 SE 95% 8me God :3: Std 5.9 3.6 33 $5.3. 98.? 2.8.5. 5.3.3 .3: mega msooaeom 83 €98 awe $6 $3; 3.3 83 “3.3 2.8% 33% his :33 $8: 98283 3.3 2: so. 3 § 8on 39 83 3.8 .c .3: a: .3: 38: £88 AX; B>oo 3093/ :32 3T5 EH ans S>< ans 5,? 3.1.5 53 Qua Bo< 2me> 3:33on mad—o own 393:. E @8582?“ .690 “SEN: 388 +2 2%; Dead vegetation 20-vear age class. Characteristics of DWD, snags, and stumps also differed among habitat types within the 20-year age class. Specifically, differences were evident in the average volume of each piece of DWD (p = 0.055), snag decay classes (p = 0.020), snag diameter (p = 0.045), stump density (p = 0.054), and stump diameter (p = 0.077) (Table 1.15). Average volume per piece of DWD ranged from 0.01—0.04 m3 and was highest in QAE and TM habitat types. Snag decay was highest (36 = 3.2, S.E. = 0.04) in the QAE habitat type. Other habitat types did not differ from each other in degree of snag decay. Snag diameter was largest in the AOC habitat type (i = 13.19, S.E. = 3.43), which was not surprising because overstory tree diameter was also highest in the AOC habitat type (Table 1.15). Although stump density and diameter differed among habitat types (Table 1.15), confidence intervals of average stump density and diameter overlapped significantly between different pairs of habitat types. Stump density, for instance, did not differ between AOC, AVO, and TM; ATD and AVO; AQV, ATD, AVO, and QAE habitat types. Diameter of stumps differed from other habitat types only in the ATD habitat type (Table 1.15). 50-vear age Chis. Within the 50-year age class, there were no differences among habitat types in variables characterizing dead vegetation except snag density (p = 0.012) and stump diameter (p = 0.045; Table 1.16). Snag density was highest in the ATD habitat type (i = 200.00/ha, S.E. = 35.78) and was not different between AOC (i = 88.89/ha, S.E. — 19.01) and TM (2 = 46.15/ha, S.E. = 19.13) habitat types. Stump diameter was smallest in the ATD habitat type ( i = 14.39 cm, S.E. = 1.39) and was not 57 different between AVO (x = 23.58 cm, S.E. = 2.19) and TM (32 = 32.57 cm, S.E. = 7.03) habitat types (Table 1.16). 70-vear age cla_ss. There were no differences among habitat types in the 70-year age class in variables characterizing dead vegetation except average volume/piece of DWD (p = 0.002), snag height (p = 0.062), and stump density (p = 0.042; Table 1.17). Volume of DWD pieces was smallest in AQV habitat types (35 < 0.01 m3, S.E. < 0.01) and largest in the ATD habitat type (i = 0.18 m3, S.E. = 0.07). Volume/piece was not different between AOC and TM, and ATD and AVO habitat types. Snag height was greatest in the AVO habitat type (i =7.98 m, S.E. = 0.60) and smallest in the ATD habitat type (i = 2.44, S.E. = 1.22). Height was similar among all habitat types except AVO. Stump density ranged from 0 in the AVO habitat type to 182.5 (S.E. = 166.61) in the TM habitat type, but was not different among the following groups: TM, AVO, and AOC; TM and AQV; ATD and AQV; and AQV and AOC (Table 1.17). 58 $08 5.3 68.8 $2: 28.8 698 59 m :5 Se 3... 23 new 5% Se as Ewe: 8&8 cane AS: 3:: A8: 55 S.E Sod .Se seam 2.35 ea: :32 .24: 283 seSE $38 62.3 @583 am. a: fii : 835 3.8 Red 56% 36% 3.25 8.2m 8.2% 238 3% bacon Gas 98.8 § .8 A2,: 2 :8 9m _ .8 3.8 owed .3 93. a? :3 .3 .3 388 mwesm A86 5.3 83¢ a; Sad 5.3 3.9 $3 3:: 82 a; 8.3 2.: at: 35.5 2:22 Eek no.8 :98 :98 :98 8:8 see 3% $3 3.8... as... 2.8... as... as... as... are: eeaa\e§_e> $28 68.8: 95.2 s 88.3: $88 $33 S.E acme $.38 8.2.8 8...? 2:5. 3.2% 8...? Age Q88 3.8 :39 ans 33.8 63.8 5:: 8me $3 an e... on an 3 we been mtnov .383 ESQ :32 fine NE. Que ~56 one 85. ans 8:. 8T5 99.. ans 89.. eseee> 3233on $30 own Bowflom E @8535»: 390 “SEN: .535333336M3Q U SE. meocmEMLmoVéngzm u m< nutmiontomsaHimoV n QH< ”535.83; taxoxmzmimov. u >O< mEzzaxmo?ebteNELoEinmoV u OO< .ooomlvoom wag—v "89:22 63233 Saab 5383 05 E 98— Sfim no womb 53m: 258%? E mmflo own Soméw 05 E mcgm comma :3?» 383m 98 .mwmcm .mtnou €003 gov mo 2305.3 wamNtogomhfio 03mg? ho,“ 88.6 @8986 was mowfio>< .2; 2an 59 :33 5:325 28 _awflfl 0:2:me 822868 2922: was 3-18—3.05 8.6 A 3 fiesta 8: on. .882 2:8 05 :23 0:: 0E3 2: .8 mo=_m>* .Gwa aa=8mao «:8 _owflmv 35%.? mo mix—£8 .8365 m_=a>»-_8_m:.0_ 05 53> noun—:28 203 39898 $32 b=Eano£~ 22.8 22¢ 5.5 5.3 23.0 620 2.me 2.2.2. am... when «.33 8.2. 2.3 3:. 2.3 22.2 22 .$ $29 28.0 3.8 5.5 2: : 2.me ES 2.2:: 22.2.. 2.3.8 .3.» .2:— 2.m_...~ .283 225.5 63: 22: 820 9:. E a... 5 am. _ : 2.me 2.8.0 .22.. .83 2.3.2. 2232 .23.. 2.2.3 2.2... 22.53 22: 898 22.: 298 23.8 ewe 39 $3 3. 2m hm Z N... m... 285 383m .252 275 NEH fins 33 Aqua ~o>< ale 2:... 873 «>0... Aqua 80... 05.2..» 5:33on 3.20 own Sumém E @8238an womb 55mm 3.88 .22 2.3 0 6 898 5.8 G98 5% £3 3m «3. 8.". Es “as: am. a and 5. c $9 9. _ .o 2.: cafi 3.: Es H9:353 2;: :23 5.3 S.E .23 ‘26.. can; .353 Laéi £28 6:: 938 A88 3.8 $3. a... 3 3 $89 mwmcm sot 30.3 3.3 3.8 and «NS 3.3. 8.3 ESE 2:29, 309 1 20.8 :98 :98 3.9 6 ~23 8... 3... 2:. £5 oooa\o§_o> Gm. _ a 638 A838 S.E #3 3mm. ~33 8&3 Eta £25 3 _ .8 A _ as 808 3% E 2 hm a... S” 38a $.53 3003 EEO :32 Eng 9: a TE 8: ans 8,? Q§E> b25305 mam—o own as??? E 38823“ 893 Hfimnmm .Ezfimfitcmcgémas u EH SnESEmQéBS {now u O>< fitmaonéwzfiiwuV H QH< .ooomlvoom waist 5&332 £39.93 Saab 5883 06 E 9:: 88m :0 $93 “952 acouotmu S 820 own Hohém 05 E manna comma 5:33 388% can .mwmcm .253 383 8506 (Ho 0.883% wENuBoEEo ofimtg Sm £83 @5ng can mow—23¢. .2; 033. Exa— §=8mmu Ea _mwflma ocmzflm sewing—=8 23:2: $3348.36: 8:9 A 3 =8th 8: 2a 3:2 2:3 2: 53, 0:: 253. of so 82.9... .933— :m=8mau EB _owflmv 35:? mo max—mam $3-25 £33-3me 2: 55 3823—8 20>» Stone.— m_o>o_ szmnoiw 3.9 £39 3.9 3.9 was $9. $5.. 8.9. 359 £205 A85 §9 am: 3.9 33 .53. .amQ can: .589 555a 5.29 83.9 8:: 3.9 a; 2.3 3.3 8.2. 399 $28 629 3.9 899 3.9 83 a... 3. an 585 mafia—um :92 ans 92H 9T5 mo>< Gué 85. 2fifi> €385 mmflo owm 393% E 338058 .699 5.3mm 389 .02 2,5 2 6 629 809 am. 9 $9 :39 3.9 Sod :3 .33 :33 :8... an} .99 3923 A3. a 5.9 82 9 $39 $59 3.9 83 3.: 3.: 3.: a: 2.2 39 Egan $39 539 839 $39 6&9 3.9 $3 :32 $6.: 8.9. 8.9: 2%: 359 £85 2N9 3&9 an : 2&9 $9 3.9 a _ .o E 3 3 3 3 38m mmmam $3: 3.. 9 659 A39 $39 3.9 on: 5.8 3.3 2.6 3.3 8.: 95%: 2:29, 33 :99 :39 A89 :odvv $9 3.9 3 wood «3.: as... any... Oadv «we... *AmEV oooEBEBo> 6 $39 €39 329 529 A869 3.9 $3 3.5. 5.23 mama" 8.3m 8.9... 3&9 €59 :29 $9 $9 E9 59 3.9 32 3 an 3 9m 9m $89 95% 3503 ESQ :32 SEQ EH ans B>< ATE BS. Qué 53. Qua 89.. 2%?» 5:33on 320 own a??? 05 E @858th 890 «Swan: .Ezfimfirfiagéwzmm U SE. MEESEHQ -395 ASV n O>< HtfichQéMSR 18v. u QH< ”£523.33 -mzuxmzmimuwx u >O< m533930339-afifefinbiwuV n UO< .ooomléoom 396 33:22 .mBmfiaum 8&3 5833 05 E 98— 03% so 393 “833 “:20th E awn—o own a??? 05 5 £283 comma £515 magnum Ea .mwacm 9.5% @003 85% mo 8803.5 wfiutoaofimso 033:3 .8.“ $8.6 @8983 98 mowfio>< 8 fl ._ 033. ma 5:880 can _awflfl BEBE :omtaquo 29:2: m___m3-_8_mavc Q: .o A 3 “:80th Ho: 08 5:2 2:3 2: 53, 2.: 2:3 2: :o mus—H5; Amwa— 5:330 28 63:8 35:? mo $92.25 33-28 £33-3me 2t .23 33.3.8 953 Eton“: 226. 5:538} 3mg . $5 5.3 3.0 «mg $3 3.5 . «man an: 8.3. 98V 33$ 3. c . 55 6mg amt S.E om _ .o 3.: . an: 3.8 3.: A83 538% :98: as age 833 and: 3.3 ~36 3:33 we £233 2.8.? 28.2 :33 base 368 . A39 2 we Gas 3.9 $3 3 . 9m 3 3. 55 383m :32 GT5 :5 Qua S>< fins 52 ans 53‘ aka Bo< 03%; bsfimnoi mmflo own a??? 05 E 325858 890 «83m: 3.58 .5: 2an 4 6 Comparison of species composition among habitat types and age classes The greatest differences among habitat types within a specific age class were evident when assessing stem densities and basal area of overstory species (Tables 1.18— 1.23). Nineteen overstory tree species occurred in vegetation sampling plots across all age classes and habitat types. American basswood was present in 50- and 70-year age classes and only occurred in AVO and AOC habitat types. Stem density (SO-year, p = 0.031; 70-year, p < 0.001; Tables 1.19—1.20) and basal area (SO-year, p = 0.031; 70-year, p < 0.001; Tables 1.22- 23) were higher in the AVO habitat type than in the AOC habitat type. American elm was only present on loamy soils in AOC and AVO habitat types across all age classes (Tables 1.18—1.20). Significant differences in stem density (p = 0.027) and basal area (p = 0.025) only occurred among habitat types in the 20-year age class (Tables 1.18, 1.20). Density of balsam fir was highest in the TM habitat type across all age classes (20-year, i= 426.67/ha, S.E. = 94.24, p = 0.020; 50-year, i= 1067.69/ha, S.E. = 153.62, p < 0.001; 70-year, i: 767.50/ha, S.E. = 134.57, p < 0.001; Tables 1.18—1.20). Not surprisingly, basal area of balsam fir was also highest in the TM habitat type across all age classes (20-year, s: 6.59 mZ/ha, S.E. = 2.92, p = 0.008; 50-year, x: 8.69 mZ/ha, S.E. = 1.38, p = 0.001; 70-year, i: 5.12 mz/ha, S.E. = 0.70, p < 0.001; Tables 1.21— 1.23). Balsam fir density (32: 160.00/ha, S.E. = 61.10) and basal area (i: 2.70 mZ/ha, S.E. = 0.99) was also relatively high in the 20—year age class within the QAE habitat type (Tables 1.18, 1.21). 65 Balsam poplar was not present in AQV or QAE habitat types in any age class. Balsam poplar density and basal area were highest in habitat types characterized by loamy soils (i.e., AOC and AVO) (Tables 1.18—1.20). Differences in balsam poplar densities (p = 0.025; Table 1.18) and basal area (p = 0.022; Table 1.21) were only significant in the 20-year age class within AOC and AVO habitat types because it was only present in those types. Bigtooth aspen densities were different among habitat types in the 20-year (p = 0.54; Table 1.18) and 70-year (p = 0.010; Table 1.20) age classes. Densities appeared to be highest in QAE and AQV habitat types, which occurred in sandy soils. Similarly, bigtooth aspen basal area was highest in 20-year-old QAE stands (p = 0.040; Table 1.21) and 70-year-old AQV stand (p = 0.006; Table 1.23) Black cherry occurred in all habitat types and age classes, but the only significant difference in densities were found between AQV and AOC habitat types in the 70-year age class (p = 0.090; Table 1.20). The only significant difference in black cherry basal area was found between AOC and TM habitat types in the 70-year age class (p = 0.085; Table 1.23). Eastern hemlock was present only in stands 2 70 years within the ATD habitat type. lionwood densities differed among habitat types Within each age class (20-year, p = 0.055; 50-year, p = 0.061; 70-year p < 0.001; Tables 1.18—1.20). In the 20-year age class, ironwood density was highest (i: 66.67/ha, S.E. = 48.07) in the QAE habitat type, but was not different from densities in AOC or ATD habitat types (Table 1.18). Ironwood basal area followed the same pattern in the 20-year age class (Table 1.18). In the 50-iand 70-year age classes, the AVO habitat type supported the highest ironwood 66 densities and basal area, but was not different from AQV or ATD habitat types (Tables 1.19, 1.20, 1.22—1.23). Northern red oak was only present in AQV and QAE habitat types. The QAE habitat type had higher northern red oak densities (i: 40.00/ha, S.E. = 23.09, p = 0.012) and basal area (“)Z= 13.87 mZ/ha, S.E. = 9.06, p = 0.006) than the AQV habitat type in the 20-year age class (Tables 1.18, 1.21). Northern red oak density and basal area was not significantly different among habitat types within the 50- or 70-year age classes (Table 1.19—1.20, 1.22—1.23). Paper birch was present in all habitat types except QAE. In the 50-year age class, the TM habitat type had the lowest paper birch density (i: 9.23/ha, S.E. = 4.87, p = 0.079) among habitat types (Table 1.19). Birch densities were not different between 50- year-old stands within ATD (X: 160.00/ha, S.E. = 88.54) and AVO (i: 71.11/ha, S.E = 38.93) habitat types (Table 1.19). Birch density was not different among habitat types in the 20- and 70-year age classes. Basal area was not different among habitat types within any specific age class. Quaking aspen densities were significantly different among habitat types within all age classes (20-year, p = 0.056; 50-year, p = 0.004; 70-year p = 0.009; Tables 1.18— 1.20). Quaking aspen basal area differed among habitat types in the 20- and 70-year age classes (20-year, p = 0.008; 70-year, p = 0.062), but not in the 50-year age class (Tables 121—1.23). In the 20-year age class, quaking aspen densities were similar among all habitat types except QAE. The QAE habitat type supported the lowest density (8 = 26.67/ha, S. E. = 26.67) and basal area (i: 0.36 m2/ha, S. E. = 0.36). In the 50-year age class, quaking aspen densities were similar between ATD (i: 552.00/ha, S.E. = 106.88) 67 and TM (2: 507.69/ha, S.E. = 81.57) habitat types and lowest in the AVO habitat type (i= 244.44/ha, S.E. = 38.93; Table 1.19). In the 70-year age class, quaking aspen density and basal area were highest in the AOC habitat type (density, x = 415.00/ha, S.E. = 91.63; basal area = 17.64, S.E. = 4.79; Tables 1.20, 1.23). Densities and basal area were similar between 70-year-old stands within the AQV and ATD habitat types, and among AQV, AVO, and TM habitat types (Tables 1.20, 1.23). All habitat types except QAE supported red maple. Red maple densities were different among habitat types in 20-year (p = 0.045) and 70—year (p = 0.001) age classes, but were not different among habitat types in the 50-year age class (Table 1.19). Basal area was also different among habitat types in 20-year (p = 0.035) and 70-year (p = 0.007) age classes. Densities appeared to be highest in the AQV habitat type (20-year, i: 204.00/ha, S.E. = 117.67; 70-year, §= 410.00, S.E. = 84.09), although densities in the 20-year age class were not different from AOC and TM habitat types (Table 1.18), and densities in the 70-year age class were not different from those in ATD and TM habitat types (Table 1.20). Red maple basal area was similar among all habitat types in the 20-year age class except QAE, and appeared to be highest in 70-year-old stands within AQV and ATD habitat types (Table 121—1.23). Sugar maple was present in all habitat types, but was not present in 20-year-old AVO stands, 50-year-old TM stands, or 70-year-old AQV stands (Tables 1.18—1 .20). In the 20-year age class, the QAE habitat type had the highest (p = 0.002) sugar maple density (8: 226.67/ha, S.E. = 167.07) followed by the ATD habitat type (i: 50.00/ha, S.E. = 37.86). Basal area was also highest in the QAE habitat type ( i= 1.66 m2/ha, S.E. = 1.11) followed by ATD (i= 0.21 m2/ha, S.E = 0.18; Table 1.21). In the 50-year age 68 class, all habitat types supported significantly different densities (p = 0.002) and basal area (p = 0.002) of sugar maple. The ATD habitat type had the highest density (8: 200.00/ha, S.E. = 108.07) and basal area (§= 3.93 mZ/ha, S.E. = 2.14) followed by the AVO habitat type (density, i: 49.89/ha, S.E. = 32.10; basal area, i: 1.07 mz/ha, S.E. = 1.03; Tables 1.19, 1.22). In the 70-year age class, sugar maple density was highest (p < 0.001) in ATD (i: 493.33/ha, S.E. = 53.33) and AVO (R: 320.00, S.E. = 128.58) habitat types (Table 1.20). The AOC habitat type had the second highest density (i = 100.00, S.E. = 64.59), and TM (32: 37.50, S.E. = 27.20) and AQV (i= 0) habitat types had the lowest (Table 1.20). Basal area of sugar maple exhibited similar trends within the 70-year age class (Table 1.23). White ash was only present in 70—year old stands within the AVO habitat type (p = 0.020; Table 1.20). Basal area was low (E: 0.07m2/ha, S.E. = 0.07) and was not different from 0 (Table 1.23). White pine only occurred in dry, sandy habitat types (i.e., AQV, QAE, TM) and densities only differed among habitat types in the 70-year age class (p = 0.045), with the TM habitat type supporting the highest densities (i: 117.50/ha, S.E. = 40.41; Table 1.20). White pine basal area was also highest in the TM habitat type, but did not differ significantly from other habitat types (Table 1.23). White spruce densities differed among habitat types in all age classes and was highest in the TM habitat type within all age classes (20-year, i: 116.67/ha, S.E. = 61.59; 50-year, 3:: 104.62/ha, S.E. = 26.23; 70-year, i: 117.50/ha, S.E. = 40.41; Table 1.18—23). Similarly, white spruce basal area was also highest in the TM habitat type (20- 69 year, i: 3.07m2/ha, S.E. = 1.75; SO-year, x: 2.42 mz/ha, S.E. = 0.87; 70-year, i: 2.92 mZ/ha, S.E. = 1.22; Tables 121-123). 70 333333.533 3.3.x.“ coo; o o o c o o o o o c o o 838 8E3 E0552 3.55 “inseam N56 o a: oofim :86? o a: o a: 8.2 «3&— ood a: .330 to.— 505.52 3:33:33 332.3530 m mod m m. m «and no.3 £593 0 a: oodm .13.: o a: o _. $3.3 $503.5: 3.2333323.» 3935 o8.— o c o c o c o a o e o c x0035: =5“QO 32.33am“. .SSEE vomd 3.x 5.3 o c o e 2.33 3.3. SNN 8.3. 36m 8.9m atop—0 xoflm 333333.3333LM 3~3Q3K 336 .8. :1 «2:5 3.3 :36”: 8.9 «2...: oodm «8.3 3.th €8.me code «8.8 $533 cwooawa 333553.333 “Snack mod o a: o a... SS ,8...» c a... o as: 2: .8.8_ 1.838 833m 335333 8an omod 3.3 93.33 2.3 2.8.8— 3.3. 2.8.3. mmd. «2.63 moém accdm oodm aaocém "LG Sam—mm 333.2353 .355 mmod o a: o an: 3.: «3.3 c an: o an: 8.2 «2...: *EE quote—Ex 32392383 3.55 coo; o o o c o c o c o a o c wooBmmaa smotoE< __o>o_ .m.m M .m.m M .m.m M .m.m M .m.m M .m.m M @833 £33339 Eisea QIWEINE Gus 8.4. a A J”: .181>.I<. 2 n3 8: 8|_ n.=....3I~.>..d<... all: mluo. <. agape... 8:. med—o own SoMéN 05 E 38:323.. 890 “SEN“: .Eafimit3w3g3msufi n 3:. M333MENL33V -wzuxuzmu n m<0 ”33.895.033.33 -xmuV n O>< HESQGEQ3M3Q {wow u AER ”53.333333 -maexmzmimuw H >O< m53-§n~3~330 -3NESEnQi8V u UO< .ooomlvoom waist 53:32 .33an& 5&3 5833 05 5 “ES 336 :o 393 3333: aqueobmu E mmflo own oném 05 E 355 comma 32:3 86on out M38896 com macho @3986 98 £58on mo 598:: ow80>< .w ~ A 033. 71 ma 5:330 can _0w0_m_ 05:80 38:09:00 032:8 m=_03-_§mEMV Ao_.o A 8 E000t=0 8: 03 00:0. 0E0m 05 £3» 0:: 0E8 05 no 0020?. .Qwfl 30:03.00 can _0w0_mv 0030:? .«o max—0:0 33.28 050340me 05 5:5 “00830—00 0003 “00:82 m_0>0_ homage} hooo 53.3: sand—e— wodo— anflnm: 336: 0863. oon 086?. end—N naooowv— end—N «8.3.2 #0033280 860% =< m~§0~=3¢m0=3 .m ow_.o o o o o oo.o3 oo.ov o o o o o_ oo.o— :83 Bo:0> 3033M 300i ovo o am _o 03.3. o go o go o go o no o aao ’00:.50 0:55 «SQOhu. gfimm omoo mm.o_ mn.m— boom boom o o o o 39w oo.n— oo.o o 0:3 0:55 QSBUNKNEG §=m§kk 83 o c o c o a o c c a o 0 5.0 853 533.3330. 0033. moo o o no no Nb— 0$6NN o ago on hm uncoom oo 2 A_aoo.n— 8.2 «3.3 *2me hum—0m 30.33.03 SEE ~25 o o o a o c o a mom 8.» 3o 0 2:0 8m Saxon; x03. 25 o mm a 2.08.3. o no 8 2 88.2 o a: S.E £8.03 8.8 £a85£ .2908 00m 0.0333803 “SEER omoo 3N2 ammdoo Sow abodu 3o. _ m goodnm mmév aooomm Simo— aood: 8%: 08.9mm *coama main—00 3K0KNKAQ3Q £50m Sumo mm.m and o o o o o o o o o o :85 00%.”— __0>0_ .m.m M .m.m M .m.m M .m.m M .m.m M Mom M 333: 3%3033 £58on dmloloua NEH jug mm? duvuwawo>g< fine a: 860% 8; mmflo oww an??? 5 3332mm: womb H833; .Ezsmfinficgémak H 2% SnESEnQéBS imuV u O>< nutmionéMSWKmuw n QH< .ooomlgom wfibfi :meosz Eggnog 5&3 5383 05 E 98— 03% so womb “83¢: :58me E 330 own 336m 05 E 3.8% comma 55$» 86on cob 5.8633 8m E95 335% Ea mEBm mo “38:: omfio>< d H ._ 2an 73 .232 5:880 can _mwflfl ozmzflm savages 032:5 m=_m3-_§mavc 8 _ .o A 3 828.33 8: 2m .83— 088 2: .23 on: 089. 05 :o mo=_m>... Ammo— 5:830 28 _owflmv 35:? no mix—£8 €3-28 £33-.“me 05 53> wean—=23 22$ v8.53. 203. 523329 vmoo 3 .vf £3.33 mndo— «aha;— — om.ww_ aaooéomu *oofinEoo 86on =< ammtfiaaom‘mza 35mm m o mo o o mo.m 3... oodm ooém :23 26:0? 835% 39$ ooo.o mmom anode NSWN somwm owd a3.3 .325? 833 3595. SEE 83 o c o a o a 8a 323 gautmfic 3:3de ooo.. o o o o o o :3 BE? $53393. .Suv‘ Nooo o go o _ .mm .3de nowoo aoooou *oEmE .awsm an??? $35 ooo._ o o o o o o 3E tom :5on xmuV ovvo vwom :6: m _ .wo S.EN www— _ oo.ww~ 292: tom 333:3?» “Samoa wood 3.; «oofiom mowm :33: 3.03 aoofimm *commm mew—«=0 Embhaum 35mm ohoo 3% :35 h _ .mm a:.: 3.3 woode— 283 59E :32 .m.m M .m.m M .m.m m €28: umbEmBmo £32on 3 a T5 8: g 369% 8: mmflo own “Show E @8535“: 393 “855 ER: 5: 29¢ 74 3333380 9:35 mwflo own 893:. E 323898 8% 53mm 033 omN cmd o c o c o c o c 380 BEE :85qu ESE magnum mid o c o c o a N2: 36— o c Mao Bu 805.52 3:32:95.» “£23.89 oood o a: m ~ .8 :35: mm. m _ ammdfi o 2...: cow .36 *nooaco: Enamhatg QMEH omod o c o e oodv 3.3 o c o c zoo—Eon Eowmmm 3:238. SEER oaod om.m cm.~ oodv 8.3 o c cod 3% cmwv 3.3. 90:0 xoflm Saunmhfitaxm “338% 0 5d omém ficmfim o go o a: Ramp 92.63 8.2 «3.2 *5me floowwmm Embfiaflaa ”Snack g _ .o oaém cmfim o c o a o e mmdm 3.3 Hmaom Sam—mm 3533a 8.3V oood $.32 camfifi. o a: oodom 2.563 8.9 93.3 modm «3.3 *5 Sam—mm anautmfic Miss :33 o c o e o c o e 093 3.3 86 525:2 azgtmfiu SEN coed o a: 3.9: 93.3w o .1: o 2:: mmdv .cc.mm *600383 sucrose :32 new m .m.m m .m.m m .m.m m .m.m m €25: ismsmv £585 3 334% 3 fig 3% 360% 8: .E3Em£~EGNBV< -QMEN H SE. SnESEmQéBS .SQV .I. O>< HESRQEQ$M=Q {mow H QH< ”£3583; -wzuxmzuimuwx H >O< ”53:933ch -QNESEinmuV u UO< .coomlvoom wcgv 53522 $32.83 Saab 5833 08 5 2:: 05% so 890 53a: Eobbfi 5 $20 own 89?? 05 5 £5de comma £53» 88on ooh b880>o 8.“ 32.8 @8983 98 mEBm mo Bags: ow80>< .34 2an 75 .232 5:380 nan _mwflfl ozmufim :omtamEou 29:38 m=_a>»-_8_mavc 2: .o A & E80Eu 8: 8a .832 2:8 05 5:5 on: 088 8: :o 82.3... .9323 5:330 98 _owamv 35:? mo $92.28 33-88 m___a3-_8_mav_ 05 53, noun—=28 88> 3:38 m_o>o_ .3338} Sod omdi £8.33 No.2— ..Séwm— mm.m_m €36:— nmd: «3.3» wwdfl .3.ch *vofinfioo 86332 8.4533883 33mm 83 o c o c o c o 8 o a :23 323 835% 38.5 28.8 3.8 .33: 2.2 $8.2 o .8 8.8 .88 8.2 8.2 802% 333 8538 832.5 8.8 8.8 ‘88: o a... o e... m? :88 o .8 .28 as? 3:63.553 hztmkdkk 88.8 o a: 8.8 .88 o a... o a... o a... £3 383 5852556“. LMUV. 88.8 8.8 :88 8.82 8.8m 8.8 98.3. o a: 8.8 .888 .288 ~88 56:33 35$ 82 Ra 8.8 o c o 8 o 8 o 8 28 8m Sakeak waV :88 8.8 98.8 8.8 28.8 8.8 £8.88 8.8 2.8.8.. NS 8.2 .285 8m hmfimgagmb HENRRQK 88.8 3.8 98.3: 8.8 98.8 8.: 3:2 8.: 28...: 8. a «8.3. 28% 8880 935.8ch 558% 83 8.8 8.8 o c o 8 8.: 862 8.2 8.2 88 “88 :82 .m.m m .m.m m .m.m m .m.m m .m.m m @282 £888 88298 3 g g g g 860% 88 momma? own 20-8052. E “885858 390 “9582 gee .8; 22¢ 76 38sz80 3:2? 77 coo; o c o c c c o c o e o : €00 323 505.52 6.35 mauxmau coo: : a: cod :54”— : a: o a: 8.: «3.: o a: .330 to.— 505.52 62533.8... 35%ch moo: 3.: 2.3.: Q .: «3.: : 2.: M: .: Saw—é : o: 8.: .38.: £503an flw=mm8§u swank 83 : : : : : : : : : : : : :55; 688: SEE»: “SEAN 3m: ::.: 5.: : : : : 5.: mm: m _.: 3.: 2.: 2.: bump—o xosm SStmEhzaxm N 8:: 3.: «8.: :0: :82 2.: «:1: 8.: «8.: no: as; :5 3: 20%: :8an cumbficgu: :33on m8: : a: : :8: _E 2.8.— : 3: : E: :3 2.»...— tuao: 833m 333.3: 3.5V :8: mod :mé ::.: EN mo: _.: mmd an." 2.: mm: mm: 5.: .1: Sam—am : a a an a s 353.225: .355 2:: : e: : B: 8.: 2.3.: : Ba: : E: 8.: 3.8.: .63 525:5. status: BEN 89. : : : : : : : : : : : : 9:03:83 5312:" :32 mm m new m .m.: m .m.m m .m: m new m 3:8: @332: mum—Banana GTE NSC. 5H5 Nm< QH=HNDH< STE N>O< QUE NUO< momooam 00C. 320 um: Seaém E 32.8892 890 «55:»: .EzfimfizficEéMfim u SE. mamcwEMimo Vézukmzm n m<0 SuESEwQéBS 383... H O>< WESQGAxQ$M-§i8V H AER ”£33.86: -aauxmzmiwow u >O< mEaszmfizabéfifofiwbin.»V u UO< .ooomlvoom mats: sawfloaz .Szmficom 5&3 8833 05 E :5: 88: co 890. “85:: Bogota: E awn—o own 323K 08 E 353m 5%: 55:3 :20on nob @8296 8.: Echo @3988 EH: no.3 Emma omfio>< AN; 2an .wai 5:380 us: _mwomma 0538: 5:59:00 2:22: m___:3-_:xm2vc A2 .: A 3 Bogota 5: 2: Sta. 2:8 2: 5:5 0:: 08mm on: so mos—.9... A32 5:830 can .035: 35:? mo nabs—8 $3-28 £33-3me 05 53> wen—:28 22> :8:an £32 b25389 8:: :5 .3: a: .85 8.: :3: 8.: :8: N3 :82 :3 .8:— 88558 330% __< flmtfigmuzc .m a _ .: : : : : 8.: 8.: : : : : 8.: 8.: :23 32.; 833M 383 :8: m5 :8: : :.: : a: : ::: : :: : a: .829: 823 «3&th 32mm :8: :2 8.” 8.: m3 : : : : 2.: 8.: : : 2:: 223 Btcbtmsa gfiuekm m: .: : : : : : : : : : : : : :8 32.3 Egbuuu. .3va 8:: : :: :2 :8: : a: w _ .: .3: 8.: £8: 8.: :8: :23: 38: 362.23 33% m8: : : : : : : : : :m: 8.: : : 2:: Bx EEQE LNUV :8: 8.: :8: : :: 8.: .8: : .: 8.: .8: 8.: :8: .292: :3: hmfimcfiusmh .K 8:: 2.2 :38 :m: 8.: 8: :23 8.: 2.8.: m2 :8... 8::. £88 23:: 8:80 metamum 55mm 2:: _:.:v _:.:v : : : : : : : : : : :23 3%: .32.... .m: m .m: m .m: m .m: m .m: w: .m: m :53: £388 3588: 3 Emma: and an: 85:. 6 T5 9% fig 860% :5 Mo|><| 82: um: 893K E 883.8%: :25 53:3 3:8: .8: 2:3. 78 waszmfiauc 3&3»: :o:._ : : : : : : 3:8 833 805.52 932.: azuxmam :::.~ : : : : : : Mao :2 805.82 3:32:38 $5539 3:: : :: mm: .3: : an: *cooBao: flatmhczau 3M5“: :::._ : : : : : : zoo—Eu: Eommm 3:8me Mitzi :8: mo: 0:: o: :m.— N:.: n:.: abuse :85 Eazwfifinsw ninmok in: 00.: :4 2.: 3.: mm: mm: 5%: 508me 3:55:33: 333% gm: 8.: m:.: m _.: 2.: : : Bio: 883m 353.3: 8.5:. 5:: mm; :3.” :0: .34 mm; .3.“ *2.“ Sam—mm 3:32:53 3:85 2 fl: : : 3 .: mu: : : EB 33:58:. 3:82:53 3:: _m:.: : :: 91 «ad : :a: *uooammmn 50:86:... :32 .m.: m .m: m .m: w. €53. €3.83 £532: 3 a _us 83.. g .28.: 8.: mmflo om: 8%-:m 5 3688mm: womb «95:3 .Ezsmfiafiuxxémah U SE. mqugofiaOéBS (8:. n O>< mutmich$M3§fi8V n 0.3:. .ooomlvoom mats: :meom—z .flsmfidom Saab 5883 2:. E :52 28m :0 889 :3an 68:56 E :33 um: Somém on: 3 mung: 5%: E53» 860% ooh .0880; .8: fiche 9:39: :5: :2: Emma 0333:. .NN._ 2:3. 79 dono— §=8mmu new _amomma ozmsfim acmEanQ 23:2: m_=m>>-_8_mavc 81o A 3 EoBtE 8: Pa .532 06% 05 5:5 0:: 2:3 2: :o mo:_m>_. Ammo— 5:830 can _owo_mv 3.8:? oo mam—«SN 33-28 £33-3me 2: 5:5 nous—=23 0.63 3:82 m_o>o_ b25305— Roo 2d avwdm omd good" 8N saw—.3 *oofipfioo 860mm =< flatmES‘Mmaa 55mm mmmo o o m o .o 36 mod oo.o :23 30:0.» 333% 68.5 ooo.o two amvd who «36 mod amoo *ooamm SE? ”395. ”:25 83 c c c e o 5 Ba 32? czcutmfic wafixuxk ooog o o o o o o :3 BE? 33.8%qu Law mood o Uo mo; :2: EN «mod *oEaE Swam 98$me .325 83 o a o c o a 8a 3m SEQ: Law ono 3: NN... ono 34 mm; bwfi 2&8 3% 8333592 «Snack nmmo omd 3.3 E: hm; fl o o .m 2.5 59% @5130 EmbeSE. Sémm ammo on; a; $4 3% 5: mod :25 89$ :32 .m.m M .m.m m .m.m m €23: umxcafiuwv £33on 3 a T5 mo>< g 860% 8; 820 um“ auxom E @2585“: 89? 55am 388 .8; 03.; 80 3:33:333333 3.33rm ovwo Nmo mm... o o o o o o o o .838 BE? 3.85.52 3.33.3 333.83% moo .o o o o o o o mvo av... o o xwo 38 3.55.52 323.33.33.83 333.353,.3 ooo.o o A.o mmo 9:...— woo «woo o ago So 9.36 3.38330: 3332333333 3M3Q omoo o n_o o ago mm; “MN.“ o A3o o aao Eco—Eon Eogmwm 33.33.38. 3.352% mmoo moo gNo... moo £mo.o o 3o woo unawoo mmo aumvo *Eono gum—m 33333333333M §N3Q3K ooo.o SN ENQN o no o Eo N _ .3 :36 om; 8on4 .2333 6.03me 3355333 3.538% o_m.o go go o o o o o o 34 FA anion Sufism 335333 33.2? ooo.o ono unfim o 3o 3m.o oaoeo moo A_Noo moo at: *5 Emmfim 3333,2233 .4355 :uvo o o o o o o o o mmo Nmo 3:0 Swerve/w 3:33.2353 3.55 ooo.o o oo 36 avodm o fio o ago Nmo «on... *300333 swore: =32 .m.m m .m.m m .m.m m .m.m m new m 35% “$23.33 £525 3 3% 3% g 384 .203. 8: mmflo own Eamon E 325852 393 “8333 33334333333 u 2. ”3333633.: 43‘ n 92 ”Emaebqggéuw n a? ”55583: -3333333 u >o< E m555335-33E28333 n oo< 383.com mass 53322 .2358; 5&2 some; 05 a 22 2% 8 3%. 3352 2&6 5 £20 093 “8%.? 05 5 $33 598 £55 360% 83 @333 So £23 365% 98 85 3.2 $802 .mmg 053. 81 .232 5:380 333 _meE ucmtfim 53868 23:28 m___m>>-_mxmavc 8_.o A 3 Echotzu 8: 23 3:2 083 2: 5:5 0:: 2:3 2: so 823?. .933— 5:280 3:3 _owflmv 35333 no mix—SE 333-05 3:33-3me 2: 55 3823.8 203 38.5%.. m_o>o_ b=338m F mmflo own 893:. 5 3338852 880 33532 oovd 3N ”.mm Nwd $.33 mod cadm _N.w 3.3m ové Exam 303588 860% =< 3233.5393623 333m ovwd E .o 3.: o c o c o c o c :85 30:0.» 3333M 333.2% Sod mm; aNQN cod €36 mm; fiNMA cod «wad m md «mm... *uoEmm 823 3333.5. “.32: wmod 3d a-.~ o fic o as: mm; a3: o a: .353 823 3333.22.23 333.382K omod o a: 36 «3.: o as: o ewe o as: .23 BE? 2323:3333 233V ooo.o Sac :36 ca; «5.3 mm." gang: 0 a: mN.N «33 $338 Bwsm 3.8582 332.2% 82 86 a... o c o a o c o a Ea 8m 53332 233V mood no; «QWN h fl .c «:6 mad as”: mm.m aund mud «2: $338 com 83.83533 §~3§3m mood cod «3.2 ova infia mod awa.N 2 .N :36 0:3 «3.: .7333 @5130 3.532233% 333% 026 m _ ._ mad o c o c w: wad mmd MN... :85 .893 .35 new m .m.m m .m.m m .m.m m .m.m m 929. £3.39 £520: 3 ENE/l... g 3% 38'... scams; 358 a? 2.5 82 Timber volume Timber (pulpwood) volume (m3/ha), as determined from Operations Inventory data was different among habitat types in the 50-year and 70—year age class (p = 0.069 and p < 0.001 for the 50-year and 70-year age classes, respectively; Table 1.24). Volume of pulpwood in stands within the AQV and ATD habitat types were highest in the 50- year age class, but not different from volume in the AVO habitat type (Table 1.24). In the 70—year age class, volume was highest within the ATD and TM habitat types (Table 1.24). Volume of pulpwood in stands within the QAE and AOC habitat types were generally lowest in both age classes. All stands increased volume in 50 to 70 years, but stands in the TM habitat type increased the most. 83 .332 EEBQU ES _awofia aunt—am :omEano 29:2: mEaBAmV—EV 8— .o A 3 inhuman Ho: 8a Ste. 253 05 .23 0:: 2:3 05 :o 333?. .932 3:880 28 _owflmv oo§E> .«o mam—a5 $3-28 £33-3me 05 £3, coon—:28 053 Eton“: 20>“: M23305? as 3.3 one 3.: _2 3.2 a; 3.2 :3 5.3 33 3.: Amuse Sqov NE .35 EN 23.3 o2 33.3 m .353 NS 2.53 a? .3: 355 o§_o> =32 .3 m .m.m m new H .mm m .m.m m .3 m £5305 jug 2H fins 56 $75 92 are 82. gunfire... smua uo< mmfio own 393:. E 330852 390 “853* 34 md— no; 3.3 and wmdm om: no.5 mm; no.3 EA wwfifi AmEoov ooo.o 2+ «3.5% 2:. «Exam owd augfih Ed .19in cow amafiw 56 aka—é 35mg: o820> __o>o_ .m.m M .m.m M .m.m M .m.m M .m.m M .m.m M ©253on own: SE. 8T5 m< Que“ e< Tunic >O< nmvufi UO< mmflo own a??? 5 338298 890 «55m: , .E:Em£§.5~<émak u SE. SmQMEMimo Véaoxwzm H m< matmaokaéMEH -LmuV n QH< ”£55.86; -wauxmzmimuV H >O< ”Ezaémszubéfifofiud ion n UO< awo— .«o xofim 5-2: x m: on $4. a E 8an be 28 68>» 2H3 wows—05 BE ME mod u 28 0:0 .3383 b852— mcoufiomo moon—80M 12:32 .«o Hangman 53:82 2: E 3288 mos—g 03380 05 many—268 .3 wean—suing 33 «6%: E 0832/ .ooomlvoom macaw naming: dBmGEom 5&3 E833 05 E 32 396 no womb 55m: fiesta E momma—o own 393K use .3 05 E 3:36 comma :33? 88on ooh E8330 e8 fiche @8986 was 9258 BE ofimgoov 0829, owmeo>< Jun; 033. 84 DISCUSSION Traditionally, effects of forest management decisions were only considered at the stand-level. Only relatively recently have forest managers begun to examine effects of management within a different spatial (Crow and Gustafson 1997) and temporal (Bissonette and Storch 2007) context. Although harvest or other forest management practices occurs at the stand-level, it is important to recognize that stand-level treatments may have implications at broader spatial scales, such as the removal of unique features of vegetation types or changing the spatial arrangement of vegetation types within a landscape, and thereby potentially restricting their availability to wildlife. First, however, it is critical to have knowledge of the potential variations in vegetation composition and structure due to abiotic spatial characteristics (e. g., soils) and successional changes among habitat types. These results describe the composition, distribution, and temporal changes of the aspen vegetation type within various habitat types and provide a foundation for understanding the distribution of areas that can potentially provide specific characteristics of aspen, how aspen physiognomy may change temporally within different habitat types, and how well aspen management strategies on public land are meeting ecosystem-based objectives. This understanding contributes to improving knowledge upon which sound forest management decisions are based. In the initial study design of this project, stands represented the evaluation unit, but because some stand boundaries overlapped habitat types (Figure 1.2; this occurred in approximately 20'% of my sampled stands), I modified the evaluation unit to be age classes within specific habitat types rather than stands. This modified sampling design probably reduced the variance associated with measured vegetation variables due to 85 clustering of samples. However, significant differences in composition and structure of aspen in different age classes and habitat types were still evident, and the design raised awareness of the significance of maintaining stands within habitat types. It may be difficult, for instance, to implement stand-level management to maintain or enhance habitat conditions for wildlife that require specific stand-level characteristics. For example, a shrub-nesting bird might require 20 ha of interior vegetation. If the stand being managed is 40 ha, 25 ha of which consists of a habitat type with low shrub density and 15 of which is a habitat type that supports high shrub densities, habitat requirements for the bird will not be met with management of this stand. Even though management occurred within one stand, vegetation within different habitat types in that same stand will respond differently to the same treatment. Stand boundaries should be consistent with boundaries of inherent landscape features, or habitat types, so that managers can fully understand and realize the effects of management on resulting within-stand composition and structure, and landscape composition and structure. Results of this analysis have several implications relating to management of forests for a diversity of structure and compositional characteristics in aspen vegetation types to provide habitat for multiple wildlife species, to maintain timber production, and for maintaining Michigan’s certification through the Forest Stewardship Council (F SC) and the Sustainable Forestry Initiative (SFI). Three broad issues were evident from the results regarding habitat provisions for a diversity of wildlife species and timber production. First, general structural and compositional differences were evident between and among aspen stands in different age classes and habitat types. Second, managers may need to increase representation of managed lands in certain underrepresented 86 vegetation types. Third, the distribution of aspen within different habitat types and age classes on state land does not reflect its distribution within habitat types and age classes across the landscape. The fact that some habitat types contained very little aspen in certain age classes also may have implications for wildlife habitat and timber sustainability. The information provided through this research can help Michigan enhance forest management plans to maintain certified forests. Differences in vegetation structural attributes across habitat types Soil moisture and texture, and landforrn characteristics contribute to differences in vegetation physiognomy within habitat types. Natural successional processes change vegetation physiognomy between age classes within the same habitat type. The most distinguishing differences among habitat types were evident when evaluating composition (Figure 1.7). Generally, habitat types supported by xeric soils (i.e., QAE, AQV) supported more bigtooth aspen and oak than other habitat types. Spruce and fir were much more prevalent in the TM habitat type than any of the others. Hardwood species such as basswood and elm were significant overstory components of AVO and AOC habitat types; especially in later successional stages. Hemlock was only observed in the ATD habitat type. These compositional differences were more evident in older (2 50 years) aspen stands because the aspen is beginning successional transitions to other vegetation types. Structural differences were also evident among habitat types (Figures 1.8—1.10). The most notable structural differences were the presence of a thick shrub understory in AVO (and AOC) habitat types, a relatively open understory in the ATD habitat type, 87 higher stem densities of coniferous trees in the TM habitat type, and higher aspen stem densities in the TM habitat type (Figure 1.9). Aspen stem densities declined sharply in old (> 70) age classes with the ATD habitat type, most likely due to competition with maple for overstory dominance. Because forests are managed for multiple uses and ecological benefits, representation of all habitat types and their associated vegetation types within the landscape is important to ensure a range of vegetation conditions exist for wildlife habitat and timber production. It is important for mangers to understand these structural and compositional differences among habitat types when designing management plans to meet multiple objectives because it is contrary to traditional belief regarding relationships between aspen harvest and wildlife habitat suitability. For instance, < 20 years ago, aspen management recommendations to benefit wildlife included harvesting small areas, maintaining stands in various ages of maturity, and shortening harvest rotations to foster favorable wildlife conditions (Kidd and Koelling 198 8). These recommendations would likely have provided habitat for deer, grouse, and wildlife preferring edges, but would not have maintained conditions for edge sensitive species and species favoring old aspen. For example, woodpeckers and other cavity nesting species may use snags provided by old aspen stands (Harnelin 1992). Additionally, traditional recommendations did not distinguish among aspen stands in different habitat types that provide various structural and compositional characteristics, to which wildlife may respond differently. For instance, regenerating aspen in QAE and AQV types may experience greater impacts due to deer browsing because bigtooth aspen, which grows in the xeric sandy soils of QAE and AQV habitat types, is preferred over quaking aspen by deer (Heinen and Sharik 88 1989), and also has lower stand vigor and stocking than quaking aspen at various browsing intensities (Campa et al. 1992). Today, research has advanced knowledge regarding the relationship of different ages of aspen forests to wildlife, and the relationship of aspen in different habitat types to wildlife. More research is needed, however to continue improving our understanding of wildlife responses to physiognomic differences among aspen in different age classes and habitat types. These relationships are further investigated in Chapter 2. 89 ‘ Habitat tmes xeric sandy soib mesic loamy soils Tree species QAE - AQV g LIV! V ATD fiw AOC American basswood - i , 7 American elm Balsam fir Balsam poplar Bigtooth aspen Black cherry Eastern hemlock Ironwood ‘ , ‘ Northern red oak haummma .. fl, . _‘ Paper birch _ _ LN“... h . Quaking aspen ' _ Red maple ' ”i it Red pine (White spruce Yellow birch Figure 1.7. Relative differences in overstory tree species presence and dominance across habitat types sampled on state lands in the western Upper Peninsula, Michigan during 2004-2006. The location of bars represents habitat t ypes in which tree species are present. Darker shades indicate dominance. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola- Osmorhiza; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum. 90 4000 3000 « - («a 2-} 20-year 5 "TM 1 .: age class i 2000 w nATD 1 1 AV 1000 - *9, A0. Coniferous Deciduous Shrub stems/ha sierra/ha sierra/ha 4000 3000..,____ ~~ ~-~ fir 50-year a, age class i 200° “‘ _" "" '“ _"" 1000 « , — in 0 fl . ,,,_, - Coniferous Deciduous Shrub stems/ha stems/ha stem/ha [T TH”? 70-year “ {I 1 age class é E” Am i EAVO; Coniferous Deciduous Shrub stems/ha stems/ha sterrs/ha Figure 1.8. Summary of vegetation structural differences among aspen stands within 3 age classes and habitat types in the western Upper Peninsula, Michigan. Vegetation was sampled in 2004-2006. ATD = Acer- T suga-Dryopteris, AVO = Acer- Viola-Osmorrhiza, and TM = Tsuga-Maianthemum. 91 1400 1200 . 1000 — ‘ [ITM } IATD! LmAvo‘i 600 4* 400 200 w- A Aspen stems/ha 20-year 50-year 70-year Aspen age class Figure 1.9. Summary of differences in quaking aspen stem density among different age classes and habitat types that can support aspen. Vegetation was sampled in 2004—2006. ATD = Acer-Tsuga-Dryopteris, AVO = Acer— Viola-Osmorrhiza, and TM = T suga- Maianthemum. 92 IYMT 20—year u ATD . age class ’ ’ a AVO Basal area % conifer % Height (m) (n'Q/ha) herbaceous .- TM ‘1 50-year n ATD 1 age class 5 AVG; Basal area % conifer % Height (m) (mZIha) herbaceous I TM 70—year . ‘u ATD age class a AVO Basal area % conifer % Height (m) (HQ/ha) herbaceous Figure 1.10. Summary of vegetation structural differences among aspen stands within 3 age classes and habitat types in the western Upper Peninsula, Michigan. Variables presented are important in distinguishing habitat types. As aspen stands age, more variables become important in distinguishing vegetation types. Vegetation was sampled in 2004—2006. ATD = Acer-Tsuga-Dryopteris, AVO = Acer— Viola-Osmorrhiza, and TM = Tsuga-Maianthemum. 93 Representation of habitat types Certain habitat types appear to be underrepresented on public land within the study area. That is, they occur on public land proportionally less then their availability in the landscape. For instance, the relative amount of area in TM, ATD, and FMC habitat types was higher in the entire study area than on public lands within the study area (Figure 1.5). Therefore, the range of vegetation conditions provided by those habitat types for wildlife, biodiversity, and timber are likely underrepresented in the landscape unless private landowners maintain those conditions. This is the case with specific vegetation types such as mesic conifers. Mesic conifers were dominant in the overstory on 39% of the western UP landscape during presettlement times. In May 2001 however, the DNR estimated that mesic conifers represented 6.7% of the state forest land in the western UP (R. Doepker, MDNR unpublished report). As such, the DNR initiated guidelines to emphasize increasing the mesic conifer component by 100% over the next 20 years on state forestlands in the western UP. These guidelines specify increasing representation of the spruce-fir vegetation type by 50% and increasing area containing hemlock by 1.5 times the current area. Guidelines for white pine and red pine are also specified. The TM and ATD habitat types can be managed to increase mesic conifers because spruce and fir dominate the middle successional stage of the TM habitat type, and the ATD habitat type has the potential to support hemlock. Both of those habitat types, however, are underrepresented on state lands. Approximately 9000 ha or 3.6% of state land occurs in the ATD habitat type (versus 15% of the landscape), and roughly 27,000 ha or 10.8% of state land occurs in the TM habitat type (versus 20%). It is area within those habitat 94 types that could be considered for land acquisition and where management efforts for mesic conifers should focus. Perhaps Michigan State University Extension could work with private landowners to strengthen cooperative programs to help meet the state’s mesic conifer and other ecological goals. Stands within the TM and ATD habitat types generally produced higher basal area (Tables 1.21—1.23) and timber volume (Table 1.24) in the 70-year age classes than in other habitat types. Basal area may be an important habitat component for wildlife species such as ruffed grouse (Zimmerman 2006). Increasing the amount of area on state lands within TM and ATD habitat types through land aquisition may also increase timber potential, which may help in designing management plans for sustainable timber harvest. Additionally, increasing area within TM and ATD habitat types may increase the potential of state lands to provide habitat for wildlife associated with hemlock (e. g., pine siskin [Carduelis pinus], evening grosbeak [Coccothraustes vespertinus] , yellow-bellied sapsucker [Sphyrapicus varius]; Yamasaki et al. 1999, see Chapter 2). The proportion of other habitat types (i.e., QAE, AQV, AVO, AOC, TTS) was higher on state land than within the entire study area, and the proportion of aspen within those habitat types was greater than the proportion of area within each of those habitat types. Therefore, the potential for management of vegetation diversity within those habitat types is greater than within the underrepresented habitat types. Unique characteristics provided by each of those habitat types include dominance of bi gtooth aspen on xeric habitat types (QAE), an oak component in QAE and AQV types, thick shrub layer in AVO and AOC types, and wetland components in the TTS habitat type. 95 Representation of age classes The representation of aspen in the 50- and 60-year age classes was relatively low (approximately 7%; Figure 1.4); and the representation of these age classes within some habitat types was almost non existent. Most notable, for example, was the small proportion (0.04%) of 50-year-old aspen within the ATD habitat type (Table 1.6, Figure 1.5). The fact that some habitat types contained very little aspen in certain age classes also may have implications for wildlife habitat and timber sustainability. Landscapes without specific age classes of aspen may lack vegetation characteristics important for certain wildlife species. For example, research in western North America has shown that bird species richness and diversity increases with ecological age of the aspen seral stage (Pojar 1995). In eastern North America, declines in several wildlife species such as woodcock (Philohelo minor) and bobcat (Lynx rufiis) have been attributed to maturing of forests and subsequent declines in the amount of early successional forests (e. g., aspen; Trani et al. 2001). My results suggested that wildlife species in Michigan would also respond differently to the significant differences in structure and composition of aspen within different age classes and habitat types. Specifically, older age classes of aspen in all habitat types showed an increase in basal area, DBH, snag decay and diameter, and a decrease in deciduous stem density and herbaceous cover. Cavity-nesters (e.g., woodpeckers, owls, bats) are likely to occur in older age classes of aspen due to those characteristics (N egn' 1995). Although stem densities were similar among all age classes and habitat types, the composition of overstory tree species changed with age. In the TM habitat type, balsam fir, white spruce, and maple species increased. Wildlife species (e. g, brown creeper [Certhia Americana], golden-crowned kinglet [Regulus satrapa], common 96 yellowthroat [Geothlypis trichasl) requiring a mixture of conifer and deciduous cover would likely find habitat components in 50—70-year—old aspen, as it is a transition from aspen to spruce and fir (Blake et a1. 1994). In dry, sandy habitat types (e.g., AQV), red oak increased, thus potentially providing a mast component in later stages of aspen succession. In mesic habitat types (ATD, AVO, AOC), hardwood species such as basswood and maple increased. It is plausible that communities of wildlife species would differ within and among age classes and habitat types. Therefore, it is important to consider the range of structures provided by different ages and habitat types and maintain them in forests to meet wildlife habitat objectives. Problems with consistent or sustainable timber production may also result fi'om lack of representation of aspen in specific age classes. Currently, levels of timber harvesting in the western UP are not sustainable because of the relatively low amount of 50-year-old aspen in the landscape (R. Doepker, MDNR, personal communication). Forest managers have had to start harvesting 40-year-old aspen stands to maintain harvest levels, which will likely lead to future declines in area of aspen stands 2 SO-years old in various habitat types. Managers may need to modify plans and perhaps harvest in older stands for a period of time to allow stand recruitment into the 50-year age class (see Chapter 4). Implications for Michigan’s Forest Certification Initiative Michigan’s forest certification initiative is a comprehensive system of principles, objectives, and performance measures to assist planning and forest management to sustain harvest while conserving wildlife and their habitats (Michigan Forest Products 97 Council 2006). Specific objectives for sustainable forestry have been identified through the SFI (Michigan Sustainable Forestry Initiative 2004) and the FSC (Forest Stewardship Council 2005), and results of this project can assist in meeting many of the objectives. For example, one objective is “to broaden the implementation of sustainable forestry by ensuring long-term harvest levels based on the use of the best scientific information available” (Michigan Sustainable Forestry Initiative 200424). Indicators for meeting this objective include current digital maps and a land classification system. The habitat type classification is an ecological classification system developed by Felix (2003) and refined in this project that used the most current information available to define boundaries of habitat types. Ecological classification systems can provide maps of land units that have similar biotic and/or abiotic characteristics, provide data that can be used to help describe the potential of geographic areas, and they can integrate biotic and/or abiotic information at multiple spatial scales to help understand dynamics of ecological processes and wildlife-habitat relationships (Felix et a1. 2007). This habitat- type classification system, for instance, is USCflll for understanding spatial distribution of habitat types and temporal changes in vegetation structure and composition throughout succession. It is useful for long-term management and planning because boundaries of habitat types are inherent landscape characteristics defined by soil moisture, texture, and landform characteristics, and explain much of the inherent variation associated with structure, composition, and distribution of potential vegetation throughout landscapes. Although boundaries and dynamics of habitat types are not static, they define a relatively narrow range of environmental conditions (Kotar and Burger 2000) that can provide a basis for predicting vegetation change over time with natural successional pathways or as 98 a result of certain management practices. Linking these predictions with habitat suitability modeling (Chapters 2—3) and timber production (Chapter 4) can aid in developing management plans that ensure long-term harvest levels that are sustainable and consistent with ecosystem-based management goals. Another objective is to manage the quality and distribution of wildlife habitats and contribute to the conservation of biological diversity by developing and implementing stand- and landscape-level measures that promote biodiversity and conservation (Michigan Sustainable Forestry Initiative 200426, Forest Stewardship Council 2005). The standards specify that program participants shall implement programs to promote biodiversity at multiple levels of biological organization including those that identify areas with unique vegetation characteristics, or that provide specific components of wildlife habitat (e.g., snags, mast trees, down woody debris). This project aids in accomplishing that objective 4 by quantifying characteristics of stands within different age classes and habitat types, and linking those data with the habitat type classification system to identify areas in the western UP landscape that can potentially provide specific elements of structure and composition important for wildlife communities over time. Without use of an ecological classification system such as habitat-type classifications, managers cannot effectively predict temporal changes in vegetation as a result of successional processes or certain land-use and management practices. Linking predictions with habitat suitability modeling can aid in evaluating the probability of wildlife species occurrence or persistence, and timber sustainability during a given time frame and location in a landscape. This approach can be useful for 99 identifying areas where management could enhance or maintain wildlife habitat, or unique features important for conservation of biodiversity. Davis et al. (2001 :77) wrote, “The empirical core of our professional claim to manage land scientifically and to ensure that owner objectives are met lies in our ability to predict the conditions and outcomes of current and future stands and stand types when managed under a specified prescription.” In essence, if managers cannot predict with acceptable accuracy the conditions and outcomes associated with implementing management prescriptions, it will be difficult to determine the extent to which ecosystem management goals are being met. The habitat-type classification is a tool for understanding temporal changes in vegetation structure and composition within ecological units and across landscapes, and for quantifying potential changes in wildlife habitat suitability and timber production potential. As such, managers can use habitat- type classifications for planning management activities, predicting and evaluating management outcomes, and sustaining fiinctional forest ecosystems while meeting human demands for resources. 100 CHAPTER 2 WILDLIFE-HABITAT RELATIONSHIPS: SPATIAL AND TEMPORAL VARIATIONS IN DISTRIBUTION OF WILDLIFE COMMUNITIES THAT USE ASPEN The conservation of biodiversity is a high priority for resource managers and conservationists (Lindenmayer and Franklin 2002), and it is one of the foundation principles of ecologically sustainable forestry (Hunter 1990, Carey and Curtis 1996, Noss 1999). Whereas historically, forest management took a very narrow focus by prioritizing production (Steams 1990), research today attempts to help managers understand and consider ecological processes (e.g., succession) and ecological relationships within forests (Haufler 1990). Within the past decade, researchers have identified important questions addressing topics critical to managing sustainable forests, conserving biodiversity (Lindenmayer 1999), and maintaining production important for human use (Kessler et al. 1992). Reoccurring topics include identifying and validating indicator species in forest planning, understanding community relationships, assessing spatial and temporal variation in forest structure and composition, and evaluating wildlife response to such changes. To help address these challenges in forest management and planning, Haufler et al. (2002) suggest the use of performance measures. Ecological performance measures are metrics used to assess how well management is succeeding at meeting specific ecological objectives. For example, Haufler et al. (2002) explain how specific metrics can be identified to assess diversity of ecosystems within landscapes, the range of structural, compositional, and functional conditions as well as processes within ecosystems, and species and genetic diversity. Performance measures can be identified 101 for multiple levels of biological organization and can help managers identify specific targets, goals, and priorities for ecosystem management. At the landscape-level, Haufler et a1. (2002) recommended a coarse-filter approach (Noss 1987) to characterize the planning landscape in terms of the range of ecosystem types that occurred historically and that allow for ecosystem integrity and biodiversity to be maintained if all communities are adequately represented. At the ecosystem-level, performance measures should describe structure, composition, function, and processes to identify the range of conditions that can occur within ecosystems and whether or not the entire range is represented within a landscape (Haufler et al. 2002). Community-level performance measures may include measures of species richness or diversity. A fine-filter assessment includes measures of species habitat suitability, and presence and productivity of species that select different ecosystem types to meet specific habitat requirements. Identifying performance measures and quantifying the range of conditions necessary for maintaining wildlife habitat conditions in ecosystems and landcapes is a complex process. At the landscape-level, ecological classification systems, such as habitat-type classifications can identify boundaries of areas with similar geological characteristics and successional pathways (Daubenmire 1966). Knowledge of the spatial distribution of habitat types within landscapes can help biologists and managers understand the potential of landscapes to provide specific ecological conditions. These ecological conditions can be described using performance measures that define spatial and temporal variations in structure and composition within habitat types resulting from natural disturbances or management (Haufler et a1. 2002). Understanding temporal 102 changes can provide a basis for predicting vegetation change over time within habitat types. Linking these predictions with community- and species-level performance measures can aid in evaluating the probability of species persistence during a given time frame and location within a landscape. Birds, especially neotropical migrants, are highly sensitive to ecological and landscape changes because many tend to have low birth rates and relatively long lifetimes (Maurer 1993). Although bird species showing declines in numbers are primarily neotropical migrants (Robbins et al. 1989, Brewer et al. 1991), many short-distance migrants (species that winter in temperate regions) and year-round resident species also have shown declines primarily due to declines in habitat area or quality (Brewer et al. 1991). As such, using performance measures of community structure and composition of birds that use forest resources for all or part of the year, and assessing habitat suitability of certain species can indicate ecological integrity (Haufler et al. 1996). Relating species metrics such as habitat suitability, and bird community characteristics (e. g., species richness, proportion of community that is threatened, endangered, or sensitive; proportion of neotropical migrants, short-distance migrants, or nonmigratory species; number and relative abundance of habitat specialists) to differences in vegetation physiognomy of different aspen age classes and habitat types can help managers answer questions about the potential of habitat types to provide habitat conditions for specific bird species, communities, and for species diversity. Spatial projection of habitat potential for birds can help managers understand the capability of areas within landscapes to provide specific habitat conditions for avian species and for wildlife using similar habitat components. 103 Biologists and forest managers in the Great Lakes Region can be challenged with developing forest management plans that sustain timber harvest and also maintain habitat conditions for a diversity of wildlife species. In Michigan, for instance, biologists are attempting to design plans to maintain current levels of aspen harvest, and simultaneously maintain the range of vegetation conditions provided by aspen in different age classes and habitat types. It is currently unknown, however, what ecological conditions are provided by different aspen age classes, specifically old aspen (i.e., Z 50 years), what wildlife species and communities may benefit from those conditions, and how conditions may change temporally. Applying the use of performance measures at landscape-, ecosystem-, community-, and species-levels can be useful for recognizing areas that can provide unique habitat components for specific species and avian communities, and other wildlife species using similar habitat components, identifying areas that may be currently below their potential for providing specific habitat components, and developing forest management models, which can be useful for planning and evaluating effective practices to meet ecosystem management objectives. 104 OBJECTIVES Specific objectives for this chapter were to 1) Identify and apply ecosystem-level performance measures to characterize the variation in vegetation composition within different aspen age classes and habitat types in the western Upper Peninsula. 2) Relate community-level performance measures for communities of avian species to successional changes in aspen within different habitat types. 3) Relate species-level performance measure to describe specific ecological conditions provided by different aspen age classes and habitat types. 105 METHODS Sampling of vegetation and avian communities I selected sampling points in aspen communities within 3 age classes (20 to 29 years, 50 to 59 years, and > 70 years) and within 6 upland habitat types that were the most predominant in the western UP landscape and characterized a range of soil conditions from dry and sandy to mesic loamy soils (Felix 2003). These habitat types included Acer-Quercus- Vaccinium (AQV; maple-oak-blueberry spp.) and Quercus-Acer— Epigaea (QAE; oak-maple-trailing arbutus) on dry sandy soils, T suga-Mainthemum (TM; hemlock-wild lily-of-the-valley) on dry sand or sandy-loams, Acer- T suga-Dryopteris (ATD; maple-hemlock-fem spp.) on mesic sandy soils, Acer- Viola-Oshorhiza (AVO; maple-violet-sweet cicely) on mesic loamy soils, and Acer-Osmorhiza-Caulophyllum (AOC; maple-sweet cicely-blue cohosh) on mesic loamy or silty soils. Several avian species were selected as indicators of the range of ecological conditions within aspen forests in different habitat types. These species were selected because research has described that they utilize various structural and compositional characteristics of forests (Table 2.1), and are sensitive to changes in ecological conditions. I quantified several habitat variables that described the food, cover, or reproductive requirements associated with the selected species (Table 2.1, Table 2.2) to determine the range of conditions that the aspen vegetation type could potentially provide within different habitat types. Each variable was measured within 10-m x 25-m plots established at Z 3 sampling points within each age class and habitat type. 106 The yellow-rumped warbler (Dendroica coronata) was selected because it has a strong association with conifer cover in the shrub layer (2.5—5 m) and high overstory cover (>5 m) (Westworth & Associates 1984). Hagan et al. (1997) observed that yellow- rumped warblers were most abundant in medium-aged (e. g., 20-60 years) and relatively old softwood forests dominated by balsam fir, spruce, or white pine. Aspen stands in which balsam fir and spruce or white pine are components of the canopy may potentially provide nesting requirements for the yellow-rumped warbler (Hanaburgh 2001). The American redstart (Stophaga ruticilla) was selected because it frequently utilizes vegetation in the 2.5—5 m strata for nesting (Westworth & Associates 1984) and prefers forests with high stem densities (Bond 1975, Hanaburgh 2001). Hobson and Bayne (2000a) observed American redstarts in numerous aspen age classes, most of which were >50 years). Veeries (Catharusfuscescens) are ground-nesting birds that prefer to nest in forests with a dense understory (Winnett-Murray 1991). Second-growth aspen may provide the dense understory necessary for veery habitat (Winnett-Murray 1991). Ovenbirds (Seiurus aurocapillus) were selected because they prefer to nest on the ground in forests with an open understory (Westworth & Associates 1984). The pileated woodpecker (Dryocopus pileatus) and black-capped Chickadee (Parus atricapillus) are both cavity-nesters, but require different sizes of snags to meet their life requisites. For example, snags suitable for pileated woodpecker habitat in the eastern portion of their range should be >38 cm in diameter (Schroeder 1983a), while chickadees can utilize snags 10—25 cm in diameter (Schroeder 1983b). Yellow-rumped warblers, American redstarts, veeries, and ovenbirds utilized Michigan forests during the breeding season 107 (i.e., May—June), while pileated woodpeckers, and chickadees, were indicators of winter habitat conditions as well as breeding habitat conditions. To evaluate wildlife responses to differences in aspen structure among age classes and habitat types, I conducted bird surveys during the breeding seasons and during the winter to quantify bird community composition throughout the year. I attempted to conduct bird surveys at each point where vegetation was sampled once during the winter (January—March), and once during territory establishment and breeding (May—June); however, due to limitations in personnel and vehicles, I was not able to survey every point. Bird surveys were 10 minutes long following a 1-minute settling period and were conducted during fair weather conditions (i.e., no rain or fog, winds < 20 kph, cloud cover <70%) (Hanaburgh 2001). Species presence was recorded at each point where a species was heard or visually observed. Winter bird surveys were conducted in 2005 (February 11—13 and March 7—12) and 2006 (February 26—March 7). Breeding bird surveys were conducted in 2004 and 2005 during the last week in May and first 2 weeks in June each year. Because of limited sample sizes for bird surveys each year, data were pooled to better represent differences in bird communities using various aspen age classes and habitat types. Data analysis I used principal components analysis (Morrison 1990) to examine multivariate relationships among vegetation structure and composition variables in different age classes and habitat types. This analysis was also used to identify groupings of variables that account for the variability in vegetation structure and composition observed among 108 different age classes and habitat types, and to define variables that were most important in describing differences among age classes and habitat types. Results of the principal components analysis were used to identify vegetation attributes that explained the distribution of birds observed in different aspen age classes and habitat types. I assessed frequency of bird occurrence observed during the winter and during the breeding season, and compared differences among age classes and habitat types with the Kruskal-Wallis one-way analysis of variance using SYSTAT (2002). I classified bird species according to nesting preference (i.e., cavity or non-cavity nesters) and migration status (i.e., year-round residents, short-distance migrants, neotropical migrants [Blake et al. 1994]), and compared average numbers of birds within each classification among age classes and habitat types using the Kruskal-Wallis one-way analysis of variance. Significance was determined at the a = 0.10 level. To evaluate habitat suitability for indicator species, I used habitat suitability index models for the American redstart (Minnis and Haufler 1994), black-capped Chickadee (Schroeder 1983b), ovenbird (Roloff 1994), pileated woodpecker (Schroeder 1983a), veery (Sousa 1982), and yellow-rumped warbler (Hanaburgh 2001). I compared average scores among age classes and habitat types for each species and related them to frequency of occurrence for the indicator to assess utility of them as indicators of conditions within various aspen age classes and habitat types. To identify groups of birds that associate with similar vegetation characteristics, I integrated principal component scores with bird observational data and graphed the first and second principal component scores at each observation for each bird species recorded. I then connected outer points of polygons created from graphing principal 109 component scores to determine the range of vegetation conditions that might be influential in determining species presence. I combined information from graphing principal component scores (i.e., identifying the range of vegetation conditions used by different bird species) and frequency of occurrence data to identify final bird communities that used specific aspen age classes and habitat types similarly. This procedure allowed identification of “clusters” or groupings of birds that could be explained by vegetation variables that distinguished different aspen age classes and habitat types. 110 coves—coach AEmvv .530 95? ..\o 228308ro couoéoaoc machoficoo 2030:6052 wfiuofl wcmuoom 92.20% mambo“ E2838? 9.. mA 550 ~30an e\e nouozwofie Eu wmA 805 we :28szro :8 3.3 much me $259 bigot wmcm cogs—coach AE mvv Emma: nPEm mambo» wnmcofl AE mAv Emma: magma mega E38333»: app :5 SA 95% 5326058 @508.“ cote—£558 A8 mvv nah—m 32888 couoscoaoa 33203 $63 5656 88m o~=n§< “ween $92 5:3: zoom 3&2 982 as mass: assume 885% awe 88% sausage :52 use: 28m?“ .8333 copay—030 bog “$890003 95:95 :85: woman—.7323» penance—uni BEBE .mwfiou 309$ :38 u 939 .Ewmon 58.5 as .568wa u £5 $2639: cm: :05 $2: 8 :33 mo momma—o own 33:? um: >9: :2: 38on 23 mo 358233» 038383“ was mambo.“ com H§EE_ 2a 35 $8953 $0.8m AN 0331 111 5:268ro E330.“ c2595 83 Emma wcmuoom QED Ea 383m mo bicep couosuoaoc 2083.898 2% wacm 2.383552 A83 and out 553029: AS _vv “260 30838: fl. 2353‘ “mecca 9&2 5:33 58 323. 3% new $556 awn—£285 530599 A93 820$ 830209 swoo— $335 “Havoc .5383 393—020 be? coo—0030003 25:26 505:2 wanna—-323» veggies—n «532E 3:09 AN 2an 112 .xcooA 9: _vv :96: 250238: Ax. so? €va 28 95: .x. .xcovv XomA 32233 .xconA $3-0m .xcmnA £8228: $8-8 AE mAv .535 30:3 X. 230.0 :8 wmA 82: mo dfiERA :5 3.3 82... .3 A25: 36:0: wag an? E mvv awe: £5: a 8A a 2A E 3 £32 328 2:3 .3: :8 _mA 2:05 82 88v 88.88 ea mvv £5: 32228 méommfimm 98:28: $53 38:0: 8on 1 2:3tfi< 520m 9&2 5:5: :22 £82 98% Ba mass: cwsnefimv 283% $2 9.88 080208 :2: 223: 2922 .6333 893—030 E23 80—02683 23:26 585:3. 3:82-326» 3:32:33 382E .230: 3:83 :38 n 95G .320: “223 E 53:33 .I.. 3Q 227.33: em: :65 B2: 3 5:3 Mo 8220 own 3223 2: 3:: can: 360% 23 :8 3:082:52 «83:: :5 33339: #28 mo 252:5 .NN 033. 113 95:: o fiA 3.5:: ~on :23 2%? :3: 93 383m ‘8 big :5 8.2 so a? :3: wag :5 ”W :5 En 8: 2333. “moan $02 5:3: zoom 3&2 Q32 2: mass: fisgaamv 588:va $2 389 “2.850% $3: :23: 2.8%“: 8333 00:83:? be? :oxooguooa 25:26 525:2 Bag—2-30:0» wammoiofln 628:: 3.58 .2 2:3 114 RESULTS Principal components analysis for vegetation variables Results of the principal components analysis on 25 vegetation variables showed that there were distinct differences in vegetation characteristics among age classes and habitat types. The first 3 principal components explained 41% of the variability in vegetation characteristics, and adding the fourth component explained 50% of the variability. Principal component 1 accounted for 15.5% of the variability and represented a contrast between an ecological community characterized by high herbaceous cover, taller shrubs, and less decay of down woody debris on the positive end. Communities on the negative end of the gradient were characterized by greater conifer stem densities, higher overstory conifer canopy cover and total canopy cover, and fewer total stems/ha. The extreme positive end of the gradient described aspen in the 20-year age class within the AVO habitat type (Figure 2. lb). Samples at the negative end of the gradient described aspen _>_ 50 years old within the TM habitat type (Figure 2.1c). The second principal component, which accounted for 12.8% of the variability in vegetation conditions, described a gradient between forest types characterized by average tree diameter and tree height (positive end), and deciduous stem density and total stem density (negative end). Communities in the 70-year age class within all habitat types were represented at the positive end (Figure 2.1) and were characterized by larger diameter and taller trees. Aspen communities in the 20-year age class were represented at the negative end of the gradient because trees in this age class were smaller, had more deciduous stems, and higher total stem density than older age classes. 115 The third principal component accounted for 12.3% of the variability in vegetation characteristics and represented a gradient between communities characterized by lower basal area, shorter shrubs, and moderate shrub stem density on the positive end, and communities characterized by larger and taller snags with more decay on the negative end. Scores for principal component 3 were on the extreme positive end for aspen communities in the 50-year age class within the AVO habitat type and young (e.g., 20-year age class) communities within the TM habitat type (Figure 2.1). Communities on the negative end of the gradient were 2 SO-years old and specifically included communities within the AVO and ATD habitat type (Figure 2.1). 116 ATD Habitat type AGE $020 XSO +70 PCM AVO Habitat type AGE $020 XSO +70 TM Habitat type AGE $020 X50 +70 Figure 2.1. Scores for the first 3 principal components describing the variability among 25 vegetation variables measured in aspen stands within 20-year, SO-year, and 70-year age classes occurring within 3 different habitat types in the western Upper Peninsula, Michigan in 2005. ATD = Acer-Tsuga-Dryopteris; AVO = Acer-Viola-Osmorhiza; TM = Tsuga-Maianthemum. 117 Frequency of occurrence of breeding bird species among aspen age classes During the study, a total of 41 bird species were recorded on the study sites (Tables 2.3, 2.4). Aspen in the 20-year age class (n = 22) had the lowest species richness (30 species; Table 2.3). Species richness of stands in the 50-year age class (n = 26) was 36, and richness of aspen Z 70 years old was 34 (Table 2.3). Twenty-seven species occurred in stands in all age classes and 8 were detected in only 1 age class (Table 2.3). The common yellowthroat (Geothlypis trichas) and pine warbler (Dendroica pinus) only occurred in aspen in the 20-year age class. The brown- headed cowbird (Molothrus ater), golden-winged warbler (Vermivora chrysoptera), mourning dove (Zenaida macroura), and northern parula (Parula americana) were only observed in aspen in the 50-year age class. The American robin (Turdus migratorius) and brown creeper (Certhia americana) were recorded only in aspen Z 70 years old (Table 2.3). Ovenbirds were the most frequently observed species in all age classes (Table 2.3). The red-eyed vireo (Vireo olivaceus) was among the 3 most fiequently observed species in all age classes. Pileated woodpeckers were among the most frequently observed species in aspen in the SO-year and 70-year age classes, and the black-throated- green warbler (Dendroica virens) was among the 3 most frequently observed species in aspen in the 20-year age class (Table 2.3). Of the 5 bird species identified as potential indicators of specific forest characteristics, only the pileated woodpecker exhibited differences in absolute frequency among age classes (p = 0.017; Table 2.3). Pileated woodpeckers were most common in aspen Z 70 years old and least common in young (< 50 years) aspen stands. 118 Breeding bird species that were more abundant in old (2 50 years) aspen, included the black-and-white warbler (Mnioticta varia; p = 0.089), brown creeper (p = 0.065), Eastern peewee (Contopus virens; p = 0.020), Northern flicker (Colaptes auratus; p = 0.091), pileated woodpecker (p = 0.017), pine siskin (Carduelis pinus; p = 0.079), red- breasted nuthatch (Sitta canadensis; p = 0.094), red-headed woodpecker (Melanerpes erythrocephalus; p = 0.002), and white-breasted nuthatch (Sitta carolinensis; p = 0.094). Frequency of occurrence of breeding bird species among habitat types TM, ATD, and AVO habitat types had representation in all 3 age classes. Of those habitat types, the TM habitat type (11 = 30) had the highest species richness (36 species), followed by the AVO habitat type (11 = 20) with 35 species. The ATD habitat type (11 = 9) had the fewest observed bird species during the breeding season (24 species; Table 2.4). Twenty species occurred in stands within all habitat types, and 7 were found in only 1 habitat type (Table 2.4). The American robin, brown creeper, and yellow-rumped warbler only occurred in the TM habitat type. The brown-headed cowbird, magnolia warbler (Dendroica magnolia), mourning dove, and northern parula were only observed in the AVO habitat type. No breeding bird species were unique to stands in the ATD habitat type (Table 2.4). Ovenbirds were the most frequently observed breeding bird species among all habitat types. The pileated woodpecker and black-throated-green warbler were the second and third most common species in the TM habitat type. The black-throated-green 119 warbler and red-eyed vireo were the second and third most frequently observed species in the AVO habitat type. Of the 5 indicator species, the ovenbird did not exhibit differences in frequency of occurrence among habitat types (Table 2.4). The American redstart was not observed in the TM habitat type (p = 0.043). Frequency of occurrence of pileated woodpeckers was highest in the TM habitat type (32 = 72.7, S.E. = 7.9), but did not differ from frequency in the AVO habitat type (i = 55.0, S.E. = 11.4). Pileated woodpecker fiequency was lowest (SE = 11.1, S.E.= 11.1, p = 0.004) in the ATD habitat type. The veery was most common (p = 0.058) in aspen stands within the AVO habitat type (Table 2.4). The yellow-rumped warbler was only observed in the TM habtiat type. Other bird species with higher fiequency of occurrence in the TM habitat type than in other habitat types included black-and—white warbler (p = 0.079), pileated woodpecker (p = 0.004), and brown creeper, although it was not significantly different. Species observed more frequently in the ATD habitat type included the black-throated- green warbler (p = 0.050) and the eastern wild turkey, although frequency of turkeys was not significantly different from that in the AVO habitat type (Table 2.4). Frequency of occurrence of the blue jay (Cyanocitta cristata; p = 0.086), chestnut-sided warbler (Dendroica pensylvanica; p = 0.002), eastern peewee (p = 0.002), and scarlet tanager (Piranga olivacea; p = 0.05 8) was lower in the TM habitat type than in other habitat types (Table 2.4). 120 0020—0 0?... m... .o 3 0.0... a... 3..- s... 3.. ea... in... .8... 3. .3. 3 saw. 3. .3. .5333 8053-528 :2. :0 E. .0 m... 3. 3.. .22... Bases-52o: 8.... E E 3 E 3. 3.. 3...... 2.3 83...: as... 3. .33 3. 2.8 m... .m... .858 50...... 8:. A... 3.. on E 3 an. .3888... $3.98: 2.3. 3. .0... 3. .3. 3 .2 08532.2 88:80 E... N... SN 8... SN 3. 2.. as...» Bea-35.2.0 82. 3. 3. .2 an 3. 3. 2538 Engage: m8... 3 .3. 2. .3. 3. .3. .888 use: 8:. S :2 3 5m 3 as a... 8.: SE. 2 3... o... 0.2 2.. 0.8 as...» 50%-00.025-03.25 .8... E 0.... 3 N... 3. 2.. as...» 83-00.85-820 on}. M... Z... N... w...»- 2. 4.0m 8.3%.... 888.820 2.3. we 2...: S .42 3. .3. .53.... 0.33-0803: 8 . .o E 3- 3. 3. 3. 3. use $.85... 83. E 3 2 E 3. 3. use... 5888... 82. 0.8 0.3 as 3.. N... Em .65 5.8.3.. __0>0_ .m.m M .m.m M .m.m M 0:8: :oEEoU £588.. fin... .803 an"... 32.8 am"... 32-8 .moomlvoom :. 5w322 83220.. 0&3 05 E :88... 36029 0:. mg: :25. .0 8000.0 09. m E 00.32% 00.00% 23 .0 Adam. 2020 030:8... 0:0 82.5000 00.00am £023 .0 3:6: .0 E0203 36:032.. 0.2030 9.02 .mN 030:- 121 122 23. E E E E E E. 5...... .83 83. E E. E m... m... m... .5... 55.3 E... E .E. E .E E .35 Eon... 5.85.3.5 33. E .E. E ..E 3. .3. .555... 35.55-...E 3.... E E E E. E E. b5.5 23. E E E m... m... m... 5...... 5.5% at... E E. 3. :3 E E 5...... 8...... E3. 3. 3. E E E E 55...... 8.52.5.3. 83. E .E E .3... m... .m... .5...5....83 3.5.-.3. 2.3. E ..E. E .3... 3.. .36 .025 55-..... .33. E ....m 3.. .E. E .E~ .555... 55.2.-..3. E... E E" E E- E E. .52... o. 3. 3. 3. 3. 3. E E 5...... 2.... .8... E .E. E ..E. E .E 2...... 5.... :3. E 2.... 3.. 53... 3.. .E. 28.5%....» .552... m. . .o 3. 3.... E E. m... E. 23.55 .53. 3. 3. E E 3. 3. .35.. .5552 33. E ..... E .E. 3. .3. .55... 55......2 3...... E 3... E 3..- E EN 5......» 2.5.5.2 .53. 3. 3. E E 3. 3. 26.. wee-5...). E... E E 3. 3. m... m... 5......p 55.5.2 :05. .m.m M .m.m m .m.m M 08...: :oEEoU 5.3.8... 3...”... 5..-... GT...5-%-% 8...“... 53-3 Emma—o 0w< .58. E 53.. Aga— 5:830 23 _mmeH ocmcfim =8ng8 23:2: m:_m3-_mxmavc 2: .o A 3 Echomtc Ho: 8a 6:2 25mm 2: 5:5 0:: 08mm 2: co 82.3... .933 5:830 95 _owflmv 3.85; no max—mam 33-28 m=_m3-_§m5 on» 52> 6823.8 0.83 var—82 203— 5:538;— wwad S. 5“ M2 3. 2 we 5333 338-3%; 23 3 E to w: 2. m... 30:88 858-32; 33. E :3; mm .3 3 .3. ABEHEEE gov—003003 __o>o_ .m.m M .m.m m .m.m M 258 :oEEoU 5:33on fins 3%-? GT5 3%-? fins 322a 8320 ow< A208 .3 03,; 123 124 33. 0.. ...m 3. 3. 3. 3. .2... .5882 .2... o. ...m 3. 3. 3. 3. as...» 3.8.... 8...... m. . 3.. S. 0...... m... 3... 5.5.. .65: 8.... N... .2. 3. 3. 3 3. 3...... 3.53-528 $3. 3. .3. 5.. .~.- 3 .Nfi L23... 85.8952... .8... a. e3: 3. .32 3 .3 3...... 2.3 8...... 83. v. .. .3.. .. .. ..... N... .3 .83... 5.5 £2. a. 3.. 5.. as ..m 3 5.8%.... amiss... .33. 3. 3. ..: .... 3 .2 ......§..... 858.0 83. v... .3.. ..: ..... a... .3 .53.... .......-.=5....o .2... Q. ...m 3. 3. 3. 3. 233.. ........-§§m $3. 3. 3. 3. 3. 3 3. 5...... 5.2.. £3. N... .3... 3. .3. E .02 a... .5... 33. 3.. 32 5.. .3. m... .3... .53.... ....w-......:..-.....m 83. q. .3" E. .2.” m... .3 .33.... .2..-.....e...-.....m £3. a. 3.~ e... 3... 3. a...“ 8.8.2... 89.8.3.5 33. o. .....m 3. .3. M... .3.. .53.... 823.528... 3.3. 3. 3. 3. 3. 3 3.. .58 5.5.5. $3. a. .3.. 2.. .~.- 3. .3. .23.... 5.58... 33. N... 3... .... .... 2. v.3. a... 5.5:... .35. .m.m M .m.m M .m.m M 0.5.: :oEEoU £32.... 8N”... o>< El 3 .3383..Q-So.§ 18v. .1. O>< .mtm.o&§Q-aM:§imuV H OH... .Eafimfizzmuxx -3MEK u SE. .moomlvoom E 53:82 diagram .233 05 E 833 mfiwoob 2: mafia. comma mamtomaa 390. .832 m 5 332.5 8.8% ES (.0 émwv 22.8 $838... 9.... 88.502. 8.8% new? 3 352. we 2.3.83 8.2.2.3... 820an 5.02 6N 2an .wai 3:280 “Ea _awfifl osmzfim :omtmaEOo 29:2: mEmBAav—maxv 8_.o A 3 “5.8%: 8: 2: .852 2:3 05 5:» 0:: 083 2: :o mos—.3“. .933— :a=3mao 28 _owo_mv ouSEg no £932: $3-80 £33-3me 05 fits 35.3.8 203 3:03.. 232 323305. 33 o: S. o: 3. 9m 3 5333 832.33% $3. 3 3.; 9o 3. 2 2 388a 3:03.32; 33 ca cm 3 a... .3 3 $052025 38:80? 2: .o 3 3. 2; SN 3 3: :35 803 :3 Na :2 o.o e... an :2 =23 5&3 39° 3 i: _.: :2 MG 2: 32:2: 385323 38 ow 3 ed a... 3 G 5255:: Bfieefifi, ”mod ,3 3.3 o.o .3 3 .3 38> $2 3 cm: a: 2— ? 3 swans §§m 33 Na in _.: :2 MG «.2 3:on 8::: $3 ow em o.o 3. me E :88.on 3303-08: 23 Na :2 :2 man 3 :2 38:89,» 882-3: 2 3 Na a...» :2 :8 3 3m 8:3 33-3: ~85 a: 3; 9: «.3 3 t? 535:: 3303-8: at: 3 e.m o: a... MS 2: 5%: $3 9m 5. 5; 2a 2 QR 3%.: BE Sod v: 2.3 _.: £3 _ 3 .25 138:8? 8323 8:3 3 35 9o 3:: 3 a: 3955 82 ow 3 9o 3. 3. a... was: Eofioz SS 3 _ a2 a: 2: 3 an 32: 50582 33 2: can :.: 3. 3 man 5533 2:232 296— .m.m M .m.m M .m.m M 0:8: :oEEoU €332: 8N u $92 a u 5 PE 2 m u 5 2H 3:08 .vd 2an 125 Frequency of occurrence of breeding bird species among age classes and habitat types Breeding bird surveys in aspen in the 20-year age class occurred within the QAE (n = 2), AQV (n = 5), TM (11 = 9), ATD (n = 4), and AVO (n = 2) habitat types. In the 50-year age class, surveys occurred in the TM (11 = 11), ATD (n = 3), and AVO (n = 12) habitat types. Surveys in aspen stands 2 70 years old were conducted within AQV (n = 3), TM (11 = 11), ATD (n = 2), AVO (n = 6), and AOC (n = 3) habitat types. Only the ovenbird was observed in aspen stands within all habitat types represented in the 20-year age class (Table 2.5). Three additional species were common to 20-year-old aspen in the TM, ATD, and AVO habitat types (black-capped Chickadee, black-throated green warbler, and hermit thrush). Six bird species were observed in stands in the 20-year age class within the QAE habitat type. Eight species were found in AVO. Species richness in the AQV and ATD habitat types was 17, and was 25 within the TM habitat type (Table 2.5). Within the 20-year age class, frequency of occurrence of chestnut-sided warblers was highest (p = 0.026) within the AVO habitat type. Eastern wild turkey frequency was highest (p = 0.033) in the ATD habitat type. Frequency of occurrence of other bird species was not significantly different among habitat types represented in the 20-year age class. In the 50-year age class, 8 bird species were observed in aspen stands in all habitat types represented: TM, ATD, and AVO. These species included black-capped chickadee, black-throated-green warbler, hermit thrush, Nashville warbler, ovenbird, red- breasted nuthatch, red-eyed vireo, and red-headed woodpecker (Table 2.6). Among stands in the 50-year age class, species richness was highest in the AVO habitat type 126 (31), followed by the TM habitat type (25), and the ATD habitat type (12). The eastern peewee was the only species exhibiting statistically significant differences in absolute frequency among habitat types represented in the 50-year age class (p = 0.013), which was only observed in the AVO habitat type, and the red-breasted nuthatch (p = 0.072), which was observed most frequently in the TM habitat type. This difference was not significant in comparison to the ATD habitat type, however. Other noticeable, but not statistically significant patterns included presence of the American redstart, brown- headed cowbird, eastern wild turkey, mourning dove, northern parula, and red-breasted grosbeak in only the AVO habitat type. The winter wren, wood thrush, and yellow- rumped warbler only occurred in the TM habitat type (Table 2.6). The eastern peewee, ovenbird, red-breasted nuthatch, and red-eyed vireo occurred in all habitat types represented in the 70-year age class (Table 2.7). Bird species richness was highest (28 species) in the TM habitat type, followed by the AVO habitat type (21 species). Richness was 17, 16, and 9 in the AQV, AOC, and ATD habitat types, respectively (Table 2.7). Eight species using aspen Z 70 years old were unique to only 1 habitat type. The American redstart was only observed in the ATD habitat type (p = 0.004). The black-and—white warbler (p = 0.060), brown creeper (p = 0.060), golden- crowned kinglet, and veery were unique to the TM habitat type. The magnolia warbler and red-headed woodpecker were only found in the AVO habitat type, and the Eastern wild turkey was only heard in the AOC habitat type (p = 0.002). Other notable differences in fi'equency of occurrence of bird species among habitat types represented in the 70-year age class included the absence of black-capped chickadees in the AVO habitat type (p = 0.021), presence of chestnut-sided warblers only 127 in hardwood habitat types (i.e., AQV, AVO, AOC), and their more frequent occurrence in the AQV habitat type (p = 0.001). Absolute frequency of northern flickers was highest in the AQV habitat type (p = 0.033). Pileated woodpeckers were not recorded in the ATD habitat type (p = 0.032). The red-eyed vireo was observed in all habitat types, but was lowest (p = 0.019) in the TM habitat type. White-breasted nuthatches occurred in the AQV and TM habitat types, but were most common (p = 0.092) in the AQV habitat type. 128 and ed :6 9mm an _.: 1: od 9.: od :6 £02m 0:5 129 82 ed 3. QR 9.3 g: m.mm 3m 33 gm gm 38383 BE: 8% od 3:: od 3:: ed 3:: 98 a...» 3 3:: 2555 $2 3:. 3m 3 3. 5.2 man 2K 3a 3. a... 5333 2:332 Sod od 2...... ed :3. 3 2...: ed as... 0.8 25m 10533 £8me 33 ed .3.: 92 can 0.: 3... 2m 35 S. S. 585 :53: :2 ed a... fix gm S; 3” od 3. ed a... Ewes. 8:36-525 33 ed ea... o2 same a: 2.: od .3. od .3. 38:3 2? 888m 3: od 3. ed a... _. : _.: ed a... 3 a... 35% Beam ES od 9... 92 can _. : _.: SK 3.” ed a... axoacooa beings and ed 3. QR 9mm _.: a: 3. 3. ed a... sags? 8858 83 ed 2.2: 3 .3. S; .32 ed .3. od .3. Ease Baggage we; 3. 3. o.o 3. 9: 9mm qom :2 ed a... 3 35 mo _ .o ode 9.2: o.mm cab 5.3 ”Kb gym adv od :6 8383 soohwéoumefiioflm 32 gm 3% ed 3. _.: _.: 3m 3:. od 3. 53s.; Ssaegégosm 83 9% 9% 3m 3% 5.2 an». 2% 3:. S. 3. 83020 BEScEm «who ed 3. 9mm :3 0.: 9% OS =2 3. 3. 380 505:2 232 .m.m M .m.m M .m.m M .m.m M .m.m M 08a: .8500 £38on fl an: E fia fie ~o>< 85. 9:. 90¢. 85 mmflo owe Somém E Hog—38%.— momb “85$“ .onEgoEwQéBi -xmoV n O>< .mtmionémaHimoV n QH< .E:Eu£:£€<é%§& U SE. .Eéfiuock -mzuxmzmimow n >O< 633%“?wa V.§u.$:® H m<0 .moom Ivoom E gwfloaz .flsmficom e233 05 E c938 @5305 05 wage 390. “853 Subway GEE» mmflo own 83-8 05 E comma 5 3523 860% 23 mo Adm—.9 macho Buggy. 28 38.880 860% £023 3 mafia .«o “5803 mfloeoscofi 828% :32 .md 03an A33 5:380 can _owflmv 85:9 :0 max—£8 .3368 £33-..“me 05 5:3 35:28 203 v8.83: £92 9.5325? .mea 5:830 can _mwomma 23:3: acmtaqfion. 23:2: m___n3-_8_mav_v 8.6 A 3 E9356 8: 2a .530. 0E8 on: :23 2.: 0E3 on: so mo:_u>* 130 8:: Q: Q: Q: Q: Q: Q: QR QR Q: Q: 5333 3:5a-ao__o> 3:: Q: Q: Q: Q: Q: Q: Q: Q: Q: Q: sagas 853-38; 3.: Q: Q: QR Q:m Q: Q: QR QR Q: Q: :35 :83 8:: Q8 eQ:m Q: eQ: Q: .Q: Q: eQ: Q: a:.: as; 5:53 5.: Q: Q: QR QR S: QR QR Q8 Q: Q: 305% Ba:§-~:§ ::Q: Q: ._Q: Q: ._Q: Q: ._Q: QR :QR Q: .Q: 38> SQ: Q: Q: QR QR Q: Q: Q: Q: Q: Q: 5:28 :2an R:.: Q: Q: QR QR Q: Q: Q: Q: Q: Q: 08% Bee: RR: Q: Q: Q: Q: S: «R Q: Q: Q: Q: cacao“: 3:55-33: 3:: Q: Q: Q: Q: 2 _ 3 _ Q: Q: Q: Q: 5:89:83 3:833: 8 _ .: Q: Q: QR QR Q: Qmm QR Q:: Q: Q::_ 8E 3:33: Rn: Q: Q: QR Q:m Q: 2: BR Q:.. Q: Q: 535:: 85853:: :8: Q: ._Q: Q: :Q: Q: :1: Q: ._Q: Q: .Q::: E93: :Qm: Q: Q: Q: Q: Q: Q: QR QR Qom Q:m sass 2:: 292 Q: M Q: m .m.m M Q: m .m.: m Ba: 8868 3:532: E at: and: Gina E 83. 8.? N2: 90:. N93 mmflo owe Hotném E wofiomoao: womb HEB—mm 3:8: .3 use: :R: Q: Q: Q: Q: Q: Q: 28:: 8:582 131 :: _ .: Q2 QR QR QR Q: Q: 5:2: E2202 9:: Q2 QR QR QR QR QR 3333 2=>§z :R: Q: Q: Q: Q: Q: Q: 26: 3:335: ::Q: Q: 5.: QR 5.: Q2 :4: :25: :Esm 3:: Q2 QR Q: Q: Q: Q: 3333 8:53-828 :2 .: Q: Q: Q: Q: Q: QR was: :2389528 8:: Q: 5: Q: Q: Q: Q: :83: 23 633: 2:: Q2 ._Q:: Q: .Q: Q: .Q: *838: 823: :Q: Q: Q: Q: Q: Q: Q: 5:03:83 386338 :8: Q: ..Q:: QR ..:.R Q: .Q: $0333 32:35:05 :R: Q: Q: Q: Q: Q: Q: 2338 882338: RQ: Q2 :.:: Q: Q: Q: QR E 2:: momd NE mdm m.mm v.3 N6: ed.» .8333 coo:w63:oafi-xo:_m 2 Q: Q: QR QR QR Q: Q: 5333 33-38532: SQ: Q2 QR QR QR Q2 QR 8:325 B::8-:§m :R: Q: Q: Q: Q: Q: QR 5333 323:5:85 8:: Q: 5: Q: Q: Q: Q: 23:8: 505.5 :R: Q: Q:: Q: Q: Q2 QR 32: 5:55: 226— .m.m M .m.m M .m.m M 08:: 88800 3:332: RE: :o>< EH. 3 :::_o om: Roxém E 36323: :25 53:2 .QNESE:O-SQ.S 989:: u O>< GERQQDQQMRR £8“: u QH< .E:E:S§:QV<-QMEH u E .38 tween 5 Q3232 63:55:: 595 :5 5 5:3: 9:685 :5 meta: :09»: 3:52 380:3“: 55:3 :36 own 323% on: E 5%: E @9333: 862:: ES :0 Qmwv Robe BEES: 2:: 38.580 862:: £033 Q: $52: :0 33:03 8:05:68: 820:8: 5.62 dd 033. .232 5:230 can _mwflfl ozmtfim fixings—co 29:2: was 3-18.3.0: 8 _ .o A 3 Eubbfi 8: 0.8 .532 2:3 05 5:5 0:: 08mm 05 :0 32.3.. A32 §__3mao Ea _owo_mv 35:? («o mum—£8 33-05 £33-..“me 05 FEB 33328 0.83 v8.53. £26. gingi— 82 ed a... 3 a... 3 § 3533 832-3%? $3 a: v.3 ed a... g 2 H3883 3:03-323 ammo 3 m.» 3 a... 3 3. ABEEESV 38%83 82 o.o 3. ed a... S Sm 5.35 803 83 ed 3. ed a... 2; 3m 55§=§ 53 3 3 3 a... 3 Q Bean; B§£¢§§ ”who 3 m.» 2. 3. 3 e... 3225: 33290:; 2 3 3: Sn 3 a... 3 E EMU> N85 3 3 9o e... NS S: 5925 62.3 was 3: can o.o 3. a: 3m anew Bram $2 m.» m.» 3 3. ed 3. figmew Ensséom $3 a: 2.3 3 .32: 2; .2” agave? Bussém $2 a: 0% od 3:: I: 3» 8% cesium S; a: .2 02 an? 2: ans Loafing BfieeBm SS 3 3. ed a... ms 2: 55“ mm; 3 a... a? 0mm :1 n2 5% BE 8 _ .o :3 3m 3 a... NE 88 anu83 8325 Rod a: an” ad 3:: 3 $5 2585 20>»: .m.m m .m.m M .m.m M 0:8: 58800 $38on GTE mo>< fins 82 2 Fe m2 330 own 392% 5 333838 89¢ 23me 3:08 dd 033. 132 I33 .5... 9.. 9.. S. E: 9.. 9.. 9.. 9.. 9.. 9.. 5:33 2.82.: 2:... 9.. 9...: 9: 9mm 9.. 9.. 9: 9:. 9mm 98 525 BE: ME... 9.. 9.. 9.. 9.. 9.. 9.. 92 EN 9.. 9.. .22.... Bazoséusc 2...... 9mm .9... 9.. .9.. 9.. .9.. 9.. .9.. 9.. 59.. 38.5. 2.3 533m .2... 9mm 9mm 9mm 9.: 9.: 9.: E E 92 9mm 8%.... 523m 89.. 9.. 9.. E. 5: 9.: 9.: E E 9mm 9% 3.02683 2.922.562 2.9. 9.. 9.. 9.. 9.. 9.. 9.. 9.. 9.. 9.. 9.. 902.32.... 8858 .2... 9mm .9... S. 2.9.: 9.. .9.. 9.. .9.. 9.. .92: $0333 8.5-5532. 89.. 9.. .9.. 9.. 29.. 9.. .9.. E. .93 9.. .9.. 12.85 555 2...... 9.. 9.. EN 9% 9.. 9.. E E 9.. 9.. 3 02m 2 9.. 9mm 92. 92 5: 9.. 9...: 3: 9mm 9.. 9.. 3333 502-90.85.85 2...... 9.. ...... 9.. .9.. 9.. .92: E .E 92. e9»... .5333 B...-.§8§-§.m .8... 9.. .92: 9.. ._9.. 9.: ......m 9: .3.. 9.. .92: .8302”. 8.28-285 ooo.o ed «:6 ed as... od .3.: NS afima od «:6 “.5333 83365.-..85 .2... 9mm 9mm 9.. 9.. 9.. 9.. E E 9.. 9.. use 98:25. 2...... 9.. .9. 9.. .9.. 9.. .92: 9.. .9.. 9.. «9.. 29%.: 58:22 .. 3... 9mm 98 In 9% 9.. 9.. 92 9.: 9.. 9.. so... 5055... .33. .m.m m .m.m m .m.m m .m.m m .m.m m can: 8888 2.382.. mule. on: Gui... a... Hula Bo... 5>< ES. :5 E>o< 820 own Hobos E woucomoao. $93 233...: .539930390-33.BEwOLmuV H 00¢. 632.85.56.03 «Sow n O>< .9...m.§bQ-9M:§LmuV H QH< .E:Em£2£§<é%:u& n S: €52.83. -9:gxm:®-gmuv~ n >O< .moom Lnoom E S.E.—0.2 dEmGEom 6&3 05 a. .838 $.68... 05 mats: 89$ :3an 2.80%... 22:3 mam—o own 392:. 05 E 5%.. E voxogm 8.8% 23 .0 Adam. 3on 93:83 was 38,380 3.8% £033 3 8302 .«o .2358 36:25on 828%. 522 SN 033. .232 5:230 can _awflfl uzmtfim cemEano oat—=8 m___m3-_8_mav_v 8.6 A 8 «:20ng 8: 2a 3:2 2:8 05 5:5 0:: 2:3 05 co mo:_m>* A32 5:880 23 _owflmv coast? .8 max—m5 $3-28 m___a3-_8_m:.0_ 2: 5:» 603323 0.53 3:32 £32 3:92.05? wood 3m .33 ed .3. i. .2. ed :3. ed .3. 10333 282-323.» 53 ed 3. E2 5: 9o 3. E E 3 a... 38an B___Béo__o> SS 3 =3. o.o ‘3. 3 .3... 3: £12 a? .55 :Bmufinesv 302683 89o od 3. o.o 3. ed a... 2: «.2 2. a... £25 803 :3 3 a... «.2 gm 3. 3. E E od 3. :23 SEE 8M3 9o 3. ES 5: 3 e... 3: «.2 mm man 323% Baéfii? SS 3 :3. 3 ‘3. 9o :3. 2: ‘15 3M .3» 2035:: 38052;? was ad 3. od 3. ed a... E E 3 3. 50> N2 .o ed 3. EN «:3 ed a... ed a... 3 a... 389 53% $3 Qo e... E2 5: 9o 3. E E mmm man @8on Buzz 83 ed a... 3 a... 3 a... o.o a... 3 e... figmew Baeeaé whvd od c... 5.2 5.: ed a... ed a... od 9.: gov—003003 Eamonéom 23. on 23¢ 9o .3.: 3 a3:: oi :3». 3 2.2: *8? 836% 53 mm mam ES a...» 9% 3m 3: 3m 2m 55 £223: Bumozeom 2 2 3m 55 E2 3— 3 a... NE EN 3 a... 52mm 33 mm 3m E2 ES 3 a... 3; 3m mm mm 5% SE med 3 .3.: ES .2» ed :3. E «ma 3m «58 138%83 8325 83 ed 3:: 3. 3:: Qo 3:: ed 3:: 3 .3.: 2555 mad 3 :3. ed as... 9o :3. E aE 2m .55 183E eofioz Ed 3 3:: _. a man 3 3. 02 man 3... 3o 8333 2:532 :82 .m.m m .m.m m .m.m m .m.m m .m.m m 2:2 8888 E5305 Bo< Eo>< Lg E5 Edd mmflo own 393:. 5 3882mm: 890 3:33 388 .E 03$ 134 Frequency of occurrence of breeding bird species throughout succession Within the TM habitat type Throughout succession (i.e., from 20-year-old aspen to aspen 2 70 years old), bird species richness remained relatively constant (25 species in 20- and 50-year age classes; 28 species in aspen Z 70 years old). Sixteen species were common to all 3 age classes, and 11 species were found only in one age class of aspen within the TM habitat type (Table 2.8). The chestnut-sided warbler, common yellowthroat, Eastern wild turkey, and rose-breasted grosbeak were only observed in the 20-year age class; however, only the frequency of occurrence of the chestnut-sided warbler and rose-breasted grosbeak differed significantly (p < 0.10) from 0 (Table 2.8). The yellow-rumped warbler was only observed in the 50-year age class, but absolute frequency was not significantly different from 0. Species only observed in aspen stands 2 70 years old included the American robin, brown creeper, magnolia warbler, northern flicker, white-breasted nuthatch, and unidentified woodpeckers. Of those species, only the frequency of occurrence of the brown creeper differed from O (p = 0.085). Within the ATD habitat type Within the ATD habitat type, bird species richness decreased over time in aspen stands from 16 species observed in 20-year-old stands to 9 species observed in stands 2 70 years old. Five species were common to all 3 age classes in the ATD habitat type. Eight species were only observed in the 20-year age class, including the common yellowthroat, eastern wild turkey, golden-crowned kinglet, pileated woodpecker, ruffed grouse, scarlet tanager, white-throated sparrow, and wood thrush. Of those species, only absolute frequency of the eastern wild turkey statistically differed from O (p = 0.082; 135 Table 2.9). Four species were observed only in the 50-year age class of aspen in the ATD habitat type, including the chestnut-sided warbler, Nashville warbler, northern flicker, and red-headed woodpecker. Of these species, the Nashville warbler and red- headed woodpecker differed from 0 (p = 0.050 and p = 0.018, respectively). The American redstart was only recorded in stands 2 70 years old within the ATD habitat type (p = 0.018; Table 2.9). Within the AVO habitat type Within the AVO habitat type, species richness of birds during the breeding season was highest (31 species) in aspen stands in the 50-year age class. Species richness was 21 in aspen 2 70 years old, and lowest (8 species) in aspen in the 20-year age class (Table 2.10). The black-throated-green warbler, chestnut-sided warbler, hermit thrush, Nashville warbler, and ovenbird were observed in all aspen age classes within the AVO habitat type. Eleven species were only observed in the 50-year age class, including the American redstart, black-and-white warbler, brown-headed cowbird, eastern wild turkey, golden-winged warbler, mourning dove, northern flicker, northern parula, rose-breasted grosbeak, veery, and white-breasted nuthatch. Of these species, only the absolute frequency of the veery significantly differed from 0 (p = 0.071). Three species (i.e., the magnolia warbler, pine siskin, and raven) were unique to aspen Z 70- years, but the absolute frequency of none of these species significantly differed from 0 (Table 2.10). 136 momma? 09» 82 E E 3. a... 3 3. 32: 88:02 ooo.o :2 men :2 3m 8: 3m 5333 2:382 83 3: us :2 a; 8.: 3... :35 .85: oz; o.o «a... 2. a; o.o we... .8333 8853-828 ooo.o NS :3 2; EN :3 «.2 page. 80:38:20.8 82 ed a... od 3. _.: :2 5:3 28 583: 82 E E 3 3. a: E: 888 ass: 83 E E E 3 2 : Z _ 888803 caisson 82 ed a... 3 3. _.: 3 _ 8.55:8 8888 83 ed as... 3 s3. 2.: .EN toss; 82...-3586 83 ed 8... ed 8... od 3. 235... 888-858: 33 we :13 9o .3. ed .3. 10820 :32: £3 E ._E 2; .EN 8.: .93 .3 2;: wmmd v.3 w.mm NE v.3 :1 Qt. .8333 :ooQéoEEfiiom—m 82 E E ed a... _.: a: 5333 83888.82: N56 v.3 3.. :2 v.8 E2 «:8 08325 888-82: wood NS 1.8 2: sEN o.o .3. 12:33 22368-82: 82 E E ed 8... ed a... :32 50:05. 83 ed 3. 3 a... go e... 2382 508:2 83 a: man :2 3.. 8.: 9mm as“V 805$. 296— .m.m M .m.m M .m.m M 0:8: :oEEoD 5:532: E u s 33-2 E n 5 38.8 a n 5 33-8 .moomlvoom E 53:22 83383 .6ng 05 E 53% $63.5 05 mEB: 093 “333 Ezfimfizfiagémcefi 05 E53, 5%: :m 3323 860% 25 («o dmwv fiche Eavgm :5 32.580 86on £023 3 9:8: mo €353 momofiscofi 8203a :82 .w.m 2an 137 Amma. 5:280 “0:: _0m0mmv 005:? we mum—m5: 330:0 0:334:me 05 .23 60830.8 0:03 3:30: m_0>0_ $3305: .232 5:880 2:: .035: 2E8: :Omtaafioo 032:8 0.5034385: 9: .c A 3 2:80:30 8: 0.8 .530— 088 05 5:5 0:: 083 05 :o 0020?. I E? o: 0.: a: g o: a... 5333 838-3%; 82 E E a: 2 2 _ a: 325:3 3.3-5:; 3:: E: v.2 o: 3. q: 3. 30:23:30 28:83: 8:: E: «.2 a: 3 a: 3: :35 :83 08.: E .E 2; .EN 0.: .3. 223 be; $2: 3: 3m: 3 E :2 an". 323:: 385.323 3:: 2: «.2 o: 3. 0.: 3. 5:52: 8585.323 :5: E E 3 3 o: a... 50> R: .o o: a... E: 2: _.: :2 5:29 §§m :8: E .E :2 as». :.: .2: $32: 8::: 3o: 0.: :0... o: :3. S; .2: *fiofiew 3235-08: owed od cc... 3; «MSN v.2 «Qmm 1006030003 c0um0né0m as: Q: 20.: 3: :EE _.: .3: .323 8:93: :3: 3; :23 2; :E: _.: .2: 2225: 03305-3: 82 E: SN :2 S: o: 3. :23: £2 3; man 2: EN _.: 2: new: 2:: E: .o E 3: a: 0.8 0.: mm 38:82,: 832:: :02 Q: 3:: E .2: o: 3:: 2555 83 o: a... o: a... Q: a... 28:: E2532 __0>0_ .m.m M .m.m M .m.m M 0:8: :oEEoU £332: 3 u 5 3%.2 2: n 5 :83: a u a :83: goo: .3 03:9 138 139 22 2o 2.. 25 25 2mm 23 3.85 8.5:: w 85 2o 2... 2o 2.2: 3 2... *anuooa 3582-82 255 25 2.2: 3 2.2: 23 2.3 *8? 33-3: 82 2% 2cm a? man 3: 22m 22525 3805.3: 2:5 2o 2.. man man 23 2mm :55... 85 22 25 2.. 2o 2.. 2mm 23 5585583 832: 22: 2o 22: 20 22: 25 22: 5555 33 2o 2.. mm 0mm 25 2.. 5.55: 52202 on; 5.5 2... mm 2% 2o 2... Essa 2:252 :2 2o 2.. 3m 5: 2mm 23 .235 5:5: 2.9 3 2.. 2o 2.. 3: 2% 52.55 855895200 N23 25 2... 2o 2... 2mm 2% .2825 BE 533m E _ .o 2% 2% 2o 2.. 25 2.. 8.5% 533m 536 9% c.em ed a... 9mm c.m~ How—83603 bfiéxgom 22 2o 25 25 2.. 2mm 23 “85:32.02 8558 $2 2o 2.. «:2 2mm 2o 2.. 5333 3295536 83 2o 22: mm 55 2mm 23 8333 5056585235 ooo.o 25 2.2: 3m :23 2o 2... .5333 336285-585 83 2% 2% 2m man 3: 25m 83020 3558-585 mad 3 2.2: 3 2... 2o 2... £552 $055. 3.2 3 2.. 2o 2.. 2mm 23 320 50:05. 22,3 .m.m M .m.m M .m.m m 085: 208800 5:538: a u 5 582-2: 2 u 5 323: a. u 5 302-8 .mooNIvoom E §wEo=2 diamcom Saab 05 5 530m wfiwoen 05 mats—u on»: “832 uthRGQémasimuV 05 23:3 comma S @9333 86on EB mo émdv 38.8 32:86 was 32:58 2:89. £023 3 3502 mo “:08on 86538.5 8.38% 5.32 dd 2%... :32 5.2280 Ba 33% uses: 555558 222:8 m___a3-_8_m=£5 8: .o A 3 East? 8: 05 5:2 2:3 2: £3, 2.: 058 2: :o 835?. A33 5:830 28 _qumV oogg> mo 29:28 33-25 m_=m3-_9_m=.£ 05 2:3 35:28 203 28.5%: 592 b23305: ovmd ed 9.: od c6 33. cam :25: woo? m mmd o.o a... o.o :6 9mm c.m~ 38.:an 8825-333 m mm .o od :6 ed :6 93 Q3 gowns“: :2:QO :32 .m.m M .m.m M .m.m M 0:8: :oEEoU 5:538: a n 5 :83: 2 n 5 3.2-8 9V n 5 322-8 3:08 dd 035,—. 140 141 2...... 2.. 2...: 2. . 22. 2.. 2...: 255.... h E... 2.. 2.. m... 2.. 2.. 2.. 2.:... 52.82 Rm... 2.. 2.. 2m. 23 2.. 2.. .83... 52.52 .62.. EN 2.: .2 23 2.: 2.: .225. 222.82 5 E... 2.. 2.. m... m... 2.. 2.. 2% wages. :2. E. S: 2.. 2.. 2.. 2.. 33.5., 2.....me 2:... SN 2:: 2... 5.. 2.. 2...: 2.:... 2.50.. 8:... 2.. 2.. 2.: 23 2.. 2.. .0323 Eggnogow :2... 2.. 2.. 2: 5: 2.. 2.. .82.. 2.3 523m. 8.... 2mm 2.: 2:. 2.: 2.. 2.. 838.. 538m $2.. 3. E: m... m... 2.. 2.. 5.0389: E22238 2.2. 2.. 2.. 2.. 2.. 2.. 2.. 22:25.2...2 855.... 2.2.. 3. .5: 2:. 2..: 2.. .22: .3323 32...-5585 E... 2.. 2.. m... m... 2.. 2.. 2538 832.532.. 3:... SN 2:: 2:. 2.: 2.. 2.. 38:. 2...... S. 2.: 2: 2.:: 22: .22: .5383 §a-.§8§-._§m $2.. 2.. 2... 2... 2..: 2.: 2..: .5323 83-8.8522... 2.2.. 2.. 2... 2:. 2.3 2.: 2..: .8302”. 8223-28... E... 2.. 2.. 2.. m... 2.. 2.. 5:23 323-..§¢_8_m :2... 2.. 2.. 2: E: 2.. 2.. 55%... 505:2 3...... EN 2:: 2:. 2.: 2.. 2.. a...“V 5.5.5.. :26— .m.m M .m.m M .m.m M 25: 208800 £338.. A: u 5 3.2.2 2.. n 5 5.-.: mm H 5 202.: .moomlvoom E 8.3532 6:65:02 .233 on. 5 :83: mavens 05 mag... 25 .83.... 33.85.6233 £82 on... E53, 5%.. E 3225.: :30on EB c... 32m. Echo 23.5.: 9... 9.858.. 8.8% 523 2. 3502 no .5803 8.05:3... 8203.. 5.22 62m 03...... .23.... =m=8mmu E... .536. 2.5.9: 55.85... 032:... 5.3348396: 8. .o A 8 East... .2. o... .032 2.8: on. 5.3 2.: 25: o... so 823?. Amwa— =m=880 E... 33%. vocatg mo 5:35.... $3-2... $334832! 2.. 5.3 3.23.... 9.03 3.8.... :35— b23395. 142 32.. 2.. 2... m. : 2.. 2.. 2.. 8.23.8... 88.323 3 .2.. 2.. 2.. m... 2.. 2.. 2.. 85.58.55. 8.85.83 32.. 23 .2..: 2.. .2.. 2..: .2..: 838.53 .52.. 2:. 2... m... m... 2.. 2.. 388... 8.85.8.5? 3 .2.. 2.. 2.. m... 2.. 2.. 2.. 8852. 8:85-353 E... 2.. .2.. 2... .2... 2.. .2.. .38.. and _AN m.mm Wm Md od e... 3me 5.8% 2.2.. 2:. 2.. 2:. 23 2.. 2.. 88... 8...... 22.. 2.. 2.. m... m... 2.. 2.. .338... 8.85-83. 82.. 2.. 2... 2:. 23 2.. 2.. 8.8.803 8828.. .8... 2.. .25... 3.: .23 2.. .2.. .853 88.8.. :8... 2:. .23 2:. .23 2.. .2.. .8858 8.85-8... :2. 2.. 2.. 2.. 2.. 2.. 2.. 83... 2:... 2.. 2.. 2.. 2.. 2.. 2.. 5...... 8.. $2.. 2.. .23 ..:. ......: 2.. .2.. ...§.8..8..3 8.8.... .65. .m.m M .m.m M .m.m M 08.... :oEEoU 3.....38. G u 5 8.-.: a. n 5 .83.: .3 u 5 8.-.: 3.5... .....~ 28: Community characteristics Nesting preference Cavity-nesting birds in the study area included the black-capped Chickadee, brown creeper, downy woodpecker, hairy woodpecker, northern flicker, pileated woodpecker, red-breasted nuthatch, red-headed woodpecker, white-breasted nuthatch, winter wren, and yellow—bellied sapsucker. Several notable differences in numbers of cavity-nesting birds and non-cavity—nesting birds were observed among age classes and habitat types. The average number of cavity-nesters per observation was significantly less (p < 0.10) than the average number of non-cavity—nesters in all age classes and habitat types, except for aspen stands 2 70 years in the AQV habitat type (Figure 2.2). The average number of cavity-nesters, however, did not differ among habitat types within the same age class. The average number of non-cavity-nesting birds did not differ among habitat types represented in the 20-year age class; however significant differences among habitat types in other age classes were evident. For instance, aspen in the 50-year age class within the ATD habitat type had fewer (i = 4.67, S.E. = 0.34) non-cavity-nesting birds per observation than stands in the AVO (i = 7.17, S.E. = 0.37) and TM (i = 6.18, S.E. = 0.48) habitat types. In aspen stands 2 70 years old, differences in the average number of non-cavity-nesting birds were only significant between the AVO and AOC habitat types. There were fewer (ii = 5.33, S.E. = 0.21) non-cavity-nesters in the AVO habitat type than in the AOC habitat type (i = 8.33, S.E. = 1.33). 143 A. _S g g Icavity ‘CI ' g "BT‘EVLW u 0 §’ 5 AQV2 ATDQ Av02 QAE2 TM2 I-labltattypos 12 B. s :3 1o , E _§ 8 O 6 , fl lcavity i Dnon-cavity' at 4 W o a: 8 2 ' > < 0 . ATm AV05 Habitattypes 12 C. = o 'g 10 Z 3 a , 7 .0 . O 6 lcavnty .. W , g. anon-cavity at 4 , m, d) a E 2 . . - . ,,, ‘ 0 > ' 1 1 < or . . Habltat types Figure 2.2. Number of cavity-nesting and non-cavity-nesting birds in ZO-year (A), 50- year (B), and 70-year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola- 0smorhiza; TM = Tsuga-Maianthemum. 144 MigLatorv status The number of short-distance migrants was significantly lower (p < 0.10) than the number of year-round residents and neotropical migrants per observation in aspen within the 50-year and 70-year age classes and among habitat types represented in those age classes (except for aspen in the 50-year age class in the ATD habitat type). In the 20-year age class, the differences were only significant when compared to the number of neotropical migrants observed in all habitat types represented (Figure 2.3). Within the ATD habitat type, the average number of year-round residents per observation was highest (3Z= 2.75, S.E. = 0.95) in the 20-year age class, and lowest in the 50-year age class (i = 0.67, S.E. = 0.33; Figure 2.3). These differences, however, were not significant. The average number of short-distance migrants observed per observation was similar between 20-year and 50-year age classes within the ATD habitat type, but none were observed in the 70—year age class. The average number of neotropical migrants increased with age in the ATD habitat type (20-year: if = 3.00, S.E. = 0.71; 50- year: if = 4.00, S.E.= O; 70-year: 32 = 5.50, S.E.= 0.50), and were significantly different (p < 0.10) between the 50- and 70-year age classes (Figure 2.3). Within the AVO habitat type, the average number of year-round resident birds increased with age, but was not significantly different between the 50-year and 70-year age classes (Figure 2.3). The numbers of short-distance and neotropical migrants stayed relatively consistent among all age classes within the AVO habitat type. Within the TM habitat type, the average number of year-round resident birds, short-distance migrants, and neotropical migrants stayed relatively consistent with 145 increasing age (Figure 2.3). No differences were observed in the number of birds observed within each migration class across age classes. Sensitive species Four species were observed in the study that were defined as sensitive based on the Audubon Society’s Watch List and species with the greatest population declines in Michigan over the past 40 years (Brewer et al. 1991, Audubon Society 2002). Those species included the golden-winged warbler, pine siskin, red-headed woodpecker, and wood thrush. There were no statistical differences (p > 0.10) in the numbers of sensitive species observed among age classes and habitat types; however some trends were observed. Generally, the average number of sensitive species recorded per observation was highest in the 50-year age class (50 year: i = 0.54, S.E. = 0.11; 20-year: i = 0.23, S.E. = 0.09; 70-year: i = 0.11, S.E.= 0.06). On average, a greater number of sensitive species were also observed in the ATD habitat type (i = 0.56, S.E. = 0.18) than either the TM (2 = 0.27, S.E. = 0.09) or AVO (i = 0.35, S.E. = 0.11) habitat types. Among age classes and habitat types, the average number of sensitive species observed per observation was highest in the 50-year aspen age class within the ATD habitat type (if = 1.0, S.E. = 0), followed by 50-year-old aspen within the AVO habitat type (i = 0.50, S.E. = 0.15), 20-year-old aspen in the ATD type (if = 0.50, S.E .= 0.29), and 50-year-old aspen in the TM type (2 = 0.46, S.E. = 0.21). No sensitive species were observed in the 20-year-old age class within the QAE, AVO, and AOC types. Similar numbers of sensitive species were observed in 20-year—old aspen within the AQV (SE = 0.20, S.E. = 0.20) and TM (i = 0.22, S.E. = 0.15) habitat types, and aspen 2 70 years within the TM (i = 0.15, S.E. = 0.10) and AVO (i = 0.17, S.E. = 0.17) habitat types. 146 A. 5 8 'fi 7 Z 6 . ,2, E 5‘ lyear-round ‘ 3 4 - Ushort—distancei o c a. 3_ - ‘Zneotropical J at . W, a 2 a E 1 o > O < B. 8 5 7 'r". E 6 o . Hg 3 5 ' lyear-round i O . 5 4 " Dshort—distance‘, a. - l at 3 . uneotropical _,J 8. 2 g 1 < 0, C. I year-round ‘D short-distance ’U neotropical Average # per observation 0 —I N u # 0| 0) \l a: AQV7 ATD7 AVO7 AOC7 TM7 Habitat types Figure 2.3. Number of year-round resident birds, short-distance migrants, and neotropical migrants in 20-year (A), 50-year (B), and 70-year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus— Vaccinium; ATD = Acer- Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = Tsuga-Maianthemum. 147 Habitat suitabilimfor indicator species Overall habitat suitability scores were based on breeding habitat for all indicator species except the pileated woodpecker. Food was not considered a limiting factor for those species. Therefore, scores of foraging habitat suitability are only reported for the pileated woodpecker. YE}! Habitat suitability for the veery was highest (x = 0.70, S.E. = 0.0) in aspen 2 70 years old within the AVO habitat type (Figure 2.4), and generally was higher than ' suitability across age classes in the TM and ATD habitat types. It was only significantly (p < 0.10) higher, however, compared to habitat suitability of stands in the TM habitat type. Ovenbird Ovenbird habitat suitability was high (2 0.60) in stands within all age classes and habitat types, but was highest (> 0.90) in aspen Z 70 years old within all habitat types assessed (Figure 2.4). The most notable differences in ovenbird habitat suitability occurred between stands in the 20-year age class within the TM ( x = 0.93, S.E. = 0.04) and ATD (x = 0.76, S.E. = 0.06) habitat types (p = 0.014); and between aspen stands in the 50-year age class in the AVO habitat type ( x = 0.60, S.E. = 0.11) and the TM (x = 0.94, S.E. = 0.01) and ATD (SE = 0.91, S.E. = 0.02) habitat types (p < 0.10). Yellow-rumped warbler A Aspen stands 2 70 years old within the TM habitat type provided better (p < 0.10) habitat conditions for the yellow-romped warbler (a: = 0.54, S.E. = 0.08) than other age classes and habitat types (Figure 2.4). Habitat suitability was also higher (p < 0.10) in the 148 TM habitat type than in other habitat types across all age classes. Suitability was < 0.15 in all other age classes and habitat types, and was 0 in aspen stands in the 20-year and 70- year age classes. American redstart Habitat suitability for the American redstart was generally higher in aspen stands within the hardwood (i.e., ATD and AVO) habitat types than in the TM habitat type (Figure 2.4). Also, older (> 50—year-old) stands within the ATD and AVO habitat types provided better habitat conditions than stands in the 20-yeaer age class. Differences in American redstart habitat suitability were not significant (p > 0.10) between the ATD and AVO habitat types within the 50-year and 70-year age classes (Figure 2.4). Blaclflapped Chickadee There were no significant differences in black-capped Chickadee habitat suitability when compared among age classes, habitat types, and age classes within habitat types. Habitat suitability was higher on average, however, in aspen > 50 years old, and within the AVO habitat type (Figure 2.4). Pilefied woodpecker Habitat suitability for the pileated woodpecker was low (< 0.17) in all age classes and habitat types. It was 0.0 for all stands in the 20-year age class, all stands within the TM habitat type, 50-year-old stands within the ATD habitat type, and stands _>_ 70 years old within the AVO habitat type (Figure 2.4). Foraging habitat suitability scores for the pileated woodpecker were generally higher than reproductive habitat suitability. Foraging habitat suitability was highest in stands 2 70 years old within all habitat types (ATD: x = 0.56, S.E. = 0.04; TM: 32 = 149 0.55, S.E. = 0.02; AVO: SE = 0.44, S.E. = 0.02). Stands in the 20-year age class within the ATD habitat type also provided moderate foraging habitat for the pileated woodpecker (x = 0.48, S.E. = 0.03). Stands in the 20—year age class within other habitat types, however, did not provide foraging habitat for pileated woodpeckers. Average foraging habitat suitability for aspen stands in the 50-year age class in the AVO and TM habitat types was 0.35 (S.E. = 0.03) and 0.39 (S.E. = 0.04), respectively. Suitability for 50-year-old stands within the ATD habitat type was 0.0. 150 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0 10 A.) Habitat suitabiitity index B.) Habitat suitablltity index 0 01 O C.) Habitat suitabiltity Index oIoo , i W I . x i ‘i , ,. , . _ , , W , , , W , . H . 7, I ' I. 7 i , ,, ,. , i v : i . I t 7 ,n , a ATD El AVO Veery Ovenbird Yellow- American Black- Pileated mmped redstart capped woodpecker warbler chickadee Veery Ovenbird Yellow- American Black- Pileated rumped redstart capped woodpecker warbler chickadee Veery Ovenbird Yellaw- American Black- Pileated mmped redstart capped woodpecker warbler chickadee Figure 2.4. Habitat suitability scores for bird species in 20-year (A), 50-year (B), and 70- year (C) age classes of aspen in different habitat types in the western Upper Peninsula, Michigan during 2004—2005. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer- Quercus- Vaccinium; ATD = Acer-Tsuga-Dtyopteris; AVO = Acer- Viola-Osmorhiza; TM = Tsuga-Maianthemum. 151 Avian communities Groupings of bird species associated with similar aspen age classes or habitat types were evident based on graphs of principal component scores (Figures 2.5——2.1 l) and frequency of occurrence data. For instance, bird species that appeared to be associated with younger (e. g., 20—50-year old) aspen stands within the AVO and ATD habitat types included the black-throated blue warbler, chestnut-sided warbler, eastern wild turkey, and rose-breasted grosbeak (Figure 2.5). Based on vegetation attribute results from Chapter 1, threshold levels for vegetation characteristics associated with this community were > 70% herbaceous cover, < 100 coniferous stems/ha, and shrub height > 2.1 m. Species generally observed in aspen in older aspen (>50 years) within the AVO or ATD habitat types were the American redstart, eastern peewee, golden-winged warbler, red-headed woodpecker, scarlet tanager, and veery (Figure 2.6). Vegetation attributes associated with this community were overstory dbh 13—15 cm, height > 9.5 m, 850—1200 deciduous stems/ha, ,< 350 coniferous stems/ha, and coniferous canopy cover < 7.5%. The blue jay, red-breasted nuthatch, red-eyed vireo, red-headed woodpecker, and white-breasted nuthatch were associated with aspen Z 50 years old regardless of habitat type, although the blue jay, red-eyed vireo, and white-breasted nuthatch were associated more with AVO and ATD habitat types than with the TM habitat type (Figure 2.7). Defining vegetation attributes for this community included shrub height 2 2.4 m, 600—1200 stems deciduous stems/ha, and tree diameter 13—15 cm dbh. The wood thrush and yellow- bellied sapsucker were mostly associated with aspen stands < 50 years old within the TM habitat type, but also occurred in younger stands within the other habitat types (Figure 2.8). It was unclear what defining vegetation attributes were for this community because 152 there was no clear distinction in occurrence of birds among younger habitat types. Birds associated with aspen 2 70 years old within the TM habitat type included the black-and- white warbler, brown creeper, and yellow-rumped warbler (Figure 2.9). The pine siskin also was associated with aspen Z 70 years old, but used stands in the 50-year age class as well. This community was associated with > 10% overstory conifer cover, trees > 12 cm dbh, and > 1000 coniferous stems/ha. Other species associated with the TM habitat type regardless of age include the golden-crowned kinglet, white-throated sparrow, and winter wren (Figure 2.10). Several birds were considered generalists because they were found in all sampled asan age classes and habitat types. These generalist species included the black-capped chickadee, black-throated green warbler, downy woodpecker, hermit thrush, Nashville warbler, northern flicker, and pileated woodpecker (Figure 2.11). Other generalists not include in the figure were the ovenbird, American crow, and raven. 153 .aNNfeEwOéNQS {now u O>< 28 .mimBSDQémaruimuV n DH< .Ezsmsaafiegémsfi u SE. .ameom—z fiamaaom comm: 8883 2: E $93 “852 O>< was QH< 05 E53, comma Eoéfixémlom 53, 9833083 53:8an 533‘ .m.~ ogmfi O>< I Dh< SC. bane—u 88m .89: Ems: 9:33 e960 3053 .30.: bacon mince €83 gen: 550 3058 Bacon: 560 38035: b —. <0 n— bmmcou 83m cordon: N o N- v- £28 8% as: A a mu bicep 88m mzoonoQH See—aoaén xwonme wowmmofiogm Z VOd >83: 23, Ecummm 5333 32955850 I F 6383 osfiéowmofifixofim I g Sea—3%-: Emma: over: _ _ m “825% 08.: 635056-303 (wow u O>< use .uthRDQéwa 518v. u QH< .E:Em£§3~<é%§& n EH .emwfiaz £3383 595 8333 2: E $93 “853 O>< use QH< 05 Eases comma 20-3%-? 53, meow—~68? 5:38:80 §m>< 6N ocswfi O>< ll. GH< EH Emma—ow 88m 30.: Emma: 93.33 H960 30:8 39: bacon mtnoc €003 E500» 858 3053 8.3809 850 msooomfiome _‘ (On— bmmeou 8on 5.380., N O N: V: bicep 88m 30.: mu bigot 88m msosEooDe Zea—«Pram boo> , Swag“ 813m I I Z VOd a , I ... a c a e 5333 wowfiaéoflow . . . . . . 838m Bowmmm III I .... 33¢.on 535:2 ll . 86on E:— Sofiaohéb Ewe: oat _ F m 58.5% ova—He 155 .QNESEmQéBi 18V u O>< can .mtfiomgéwafi «Sow u QH< .Ezsmfifida: Aymara H EH .gwEE—z .23:th “RED 5883 05 5 Emma Eoéaozém 53> 823888 b33888 Swm>< KN oSwE O>< II 95‘ SF 36:3 8on 38.: Emma: canme .825 3058 got >83 mung 3003 :32: .830 32.8 Morgue B>oo gooomfiomt _. (On— bmmaoc 8on Bacon: N O N: V- bane—u 83m _Sohe _ _ m: bacou 8on 98:20er 2.7.—3%.: I T. d 585:: 3535-833 0 51.8383 393:-qu I I V 083 35-qu .. .. . . 7a 535:: @8985-qu II I _. ham 0:5 I. g Sofiaohéb £32. BE _ _ m .838me 02,: 156 .anEonmQéNoS {mow u O>< wad .wthQDQémsmm {Pew n DH< .5330va ”235:?“ Saab 5283 2: E 090 “952 C25 E=ENS§BV<$M§R 05 £53» comma Eoéaoxém ~23 328683 £5888 534a .w.~ oSwE O>< II. GH< EH. 56:3 8on 30,5. Ewmo: page $30 3980 30,: $on $.53 3003 E52: B>oo 30:8 Bacon: H269 msooomntome _‘ (On— buaou 80% 8.380% N O N: V: bmmaow 88m 130,: _ _ m: bmmcov 88m msoscmoomh Ec..3§.e~ I L _‘I .d 3 V 7v “86:33 3239323.» I l I _. fish: coo? I an Eo...au%..¢b _ _ Ems: 85 m SEE—xv 02.: 157 .GNE‘NKQEMQIUNOQ— {mew u O>< :5 £t-o§bQéM§Ki~uV n DH< dawE£E 632:0: :253 50603 05 :m on»: 235:: A25 EnfimfinsaEémgk 05 :62? 2: 6.8% on N :26: 56> 6:063:88 bass—Eco :wm>< dd Rama 8383 8:82-302“; I I :666 0:5 8:88 :32m 5583 8336:?xoflm 3 O>< ll QP< Ems; £56: >80: 663: 30:3 85:3» :38 80338:: 2H >65: :86 68.2. F< was .atfiongéwzfl (wow u 95‘ .fiwEfi—Z dame—Eon 8&5 €883 2: 5 25 233m: :2: Ezsmfitfiuééwzfi 05 GEE» comma .33 323683 $5.888 §m>< .SN 93w?— O>< I Dh< EH. bmmcov 88m 30.: Emma; page 539 3053 130,: 30% muse—V 3003 8500» .650 30:8 homEoUh 8>oo msooomfiome —. (On— bmmcow 8on 8.3509 N O N: V: bmmcov 88m 730‘: _ q m: bang 82m .050oner 35:390.: w. Z VOd :8? H853 388% 3385-833 I I Ewci 3:389:02o0 I 86on E:— 2.73??? :32 BE _ _ mu “Sonic 02.: 6380500633 -000V .1. O>< 0:: £t80§£QéM=£ L000. n QH< 83508230380085 u SE. days—0:)” 60:05:03 8333 E0803 05 :_ 099 0830..— 0090 0330: 0:0 000.88 08 0.322: 5505 :038 505 80008008 385800 83¢. A _.N 0593 O>< II QP< 2H. 30:00 80% 030.: £30: 0::ng 8>00 30:8 030.: 8000 £500 >000? 80D.— :0>00 30:8 8&8“: :0>00 80008821— _‘ (Un— b_m:00 E000 8&8“: N O N: V: >880 808 30.: _ . mubmm:00 E08 80:080e‘ 0.0-..8hén \ . I “3.00-o . 88030003 088:3 _ I W 0-....- A . l _‘u 32: 8582 W... .0 .. u. ... a.“ ,. .. 83:03 05>:me 2 .u m mm 80:0 0:803 I I . u. .u V _M f... u .. .. Z 88030003 3300 .. n. W... ... w a 0333 520 885-0020 II I .1 ... .. 1 _‘ 0000808 0038-805 I M — \inIIha . ”0% F r ‘2: 0.0.5880: £00: 2:: W _ m quoF—dmfi 0953—. 160 Winter resident birds Ten bird species were observed during the winter bird surveys and included the American crow, black-capped chickadee, blue jay, downy and hairy woodpeckers, pileated woodpecker, raven, red-breasted nuthatch, ruffed grouse, and white-breasted nuthatch. Species richness was similar in all age classes, with the only difference being absence of ruffed grouse recorded in stands within the 70-year age class. Very few differences were evident in frequency of occurrence of winter resident bird species among aspen age classes and habitat types. When comparing among age classes, the black-capped chickadee, pileated woodpecker, and raven exhibited differences in occurrence, however the differences were barely significant at the a = 0.10 level (p = 0.974, 0.091, and 0.099, respectively; Table 2.11). Within the 20-year age class, observations were made in the AQV (n = 3), TM (11 = 6), ATD (n = 6), AVO (n = 5), and AOC (n = 3). When comparing across habitat types in this age class, the American crow only was found in the ATD habitat type (p = 0.053), the black-capped chickadee was observed more frequently in the AQV, AOC, and ATD habitat types (p = 0.041) and the downy or hairy woodpecker was observed most frequently in the AQV habitat type (p = 0.023; Table 2.12). Within the 50-year age class, the only samples occurred in the TM (n = 9) and AVO (n = 15) habitat types. Of these 2 habitat types, the blue jay was more prevalent (p = 0.031) in the TM habitat type (Table 2.13). Within the 70-year age class, samples were taken in the AQV (n = 7), TM (n = 11), and AVO (n = 6) habitat types. The only difference among these habitat types occurred in the TM habitat type, with higher frequency of occurrence of the raven (p = 0.048) and red-breasted nuthatch (p = 0.058; Table 2.14). 161 .233— :m=8mwu can _awflfl oumtflm acmEaEOo 29:3: m=_a>?-_mxm_:v: Q: .o A 8 Engage 8: 0.8 .632 2:3 05 5m? 0:: 2:8 05 no mo=_m>.. Ammo— 5:380 use. _owflmv 3:35; mo flab—En .3365 £33-?me 05 53> 33328 203 3:32 232 b=Emno$~ 3.2 M: n: 3 N... S 4.: 535:: 3305-223 Rod 3 3. .3 a... 3. m... anew Baum a; no ”.3 ed 33 3. 3. €353 veneefim god 3 3.2 E: 3.3 2: even 23% Rod 2: .93 v.3 aNém vd «man 18.8383 BEBE a _ .c 2: an». 3: 3m 2: .23 300383 beings $2 3 me we 3 ed a... 385 :55 $3 3 n: 2 ”.2 3: 2m 3 35 god 2: an: 2: anwm .1: new .8325 Basgga 9.3 we a: 3 «.3 S 9.2 38° 30.55. :33 .m.m M .m.m M .m.m M 0:8: :oEEoU £335 Manager? Gmuasohfi “managers. .ooomlmoom E 535:2 Essen .5an 2: E EQBEIEBnoE BEE» 05 wfihfi momma—o owe :33 m 5 3.3.56 86on EB mo A.m.mv £85 @3283 93 95.568 86on £023 “a $53 90 E883 36:25on 328% 502 A _N 2an I62 .239 5:830 EB 39%. 32:5 somtaano 032::— m=_m>»-_mxm£ 2: .o A 3 EBabE 8: 08 .632 2:3 2: 53» 0:: 08mm 05 so 833?. A32 aux—8&0 Ea _owummv 35:?» no £3.88 $3.20 m___m3-_8_mav_ 05 £3, Bum—3.8 203 notoqfi 226. 5:538“; £3 a? man 98 :2 5: S: 62 5: ed a... 5352. 85852.3, 32 ed 3. ed a... S: 5: od 3. ed a... anew Baum a; 3 3. 9o 3. ed a... 3 a... 2m 3m €353 3.323% $3 2m mam 3m 3:. 2m mam 2N mam 2H an». 53* mm; ed a... SK :2 ed a... tam 3m 3m 55 38%83 832: SS ed 3x: 98 3; 2m 55 EN mam mam man 38383 $3338 Rod 3 =3. o.o 3:. EN 23% EN 33¢ od :33 .385 :3 mm 25.8 od :3. «mm 2:3 52 2.5.3 3 .3.: $03020 89.8285 mod od .3. 9o .3. EN 3% 9o .3. ed «a... .355 30:82 :22 .m.m m new m .m.m m .m.m m mm m 253388 £3.35 qua and 8M3 one find 89.. 83. 8.5. ms: ~>o< mmflo own 3908 E 36823“ $99 “953* .Exaémszcb-cuEonmQLmuv. u UO< .cNELoEuQéNoS 18V u O>< .wtmicfiQéwaflimuV u 9.3.. .5:Em$t3§<$%§& u 3:. €559.93; -waugmzmimuv. u >O< .ooomlmoom E :dwEBE 63223 8&3 05 E BufiElEEpoB 52:3 05 waist womb “SEE “subway E53» 320 um“ Boxéw 05 E comma E @3256 38on 95 mo A.m.mv Echo 22:86 98 90568 86on :03? 8 35a mo E853 momoaozcoa 838% 582 .N _ .N 2an I63 .232 5:2me 28 _mwflfl osmzsm comEaES 0.22:8 mink/4330: Q: .o A 3 “BEE—u 8: Pa .832 2:8 05 5?» on: 088 2: :o mos—Sf. Ame $23.30 28 _owo_mv 35:5, mo mix—£8 $3-28 £33-3me 05 5E van—=28 20>» 3288 £32 b__Emno£_ $3 3 3 od 3. £323: Bfioséfi, 33 E 3 od 3. anew out; was 2: 3." 3: man €352 Basséom owed a: 5; 0.: 3m 33.x w So 02 0mm 9: 3m 38383 BEE End 1: SE 52 man Sfiauooa bainaom aomd We 56 <2 fl: 830.8 :3on ES 3 a3 a: .3.: *3. 02m 32 we 3m 52 $5 8325 8&8285 $3 3 mi 3_ 3m 320 505:3 226— .m.m m .m.m m 08m: :oEEoU E535 Ens mo>< 3 mmflo oww Homém E 36883“ $93 3:an .QNELoEmQéBE -LmuV u O>< .Ezfimfizumagémsfi n EH .coomlmoom E gwfiomg ,flsmcaom Saab 05 E anmElgpomv SEE 2: mega womb “SEE Enact? GEE» mmflo own Soxém 05 E :38 5 wozogm 86on 25 mo A.m.mv €on @83de 98 82.880 86on £023 a 358 «c 68.83 momocosvot 8303.“ :82 .m _ .N 2an 164 .533.— 5=8m~0 Ea _awommfl ocmzsm :oflamfioo oat—:5 £=~34~me5 82V A 3 ERNEu 8: 2a .532 2:8 05 fits on: 0E3 of co mo:_m>.. .932 5:880 98 _owo_mv 35:? mo mix—SB $3-05 £33-?me 2: 5:3 38323 803 3:32 £32 bzfimnoik £3 3 3. 3 2 3: 3a 5.25: Emmoség was od .3. 5.2 may ed .3. £332 Busséom ”35 ed .3. N2 “van 3 .3. .55m ammo a? an». 2; «:2 3: 3a flanges BEE mad o.o a... :2 mfin NON “Sm Execute?» EQQMEOQ 33 ed 9... NS 3: ed a... 388 :3on move ed a... 3 2 2: 33 3 02m 23 mm 55 3. mém 9o 3:: 8325 8&8285 33 ed 3. NS 2: mi «:2 320 505:8 296— .m.m M .m.m M .m.m M 0:82 coEEoU 3:535 and :lTna fin B>< EB 59* mmflo own a??? E 685332 womb “SEE; .uNELoEwQ-£oS .meV H O>< ,Ezfimfizamagémas .I. 3:. €35.83; -wzugmzmimuw u >O< .coomlmoom E Swwfioaz 4:35:05 895 05 E EoEEIbaBnomv H853 2: mats—V womb 55m: Baobau 55:» $20 own BPIK 05 E comma E 3323 36on 33 mo A.m.mv Echo 2.35% was @2380 830% 32:? “a 3509 «o E853 meteosaot 82035 502 .Ed 2an 165 DISCUSSION Vegetation structural differences among age classes and habitat types Although there were differences in vegetation characteristics among aspen age classes and habitat types, there was much overlap in terms of vegetation structural attributes (i.e., 41% of the vegetation variability was explained with the first 3 principal components). This overlap likely was evident because age and habitat types occur on a continuum, but have boundaries that are somewhat arbitrarily defined. For instance, age class designations were 20—29, 50-59, and 2 70 years old. The 20— and 50-year age classes contained stands that may have been 9 years apart in age. The oldest age class contained stands 70—93 years old. Thus, there is inherent variability between the youngest and oldest stands within an age class that may account for the wider distributions in ranges of vegetation conditions than if the classes included a more narrow range of ages. Likewise, vegetation conditions within a habitat type vary on an ecological gradient defined by geological conditions (e. g., soil and topography). Habitat type boundaries are defined to represent a “relatively narrow segment of environmental variation” (Kotar and Burger 2000:4). However, spatial heterogeneity of soil conditions within a spatial extent defining the boundaries of a particular habitat type could create patches that increase variability in structural characteristics of that habitat type. Nevertheless, despite the overlap in variability of vegetation conditions among aspen age classes and habitat types, results of the principal components analysis indicated that differences in vegetation structure among age classes and habitat types were still 166 evident (Figure 2.1). The first principal component generally described the variation among habitat types, while the second principal component described variation among age classes. The third principal component characterized differences in understory structure (e.g., horizontal cover defined by basal area, shrub density, and height) and overstory snag characteristics. The most defining characteristic of the TM habitat type was the presence of conifers (i.e., spruce and fir), especially in older age classes due to the dry sandy to sandy-loam soils, which promote spruce and fir growth (Coffman et al. 1980). The most defining characteristic of the AVO habitat type was prevalence of a dense shrub layer, which persists because of the high nutrient content in the loamy soils of the AVO type (Coffman et al. 1980). The ATD habitat type had characteristics in the middle of the gradient. Older age classes had larger, taller trees, and younger age classes had higher total stem densities and densities of deciduous trees. Older (e. g., 2 70 years) age classes of aspen within the TM habitat type were more distinguishable in structure than older stands within other age classes. Aspen communities in the 20-year age class within the TM habitat type were more similar to structure of younger stands within the ATD and AVO habitat type. This may have implications for management in retaining conifers in stands within the TM habitat type to increase diversity of structure within aspen in the 20-year age class. Avian species associations with aspen age classes and habitat types The vegetation attributes that explained differences in aspen age classes and habitat types also explained the distribution of birds observed in different age classes and habitat types (Table 2.4). For example, species with stronger associations to 20—50-year 167 old aspen without the presence of conifers included the black-throated blue warbler, chestnut-sided warbler, eastern wild turkey, and rose-breasted grosbeak. These species frequently occurred in older age classes within the AVO and ATD habitat types (Figure 2.5), but the black-throated blue warbler and chestnut-sided warbler were associated more with the AVO habitat type. These results are consistent with habitat requirements found in the literature that suggest common habitat components for all these species are mature mesic deciduous forests with closed canopy or other overhead structure for nesting cover (Brewer et al. 1991), which is characteristic of aspen 2 50 years old within the ATD and AVO habitat types. The black-throated blue warbler and chestnut-sided warblers also require dense shrub cover for nesting (Walkinshaw and Dyer 1953), which likely explains their stronger association with the AVO habitat type (see Chapter 1). Species with strong associations with aspen communities 2 50 years old included the American redstart, eastern peewee, golden-winged warbler, northern flicker, scarlet tanager, and veery (Figure 2.6). These species prefer to nest in mature dry to mesic forests containing hardwood species (e.g., maple or beech) with a dense understory (Brewer et al. 1991, Hobson and Bayne 2000b). The golden-winged warbler prefers to nest in early successional vegetation types such as old fields, but also prefers to nest in areas with a scrubby, well developed understory (Confer and Kapp 1981), which may explain their occurrence in the AVO habitat type. Researchers suggest that the blue jay, red-headed woodpecker, and white-breasted nuthatch prefers to nest in mature dry to mesic hardwood forests (Hagan and Meehan 2002, Rodewald et al. 2005), which explains their association with older (2 50 years old) aspen stands in the AVO and ATD habitat types (Figure 2.7). These vegetation types 168 provided snags for cavity-nesters, and relatively less herbaceous cover, which is important for nesting success of the red-eyed vireo (Siepielski et al. 2001). Avian species found in this study to be associated with aspen communities _>_ 70 years old within the TM habitat type included the black-and-white warbler, brown creeper, pine siskin, and yellow-rumped warbler (Figure 2.8). Researchers suggest that conifers (e. g., spruce and fir species) are important components of communities providing habitat for the pine siskin and yellow-rumped warblers for nesting and foraging (Brewer et al. 1991, Hagan et al. 1997, Hanaburgh 2001). Black-and-white warblers have also been associated with mixed aspen and conifer forests throughout its breeding range (Paszkowski et al. 2005), but have shown flexibility using different vegetation types across its range (Dettmers et a1. 2002). Brown creepers have been found to nest behind slabs of loose bark associated with older or dead trees in conifer or mixed forests (Franzreb 1985). Species associated with the TM habitat type regardless of age included the golden-crowned kinglet, red-breasted nuthatch, white-throated sparrow, and winter wren (Figure 2.9). These bird species also appear to associate with the conifer component within the TM habitat type (Pettingill 1974, Brewer et al. 1991). Mature (i.e., Z 80 years old) conifer forests containing spruce and fir provide the best habitat conditions for the golden-crowned kinglet and winter wren (Titterington et al. 1979), however, forests with 20—50% spruce and fir canopy cover can provide at least marginal habitat conditions for the golden-crowned kinglet and winter wren (Salt and Salt 1976, Titterington et al. 1979). The white-throated sparrow and red-breasted nuthatch have also been found in other studies to have higher abundances in mixed aspen forests (Hobson and Bayne 2000b). 169 The wood thrush and yellow-bellied sapsucker appeared to be associated with younger stands within the TM habitat type (Figure 2.10), but these results were not consistent with the literature. Other researchers describe that the wood thrush prefers to nest near streams in wet and mesic deciduous forests with a dense understory (Bertin 1977). Yellow-bellied sapsuckers showed no preference for conifers or deciduous forests for nesting, but preferred birch trees for foraging (Kilham 1964). The association of these birds to young stands within the TM habitat type in this study was likely due to low sample sizes (n = 6 and 5 for the wood thrush and yellow-bellied sapsucker, respectively), or they were possibly using marginal habitat. Avian species observed in all age classes and habitat types were the black-capped chickadee, black-throated green warbler, downy and hairy woodpeckers, hermit thrush, Nashville wabler, ovenbird, and pileated woodpecker (Figure 2.11). Mature forests typically provide better nesting habitat requirements for cavity-nesting species (e. g., black-capped chickadee, woodpeckers) because of increased snag availability (Schroeder 1983a, Schroeder 1983b); however chickadees and woodpeckers were recorded foraging in younger aspen age classes on snags smaller (e. g., S 10 cm) than those required for nesting (Schroeder 1983b). Generally, their prevalence was higher in older age classes; thus they had a stronger association with aspen 2 50 years for nesting requirements (Table 2.3, Figure 2.2). The black-throated green warbler, hermit thrush, Nashville warbler, and ovenbird appeared to show no difference in occurrence for aspen Z 20 years old within the TM, ATD, and AVO habitat types. This is not to say, however, that these species are habitat generalists. They may just show no difference in prevalence among different aged aspen communities within those habitat types, but may, in fact, prefer 170 W". M - vegetation conditions within other habitat types. For instance, the Nashville warbler and hermit thrush have been documented to co-occur broadly and nest in dense ground cover in mesic mixed or dry coniferous forests (Evers 1991). The relatively higher frequency of occurrence of those birds in the dry AQV habitat type supports this habitat association (Table 2.5). The black-throated-green warbler and ovenbird, however, do appear to be generalists and have been documented elsewhere to occur in a wide variety of vegetation types, as well (Brewer et al. 1991, Hanaburgh 2001). Three avian species that were observed in the study were not put into groups associated with aspen age classes and habitat types because they were only observed once. These included the brown-headed cowbird, magnolia warbler, and pine warbler. The mourning dove, American robin, American crow, and raven were omitted because they tend to be abundant throughout their ranges in Michigan and are not of management concern. The northern parula was omitted because although it was only recorded in aspen in the 50-year age class within the AVO habitat type, it prefers to nest in wet coniferous areas or near hardwoods with hemlock or fir where the hanging Usnea lichen, which is important for nest construction, typically grows (Pettingill 1974). Thus, the northern parula is not strongly associated with the AVO habitat type, but more likely was observed because the sampling point was adjacent to a wetter area. The distribution of birds grouped by migratory status was slightly higher for neotropical migrant species in all age classes and habitat types, but was not different among age classes and habitat types (Figure 2.3). This suggests that neotropical migrants as a group may not exhibit associations with forest characteristics that reflect different aspen age classes and habitat types. Year-round resident species were more prevalent in 171 older age classes regardless of habitat type (Figure 2.3). This result is likely because 5 of the 8 birds classified as year-round residents were also cavity nesters, which were more prevalent in older age classes. Hanaburgh (2001) found similar results for northern hardwood stands in the Upper Peninsula. Species classified as short-distance migrants occurred more frequently in the 50-year age class than in other age classes, but prevalence did not differ among age classes within habitat types (Figure 2.3). Similarly, the number of sensitive species recorded at each observation was highest in the 50-year age class and specifically within the ATD habitat type. Therefore, the 50-year aspen age class (especially within the ATD habitat type) is providing features important for short-distance migrants and species of conservation concern. This has important management implications for retaining 50-year-old aspen in the landscape, especially within the ATD habitat type. Currently, asan in the 50-year age class is the least available (3.11%) in the landscape (Doepker et al. 2002), and the prevalence of the ATD habitat type on state lands in the western Upper Peninsula is lower than other habitat types (e. g., AVO, TM, AOC, AQV). F inc-filter assessment of habitat suitability It was evident that different avian species were associated with a specific aspen age class and habitat type, however, individual species may be associated with a particular vegetation type for different ecological reasons. As part of a fine-filter assessment of aspen ecosystems, I selected 6 species a priori, evaluated habitat suitability, and compared it with frequency of occurrence to investigate the utility of those species as indicators of specific ecological conditions that distinguish different age 172 classes and habitat types. The veery, yellow-rumped warbler, and American redstart were good indicators of specific vegetation conditions associated with age classes and habitat types. For example, frequency of occurrence and habitat suitability for the veery was highest in the AVO habitat type (Figure 2.4) because of the relatively thicker shrub layer (> 4,000 shrub stems/ha) than within other habitat types. Yellow-ramped warbler habitat suitability and frequency of occurrence was highest within the TM habitat type; specifically, stands 2 70 years old with greater conifer cover (> 10%) and stem density (> 1000 stems/ha; see Chapter 1, Table 1.14). Habitat suitability for the American redstart was highest in the AVO habitat type because it supports higher shrub cover for nesting than other habitat types. Habitat suitability for the American redstart was also relatively high (0.80) in aspen in the 20- year age class within the TM habitat type due to higher deciduous stem densities (> 1000 stems/ha). American redstarts, however, were not observed in that vegetation type. They were only observed in older stands (> 50 years) within the ATD and AVO habitat type. Another variable in the habitat suitability model was percent overstory conifer cover, which was lowest in young vegetation types within the TM habitat type and within the ATD and AVO habitat types. Therefore, presence of American redstarts in an area likely indicates low (< 40%) conifer cover and relatively higher (> 650 stems/ha) stem densities (Table 2.2). The ovenbird was not a good indicator of specific vegetation conditions and was not useful to distinguish characteristics among the aspen vegetation types within the AVO, TM and ATD habitat types in this study. The ovenbird had high (> 90%) frequency of occurrence in all areas sampled (Tables 2.3—2. 10), and had high habitat 173 suitability scores in all age classes and habitat types represented (Figure 2.4). The black-capped chickadee and pileated woodpeckers were selected as indicators of snag density and size. Habitat suitability for the black-capped chickadee increased with increasing age and was highest in the AVO habitat type in all age classes, and habitat suitability for the pileated woodpecker was low in all age classes and habitat types (Figure 2.4). These habitat suitability trends were not reflected with frequency of occurrence data. For instance, frequency of occurrence of black-capped chickadees was similar in all age classes and habitat types, except in aspen Z 70 years old within the AVO habitat type, in which it was not observed. Frequency of pileated woodpeckers was generally higher in older stands. As noted earlier, occurrence of chickadees and pileated woodpeckers in younger age classes may have been due to foraging activity. Research has shown that reproductive success of cavity nesters is higher in areas with sufficient snag size and density (Schroeder 1983a, Schroeder 1983b), which is characteristic of older age classes. Therefore, the black-capped chickadee and pileated woodpecker might still be used as indicators of snag presence in the asan vegetation type, but productivity (e. g., numbers of nesting pairs) should be measured and related to those characteristics. Although presence of the veery, American redstart, yellow-rumped warbler, black-capped chickadee, and pileated woodpecker can indicate specific ecological conditions, this group of species does not collectively represent the range of vegetation conditions within sampled aspen age classes and habitat types. For example, characteristics of 20—50-year-old stands within the hardwood habitat types (ATD, AVO) are not represented by the selected indicator species. The chestnut-sided warbler might potentially be an indicator of vegetation conditions represented within those vegetation 174 types because the warblers prefer early successional forests with relatively high shrub and stem densities (Richardson and Brauning 1995). Perhaps the white-breasted nuthatch may be an indicator of conditions provided by mature hardwoods (Pravosudov and Grubb 1993). Conclusions Management for multiple objectives (e. g., timber harvest, biodiversity) requires a hierarchical approach to better understand ecological patterns at multiple spatial scales and levels of biological organization (Sampson and Knopf 1982, Haufler et al. 2002) because ecological processes and species selection of habitat conditions occurs hierarchically. That is, managers must first be aware of the ecological potential of landscapes, describe conditions within ecosystems and communities, and then relate ecological conditions to species presence, distribution, and productivity. The following questions can help managers understand landscape potential: what types of ecosystems (e. g. vegetation types) have occurred historically in landscapes and what types of ecosystems are currently present? Are there any ecosystems that are not represented? This coarse-filter approach should identify ecosystems on which management efforts should focus—either to increase or decrease representation for biodiversity conservation. Maintaining a representative array of ecosystems should subsequently maintain habitat for the majority of native species (Noss 1987). Habitat type classifications and quantification of vegetation structure and compositional characteristics of different successional stages can help identify ecological representation. A limitation to only using the coarse—filter approach is that no ecological 175 A/ classification system is comprehensive enough to account for every species (Noss 1987). This is evident when investigating age class and habitat associations of rare or sensitive species (e. g., red-headed woodpecker, golden-winged warbler) or ones with otherwise low frequency of occurrence (e. g., black-and-white warbler [Paszkowski et al. 2004]). Fine-filter assessments become important to derive mechanistic relationships and understand what specific characteristics of ecosystems affect species presence and viability (Marzluff et al. 2002). It would be valuable to develop models (e. g., logistic) to quantify relationships between species presence and aspen age classes and habitat types to assess the predictive ability of forest management models on the effects of wildlife (Marzluff et al. 2002). In this study, however, insufficient sample sizes prevented development of quantitative models, but did still allow determination of avian species associations with aspen age classes, habitat types, and vegetation characteristics. Defining these descriptive associations is an important step in developing future assessment models for forest management. The Michigan Department of Natural Resources has assumed a leadership role in implementing ecosystem management to conserve, protect, and manage Michigan’s natural resources for current public benefits and fiiture opportunities (MDNR 2007). The ecosystem management plan relies, in part, on scientific-based approaches as a foundation for understanding natural resource relationships and sustaining ecological systems (MDNR 2007). As such, the MDNR has appointed management teams for different ecoregions of Michigan to develop strategies to sustain and enhance representative ecosystems within Michigan. The hierarchical approach to ecosystem 176 management described in this project will help the MDNR develop those strategies. For example, the habitat type classification system described in Chapter 1 allows implementation of a coarse-filter approach to ecological assessment, and to understand the availability and spatial distribution of habitat types (e. g., landscape-level performance measures). Results of the vegetation analysis and principal components analysis characterized the range of structural and compositional characteristics as well as potential function (e. g., wildlife habitat) and processes (i.e., successional changes) that can occur within habitat types (i.e., ecosystem-level performance measures). Ecosystem-level performance measures describe vegetation conditions provided by different successional stages within habitat types that can explain wildlife community structure. My results indicate that different aspen age classes and habitat types provide unique vegetation features important to different assemblages of avian communities. Without representation of all aspen age classes within all habitat types, it is evident that certain ecological features would be missing that directly affect presence of specific avian species or communities. Determining the appropriate representation of aspen age classes within habitat types, however, should be based on explicitly stated objectives for wildlife habitat, timber, and other human values. Results of this project determined that species composition of bird community associations with aspen varied among age classes and habitat types. The value of this to forest management lies in the ability to uses the associations to determine likely changes in species presence within forests throughout time or as a result of management activities. As such, the results can be incorporated into planning management activities for avian communities or biodiversity, predicting potential outcomes of management on avian 177 community composition and distribution, and identifying areas within unique feature for management focus. It is through understanding what conditions are provided by different aspen age classes (specifically old aspen _>_ 50 years old), how wildlife respond to these conditions, and incorporating this knowledge in timber management plans to meet multiple objectives that Michigan’s leadership role in ecosystem management will be sustained. 178 CHAPTER 3 POTENTIAL OF ASPEN IN DIFFERENT ASPEN AGE CLASSES AND HABITAT TYPES TO PROVIDE DRUMMING HABITAT FOR RUFFED GROUSE Aspen is an economically valuable resource in Michigan and is also important for providing wildlife habitat components. Specifically, ruffed grouse (a popular upland game bird in Michigan [F rawley 2006]), use of the aspen vegetation type for food, nesting, and drumming has been well documented in the literature (Bump et al. 1947, Svoboda and Gullion 1972). As such, a forest management challenge is when and where to manipulate aspen stands to sustain timber benefits and habitat for grouse. To design appropriate aspen management plans for grouse, it is necessary for managers to understand habitat relationships contributing to grouse production. State agencies (e.g., Michigan Department of Natural Resources [MDNR]) often use several tools to assess grouse population trends and potential productivity, such as spring drumming surveys, but the surveys do not document vegetation characteristics associated with locations of drumming grouse. This information may be critical for understanding how grouse are distributed throughout landscapes during the breeding season, and may provide insights as to what ecological factors are contributing to the observed distribution. Researchers in the Lake States (e.g., Gullion 1977, Kubisiak et al. 1980, McCaffery et a1. 1996) have documented higher densities of drumming grouse in 6—25 year-old aspen stands than in other vegetation types and age classes. Other studies have also reported grouse use of older (2 25 years) aspen and non-aspen (e. g., balsam fir, oak, alder) vegetation types for drumming (Palmer and Bennett 1963, DeStefano and Rusch 179 1984, McCaffery et a1. 1996). In Michigan, Hammil and Moran (1986) observed drumming grouse in lowland conifer vegetation types. Most studies have documented relationships between drumming grouse and broad vegetation categories (e. g., northern hardwoods, young aspen, oak; Cade and Sousa 1985), but few studies have documented how habitat types may influence use by individual drumming grouse. Habitat types are geographic areas with similar ecological characteristics that support the same successional trajectory (Daubenmire 1966). Aspen is an early successional vegetation type occurring in several different habitat types such as those characterized by soil types ranging fi'om poorly-drained loams to well drained sandy soils (Coffman et al. 1980). The structure and composition of aspen varies with habitat type (see Chapter 1), and provides different wildlife habitat components (see Chapter 2). Therefore, it is reasonable to assume that grouse drumming activity may differ among aspen in different habitat types as well as in different age classes. Additionally, few studies have documented the importance of older (> 25 years) age classes of aspen as alternative drumming habitat for grouse (i.e., secondary to preferred 6—25 year-old aspen). To design long-term sustainable timber harvest plans that benefit ruffed grouse, it may be beneficial to assess grouse use of aspen > 25 year old and grouse preference of aspen in different habitat types. Such information may provide insights for understanding landscape-level characteristics affecting grouse drumming distributions, and identifying specific habitat types and age classes of aspen that may be valuable in providing grouse drumming habitat longer than 25 years. 180 OBJECTIVES Specific objectives for this chapter were to 1) Assess ruffed grouse selection of aspen stands in different age classes and habitat types for drumming. 2) Determine the distribution of potential ruffed grouse drumming habitat in the western Upper Peninsula, Michigan based on aspen age class and habitat type. 3) Make recommendations for incorporating ruffed grouse habitat management objectives in aspen management plans. 181 METHODS The Operations Inventory (OI) database for stands in state forests in Baraga, Dickinson, Iron, and Marquette counties was obtained from the MDNR. I identified established 10 routes 5—10 km long for grouse drumming surveys based on stand accessibility in a 4-wheel drive truck and to maximize the number of stands assessed during the drumming survey. On each route, I identified specific GPS coordinates (latitude and longitude) adjacent to aspen stands, which were the designated stopping points for the drumming survey. I determined the age class of each aspen stand on the routes from the 01 database and assigned each stand to a habitat type classification (see Chapterl). All routes collectively contained stands representative of different asan age classes 2 20 years and habitat types. Grouse drumming surveys occurred between May 3—16 in 2005 and April 17—26 in 2006, which corresponded with the peak drumming period in the UP (April 20—May 10). A portion of the survey routes were the same in 2005 and 2006. Surveys were conducted according to MDNR protocol for state ruffed grouse drumming surveys. To minimize effects of weather conditions on drumming activity or audibility, surveys were only conducted when the temperature was between -4—4°C with no precipitation and when wind was < 19 kph. I used the Beaufort wind scale to monitor wind speed, which indicates that when small twigs are in constant motion, wind speed is 13—19 kph (NCAA 2007). Surveys began 30 minutes before sunrise and ended around 8:00 am (Petraborg et al. 1958). At each stOp, I walked at least 15 meters away from the vehicle and listened for 4 minutes, which is the average drumming interval for ruffed grouse (Petraborg et al. 182 1953). I recorded the number of drumming males heard at each stopping point and recorded the compass bearing from the stopping point to the direction where the drumming was heard. I also recorded a qualitative measure of distance: close, medium, or far. “Close” drumming occurred within 60 m of the stopping point, drumming at a “medium” distance occurred 60—130 m from the stopping point. In ArcView 3.2, I placed a 200-m buffer around each point where drumming was heard. Two-hundred meters was the radius of audibility for drumming ruffed grouse in similar vegetation types in Minnesota (Petraborg et al. 1953). Next, I determined in which age class and habitat type drumming occurred by using the Distance/Azimuth Tools Extension for ArcView 3.x (J enness 2005), which establishes a point at a specified distance and directional bearing from a point of origin. I assumed the grouse were drumming in the specific age class and habitat type occurring at the point established with the extension. Within each 200-m radius circle, I determined the proportion of area of young (0- 29 years), medium (3 0—49 years), and old (2 50 years) aspen stands in different habitat types, and the proportion of area in vegetation types other than aspen in different habitat types. These vegetation classes represented those found to be fi'equently used by drumming grouse (i.e., young aspen), and other classes in which drumming has been documented in the Great Lakes region. To evaluate selection of aspen by drumming grouse, I used methods described by Manly et al. (2002) for Design I studies where used, unused, and available resource units (i.e., aspen or other vegetation types of different ages and within different habitat types) were sampled for the entire study area and for the collection of all animals in the study area. Individual animals were not identified. 183 The following resource selection function (Manly et al. 2002) yielded the relative probability of grouse selection for different vegetation classes: \i‘vi 2 0i /1ti , where 0; = the proportion of grouse using the it“ vegetation class it; = the proportion of area available in the ith vegetation class. A selection index = 1.0 indicated no selection, < 1 indicated avoidance, and > 1 indicated selection. The standard error for each Wt was approximated with the following equation (Manly et a1. 2002): 530%) = W1 1/{0 _Oi)/(0iu+)+ 860102015} 9 where o; = the proportion of grouse using the ith vegetation class it.» = the total number of grouse sampled it; = the proportion of area available in the ith vegetation class The standard errors were used to determine approximate confidence intervals for the selection indices [wi i 2(w,,)se(v“vi) with a = 0.10]. The above procedure was conducted by age class, habitat type, and their interaction for data collected in 2005, 2006, and both years combined. A Chi-square test was conducted on the proportion of grouse drumming to determine differences in grouse selection of age classes and habitat types between years (a = 0.05). To understand the spatial distribution of selected grouse drumming habitat, I used ArcMap in ArcGIS and 184 mapped the selection indices for different age classes and habitat types on state land within the study area. 185 RESULTS In 2005, 78 points were surveyed for drumming grouse. The total area surveyed based on the radius of audibility at each survey point was 816.54 ha. We recorded 2 l grouse drumming at 32 points. Two drumming grouse were documented at 5 points, and 3 grouse were recorded at 1 point. In 2006, 85 points were surveyed and area surveyed totaled 875.42 ha. There was no difference in grouse selection of aspen age classes between years (p = 0.73); however, differences were detected in grouse selection of habitat types (p = 0.02) and the interaction between age classes and habitat types (p < 0.001) between years. We recorded 2 1 grouse drumming at 31 points; 2 grouse were documented at 5 points, and 4 grouse were documented at 1 point. Selection by age class The highest proportion of all vegetation classes sampled within the range of audibility was in old aspen (31%) in 2005 (Figure 3.1), and medium-aged aspen (37%) in 2006 (Figure 3.2) and for combined years (34%; Table 3.1). The highest proportion of all vegetation classes sampled within the range of audibility was in and medium-aged aspen (34%; Table 3.1). Interestingly, grouse selected young aspen (w = 1.36 i 0.39) and old aspen ( w = 1.30 :1: 0.29) similarly (Table 3.1). The lower confidence limits of the selection indices for young and old aspen were very close to 1.0, indicating selection for those age classes. There appeared to be no selection or avoidance of medium-aged aspen (w = 0.83 :t 0.25), although the upper confidence limit for combined years was close to 1.0 (1.08; Table 3.1). The upper confidence limit for non-aspen vegetation types was < 1, 186 indicating possible avoidance for non-aspen vegetation (Table 3.1). No grouse were observed drumming in non-aspen vegetation types in 2006. Selection by habitat type Nine habitat types were represented within the 200-m radius of audibility for each sampling point. These included the QAE, AQV, TM, ATD, AVO, AOC, TMC, FMC, and TTS habitat types (Figures 3.1, 3.2). In 2005, 38 ruffed grouse were observed drumming in the AQV, TM, AVO, AOC, and TTS habitat types (Figure 3.1a). In 2006, 40 grouse were observed in the same 5 habitat types, and additionally in the ATD and FMC habitat types (Figure 3.2a). Of the habitat types in which grouse were observed drumming, the highest percent of area sampled was within the AVO habitat type in 2005 (28%), in 2006 (27%), and in combined years (31%), followed by the AQV and AOC habitat types (Figures 3.1, 3.2). The QAE, ATD, TMC, and FMC habitat types were the least available (< 5%) within the sampled area (Figures 3.1, 3.2) Ruffed grouse exhibited selection differences among habitat types for drumming. In 2005 and 2006, grouse showed the strongest selection for areas within the AVO habitat type (V?! = 1.39 d: 0.48, w = 2.37 i 0.92), which was significantly higher than selection indices for the AQV habitat type in both years, and the TTS habitat type in 2006 (Table 3.2). Grouse appeared to select (w >1) for the TM habitat type in 2005, 2006, and combined years based the selection indices (w = 1.33, 1.17, 1.54, respectively). Other habitat types in which the selection indices were > 1 included TTS in 2005 (w = 1.14), AOC in 2006 and combined years (01/ = 1.39 and 1.35, respectively), and FMC in 187 2006 (w = 1.38; Table 3.2). Grouse avoided the AQV habitat type for drumming (w = 0.54 :L- 0.37 [2005], w = 0.17 i 0.29 [2006], w = 0.41 rt 0.23 [combined years]). Avoidance for the TTS habitat type was only evident in 2006 (Table 3.2). No selection was observed within the AOC habitat type in 2005 ( w = 0.83 :1: 0.46), and the ATD habitat type in 2006 (w = 0.84 i 1.36) and for combined years (w = 0.64 :t 0.82). Selection among age classes and habitat types Although ruffed grouse use of different aspen age classes and habitat types differed between years (p = 0.02), in both years, it was evident that grouse exhibited similar trends in selection of specific age classes and habitat types for drumming. General trends included selection of aspen > 50 years and < 30 years old within habitat types characterized by mesic soils (i.e., AVO, AOC, TM). Medium-aged aspen was typically neither selected nor avoided by drumming grouse in all habitat types. Non- aspen vegetation types and aspen on habitat types characterized by xeric or hydric soils were generally avoided. In 2005, grouse appeared to drum in old aspen > 50 years more than expected, given its availability in the landscape (Figure 3.1, Table 3.3). Grouse selected aspen Z 50 years old within the AVO habitat type more than other age classes and habitat types (Table 3.3), and significantly more than in proportion to availability (because the lower limit of the confidence interval for the selection index was > 1.0). Other vegetation classes in which drumming was observed included non-aspen vegetation types within the AQV (e. g., oak) and TTS (e. g., lowland conifers) habitat types, medium-aged aspen with the TM, AVO, and AOC types, young aspen within the AVO, TTS, AOC, and AQV 188 types, and old aspen within the TT S, TM, and AOC habitat types. The confidence limits for these vegetation classes, however, contained the value 1.0, suggesting no strong selection for or against those vegetation types (Table 3.3). Despite the lack of statistical significance between vegetation classes in 2005, notable differences in selection of different age classes and habitat types were evident. Generally, grouse selected young and old aspen within the AVO habitat type, non-aspen vegetation types within the AQV habitat type, and medium-aged aspen within the TM habitat type approximately twice as much as if there were no selection (Table 3.3). Old aspen within the TTS and TM habitat types and young aspen within the TTS type were about 1.5 times as likely to be selected as other vegetation classes (Table 3.3). Young aspen within the AOC and AQV habitat types, non-aspen vegetation types within the TTS habitat type, medium-aged aspen within the AOC type, and old aspen within the AOC type were neither selected for or avoided (Table 3.3). Vegetation classes in which no grouse were observed drumming, but which had expected values > 1.0 included medium-aged aspen within the AQV habitat type and non-aspen vegetation within the AVO habitat type. In 2006, grouse appeared to select aspen > 30 years old within the AVO and AOC habitat types more than in proportion to their availability (Figure 3.4), but the lower confidence limit of the selection index for only old aspen within the AOC habitat type was > 1.0. The selection index for medium-aged aspen within the FMC habitat type was the highest (w = 5.237), however, the confidence limit was large (0—13.72) and contained the value 1.0, which indicated no selection (Table 3.4). Other habitat types in which grouse were observed drumming, but where selection was not statistically significant 189 included young aspen within the TM, AVO, AOC, ATD, and AQV habitat types, medium-aged aspen within the AVO, AOC, and TM types, and old aspen within the TM, TTS, and AVO habitat types (Table 3.4). Some trends in 2006 included grouse selection of young asan within the TM, AVO, and AOC habitat types, and old aspen with the AOC type with > 2 times the probability of being selected than by random chance. Old aspen within the TM habitat type and young aspen within the ATD habitat type were selected similarly. Grouse appeared to exhibit no selection or avoidance of medium-aged aspen within the AVO, AOC, and TM habitat types, or for old aspen within the TTS habitat type (Figure 3.4). A greater number of grouse than expected were observed drumming in young aspen within the AVO habitat type, old aspen within the TM type, and medium-aged aspen within the AVO type (Figure 3.4). As observed in 2005, vegetation classes in which no grouse were observed drumming, but which had expected values > 1.0 included medium-aged aspen within the AQV habitat type and non-aspen vegetation within the AVO habitat type (Appendix 2). Other vegetation types in which no grouse were observed, and thus appeared to have been avoided in 2006 were non-aspen vegetation types within the AQV, TTS, and AOC (e.g., maple) habitat types (expected values = 0.99, 0.84, 0.92, respectively; Appendix 2). In both years combined, grouse selected for young and old aspen within the AOC habitat type (w = 2.30 i 1.27, w =1.982 :1: 0.91, respectively; Table 3.5), although a greater number of grouse than expected were also heard drumming in old aspen within the TM and TTS habitat types, and young aspen within the AVO habitat type. The only non-aspen vegetation types in which grouse were heard drumming occurred within the 190 AQV and TTS habitat types (Table 3.5). Vegetation types that did not have drumming grouse and were apparently avoided included non-aspen vegetation (e.g., maple) within the AVO and AOC habitat type, and medium-aged aspen within the AQV and TTS habitat types (expected values = 4.09, 1,82, 2.92, 1.41, respectively; Appendix 3). Spatial distribution of potential grouse drumming habitat Distinct regional differences in potential ruffed grouse drumming habitat (i.e., based on selection indices of different habitat types) were evident within the study area. In the southern and eastern portion of the study area, selection was relatively high because that area contains loamy soils characteristic of the AVO and AOC habitat types, and sandy-loams characteristic of the TM habitat type (Figure 3.3). The northern portion of the study area had low selection indices because that area is dominated by the ATD habitat type, which tended to have low selection indices (Table 3.2) . The remainder of the study area contained a mosaic of highly selected and avoided habitat types (Figure 3.3). 191 2: 3 2: $.82 295 as 8.0 Ed 85 2..: is m is 252 Seneca m: 02 Ed 92 3.2 one mm NS 3N8 :83 mesa ”3 who 26 as 3.8 3o mm 43 8.8% coma 2562: a2 2: w; of Sam woo 8 one 38m =83 20 “253 826; A3 Vow 9: e82: 3:on .o: 358w 838w 338% some a: 09¢ wo> :85 853250 Gong—om woaoomxm mo newtomoi mo .0 Z .«o :22?”on .ooom Ea moom wage ewwEomE .Bsmficom 8&3 88.0.03 05 E womb somwflowg commméo: 8 dogma $.89» om vv was?» 98 A989» 2198 vowméamoofi Amado» cm NV Be 5 8:on vote manage me ooeogooo 05 com 83?: seven—om no 53355 4m 03mm. 192 193 2: ov co; m E .23 £88. Nwd _oo.o oNd ovd mod mod N m E .o 3.3 m: Sim .86 MN; mm; mud mod _ Nod o2: USE 2: mod mNd an; mN.N_ Nvd S omd $62 UO< oN.m Sq.— omd RN mod and N_ Nfio NYE. O>< oN.N _oo.o Nwd £6 a _ A mod _ mod .3.: Ohm 34 3.6 mud n: wN.v N_.o m 2.0 3N0 SE. 36 bed N. _ .o h _ .o 3.: mod N aNd Snag >O< ©OON mod mm co; $.30 v.38. ooN de de 34 Ed :6 v o _ .o 3.8 wt mN._ nmd de mwd mix ONd N. de Nada UO< 9w.— Nod 0N6 om; 3.3 ovd 3 oNd mmdfl O>< ONN wed mmd mm._ exam Ed m :6 m0: SE. 36 n _ .o NN.o de mNd Ed m oNd Nd: >O< mooN SEED .530: A3 Vow :3 5?: 832m .0: 358w 025% 338% 38 a: HE E85 oocovmcoU eocoflom woaoomxm .8 cEtomoE («o .0 Z we :EtomoE .38 nguo org..— co mEooo m: can .mzom .382 .338 so 580 womb 552 USE was .UO< .O>< .m=0m 382-353 co mEooo QH< :58 $88 86 no 380 890. “852 SE. use >O< .E::M§ere%z§é%z§ H m: 08er -ufixegézfixci .I. USE mESEESEpUdNESEwQLeoV n UO< SuESEmQéNeS {mow u O>< ”mtmiebQémasimoV .I. QH< ”Eafimfiziagéwafl H SC. ”53:32: -waexmzmimew n >O< .33» @3388 EB ECON ,mOON wEBv 59:32 EsmEcom BED 8333 05 E 893 “353 EthE E 838w vote wEEEFE mo 8:05.08 05 .8» mooEE 33028 mo :EEEEmm .N.m 03$. M‘. .- .o_£mman_ 2a 806E sauce—om 05 SE 839 gamma: 8:83 cod 3 3832 soon we: BEBE 00:03:00 05 do.“ ES: 532 gamma: < F odd mm. dd; m _ New £809 RA de _md dwd NmN mod o 2d wvéw EL. EN _ood do; add 54 3d _ 5d mm. 2 USE Nb; mod NNd mm; N5: de VN VNd 5.09 UO< 32 Ed Ed 24 om.mN mmd dN Rd $.ch O>< 33 Bed dmd Ed cm; _dd E Nod S.E QH< mmN mud wed em; omd 2d 3 odd omNn SE. wed Ed 3d 3d 2..: odd 5 MNd SUN? >O< memo» doEnEoU dong: $304 A 6 vow A 9: SEE 3:on .8 83on 8:on BEES 83 a: PE ESE vacuumeoU sauce—om dowooaxm mo :EtomoE me .o Z we :EtomoE good S. 292. 4 9 35 u 307 anon-aspen 3 25 a IZSOyears 2 20 '7". 3 15 [330493965 5 o 10 inc—29m 5, 0 > E D O Q 0 2 2 F ’2 2 2 E «— B. 40 -- M _ 35,, anon-aspen E 30 3 25 o a IZSOyears .2 201 g 15 USO-49% S 10 O u0-29yaars 5 7 0 ~ > E D o 0) l— I- S 2 < 2 2 L t Xeric soils Hydric soils Sand Loam Organic Habitattypes Figure 3.1. Cumulative percent of aspen age classes and habitat types used by drumming ruffed grouse in 2005 (A) and available (B) in the western Upper Peninsula, Michigan. Habitat types listed include those that occur within Baraga, Dickinson, Iron, and Marquette counties. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga- Dryopteris; AVO= Acer- Viola-Osmorhiza; Fl— — Fraxinus-Impatiens; FMC= Frwcinus-Mentha- Carex; PCS=P1cea-C’ J ’ “ ' o PVC= Pinus-Vaccinium- Carex; PVD= Pinus- Vaccinium-Deschampsia; QAE= Quercus-Acer-Epigaea; TAM= Tsuga-Acer-Mitchella; TM— — T suga-Malanthemum TMC: Tsuga- -Maianthemum- Coptis; TMV= T suga -Ma1'anthemum- Vaccinium; TTM = Tsuga-Thuja-Mitella; TTS = T suga—Thuja-Sphagnum. 195 “ anon—aspen C Q 2 a IZSOyenrs 0 a 2 u30—49yaars 3 E 3 O a0—29years 45 -- —- ~— ~ ~ BI 40 a 35 7 DMD-85pm : 0 g > L I~50years O > 3 % also-49m E 3 O I0—29yaats > E o o to a r 2 2 2 E t Xeric soils Hydric soils Sand Loam Organic Habitat types Figure 3.2. Cumulative percent of aspen age classes and habitat types used by drumming ruffed grouse in 2006 (A) and available (B) in the western Upper Peninsula, Michigan. Habitat types that comprise < 5% of the area are not included. Habitat types listed include those that occur within Baraga, Dickinson, Iron, and Marquette counties. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; FI = F raxinus -;Impatiens FMC= F raxinus-Mentha-Carex PCS= Picea- (" J ' " ’ ,PVC= Pinus- Vaccinium- Carex; PVD: Pinus— Vaccinium- Deschampsia; QAE= Quercus-Acer-Epigaea; TAM= Tsuga-Acer-Mitchella; TM= Tsuga- Maianthemum; TMC= Tsuga- -Maianthemum- Coptis; TMV= Tsuga-Maianthemum— Vaccinium; TTM = Tsuga-Thuja-Mitella; TTS = Tsuga-Thuja-Sphagnum. 196 .o_£mman_ 2m 8065 c2823 2.: .8.“ 32S, 25%»: 3:82. cod .3 coca—QB :39 was BEBE 8:25:00 05 co... :8: .532 038%.: <— 32 2: mm :3 038 mash 3o _oo.o one 2.5 m2 8.0 _ 5o 3% >053 NE .85 as :6 23 mod N So 2.8 0038. N: .86 one :3 as 86 N 86 8% 003% *2 So :3 3o Be :6 v :5 8.5 53:» m: mad :8 3o 8+ :5 v :.o is» 05.88 8a .23 92 a? MS 85 _ 86 8.2 mt: SN 26 Rd 2: £.~ 3c m 86 $9. 003% am .85 OS N2 85 3o _ 3o 2.: mime» cam .86 2: m2 an 85 N 23 ”EN 2%? a? :6 23 3; N3 mod m mod 3:. misc mg m; 30 o: 0: mod m 85 3.2 o>33 mam z: a; $50. 8a ”3 a 23 M38 9220 8&3 5304 3» vow :3 x02: 838w .o: 838w 025% 3388 mos a: own 98 H: :85 mocodmcou couoofim doaooaxm do :oEomEm mo .02 do aomcomoi 197 .mmom omqmwao 26%: so 3580 m .51 28 .mzom bug: .288 no :68 $93 “952 00¢. 28 O>< ism $53. etc so 880 $93 55¢: 3:. Ba >O< .Ezcmcxmmémgrm -QMEH u WE. mamxcbéfizmxxézfixdxk H USE mEz-§me~§béfi50:36-69V u UO< SmESEwQéNeS -xmoV u O>< ”Easmfitfia: 19mg. 5 U SE: ”25.3.83; -waogmamimuw H >O< dumb noufiomg 3-5 commméo: can .390 fine» dm v comma was?» .808 ”28% 919$ :33 dowmébzdoe A23 may» on N comma eds—0E momma—o ow< .mddm wcgd :meoaz £33809 8&3 5833 05 E S.E womb “83m: was momma? own Eobbd E 3.:on vote mg no 082580 on... 8m 30de new—020m mo coumfismm .m.m 035,—. 633898 88 8088 cocoa—om 05 88 82? 0233: 3:83 odd 3 doom—8: :82 was 3288 8808.88 08 he :8: 632 czawo: <_ vodm dd. fl dv mod 3.3m 380,—. m _ A _ddd mvd mvd mmd mod _ odd hwdm O>O< £30m m8moT€§m :o 8:80 QH< ism 38% S8 no 880 $93 8858: EH 88“.. >O< .E:=m§&mém:§-um§& H m: Maegabéfizmgézfixdi n USE mEE~§Q£§U$N£888Qimow n UO< SwESEmQéBS {8% n O>< matmiebQémsmm .888 n QH< ”8388888: -8325 u EH 8888:; -mauumzmimuw H >O< .momb aowfiowg 8-5 838.88 88 AwEO 83» On v 838 was?» .258 M83» @799 comma 33-8888 A23 83% dm N comma “6208 momma—o ow< doom $888 8%282 .mEmEHHom 6&3 5383 05 8 CFC $93 8852 was momma—o own 880b6 8 38% vote w88§ mo 388.880 2: 8m 8088 80:023. mo :oumgmm .vd 2an I98 .0358qu 93 8068 canon—om 05 Sm mos—g gamma: 8:82. 86 3 c8582 soon was 3838 38.688 2: .8 :8: .632 03:82. 35... 8.. .85 $5 85 53 S5 . 555 525 8.88 3.. 555 25 85 85 85 m 85 8.8 835:. t... 555 R5 85 85 85 5 85 55.8 >088 85 .85 85 85 8.. .55 . S5 5...: 3.558 8.. 85 555 85 8.: m . 5 5. m . 5 5552 0:85 :5 .85 55 1.. 85 .55 8 S5 8.: 8:58 5W5 .85 555 58 8.. 85 5 555 $55 >038 _55 85 85 555 $5 85 m 85 8.: mg... 85 555 85 a: 85 85 5 85 5555 38.2.. 555 85 555 8.. 5.5 5.5 w 85 8.8 035.0 85 8.. 85 8.. m3 :5 _: 85 8.8 003.3 555 :5 85 8.. .55 85 5 85 2.8 0:58 85 .85 85 85 85 S5 . .55 25 2,558 85 85 85 555 555 85 m 85 5555 .255 85 8.. R5 85 85 25 5 85 555.. 09558 5555 .85 35 5.5 555 .55 . .85 .55 0282.. BED 5264 Tu» vow 3; 5888 8:on .o: 85on 33on 3388 585 a: own 88 PE :85 vococccoU cotoflom 380qu .8 .8anon mo .0 Z mo :oEomoE 5:8 ”8me0 0383 so 838 mHH 88 5:8 m852 0888 so 858 $93 8835: USE Ea .UO< .O>< 5:8 >852-.€85m so 8:80 QH< 5:8 .888 {cw so 880 893 “835: SE 88 >O< .Ezzmgmmramsfiémzfi H mrEl WRmEUéfimmgéxfixdi .1. USE mEzzémfigUéNESEmQi8V H UO< 65388808358 {wow H O>< mntmfiefiQéwxmmimuV H 95.. ”Ezfimfizfiagémak u SC. ”88583. -89.»:Qim2x H >O< .883 cosfiowg 8-5 838.80: 88 .380 88% Om v comma wane» .808 $80» 319$ comma 33-8838 A28 was?» on N comma 08:08 momma—o ow< doom can Sea @856 Emu—22 dBmfiaom Saab 8833 05 8 $5 $83 8855; can momma—o ow“ 80.8%8 8 3:on 3&8 w8§§8 mo 888.880 2: 8m $068 88828 HO 8888umm .m.m 035.5 Ruffed grouse selection indices High (1.54) Low (0.00) Figure 3.3. Spatial distribution of habitat types ruffed grouse would potentially select for drumming in the western Upper Peninsula, Michigan. Selection indices > 1.0 indicate selection; those < 1.0 indicate avoidance. The area includes Baraga, Dickinson, Iron, and Marquette counties. This figure was based on data collected during mid-April in 2005 and 2006. 200 DISCUSSION Results of this project demonstrated that ruffed grouse respond differently to the various ecological conditions that support aspen, and provided additional information useful for understanding ruffed grouse selection of different age classes and habitat types for drumming. To my knowledge, previous research has not investigated grouse response to the successional stages and site conditions that support aspen. In this study, selection indices for young (< 30 year-old) aspen were higher than those for older aspen age classes and non-aspen vegetation types, which was not surprising given other studies that also reported grouse selection of young aspen stands for drumming (Kubisiak et a1. 1981, McCaffery et al. 1996). Also consistent with previous findings was the lower selection or avoidance of non-aspen vegetation types (Gullion 1970). Svoboda and Gullion (1972) stated that in Minnesota, if aspen is not present in an area, breeding ruffed grouse are seldom present. Interestingly, however, ruffed grouse were observed drumming in older (2 50- year) age classes of aspen more than expected in 2005 (n = 15; expected = 11.87), in 2006 (n = 15; expected = 12.09), and in both years combined (n = 30; expected = 23.17; Table 3.1). Use of old aspen by drumming grouse has not been well documented in the literature in habitat selection studies, although DeStefano and Rusch (1984) reported instances where the density of drumming grouse was higher in mature aspen with hardwood shrubs than in young aspen. A compelling result of this study was the obvious selection differences of aspen among habitat types. In 2005 and 2006, selection for aspen within the AVO habitat type was the highest, followed by the TM and TTS habitat types in 2005, and the AOC, FMC, 201 and TM habitat types in 2006 (Table 3.2). One grouse was observed drumming in the FMC habitat type, which comprised a relatively small proportion (approximately 3%) of the sampled area (Figure 3.2), thus contributing to the high selection index. In combined years, selection for aspen within the TM habitat type was highest, followed by AVO and AOC. In 2005, 2006, and for combined years, aspen within the AQV habitat type was avoided ( Table 3.2) by drumming grouse, especially for young and medium-aged aspen although there was some between-year variation (Tables 3.3—3.5). Not enough grouse were observed drumming in the ATD habitat type to determine selection trends. Selection by age class within habitat types was also evident (Tables 3.3—3.5). For instance, within the AVO and AOC habitat types, grouse tended to select for young and old aspen age classes (Table 3.5). Medium-aged aspen within those habitat types appeared to be neither selected for nor avoided (i.e., selection ratios close to 1.0). Within the TM habitat type, grouse selection of all aspen age classes was evident, but had greater selection of young and old aspen (Table 3.5). Interestingly, within habitat types containing a conifer component (i.e., white spruce and balsam fir within the TM type, and lowland swamp conifers within the TTS type), grouse exhibited greater selection for old aspen than aspen < 50 years old or non-aspen vegetation types. These findings suggest that grouse select aspen differently depending on habitat type even if the aspen within different habitat types is in the same age class. Important habitat components, which seem to influence whether or not grouse will select a site for drumming are relatively high shrub or sapling densities (Palmer 1963, Gullion 1970, DeStefano and Rusch 1984) and overstory cover (Gullion 1970). For instance, DeStefano and Rusch (1984) reported that alder and aspen vegetation types 202 with saplings or shrub densities 4,000—15,000 stems/ha supported higher densities of drumming grouse in Wisconsin and Minnesota than other vegetation types with fewer stems, but areas devoid of overstory cover did not support drumming males. Optimal drumming habitat provides overhead cover from aerial predators and allows for horizontal surveillance for terrestrial predators (Gullion 1977). It is evident that structure and composition of vegetation within the aspen type varies within different habitat types (see Chapter 1, Coffrnan et al. 1980). These vegetation differences likely contribute to the different selection indices observed in various habitat types. It is not likely that the observed selection differences were a factor of differences in detection probability among vegetation types. Zimmerman (2006), for instance, found that temperature change during a survey and the interaction with temperature at the start of a survey influenced grouse detection probabilities. Selection of habitat by organisms is dependent on their requirements to survive (i.e., nutrition, protection from predators) and reproduce in an area. Selection or avoidance of different vegetation types, age classes, and habitat types by drumming grouse can be explained by their feeding and breeding habitat requirements. For example, the AVO and AOC habitat types support a thick shrubby understory (> 4,000 stems/ha), which is a structural component of optimal drumming habitat (Cade and Sousa 1985). Additionally, Svoboda and Gullion (1972) documented grouse use of aspen for food was greater in areas supported by loamy soils and peat than sandy loams and sand. The AVO and AOC habitat types both are characterized by loamy soils (Coffman et al. 1980). The TM habitat type and the old aspen vegetation within the TTS habitat type likely provide grouse drumming habitat because the conifer cover within those types 203 provides overhead cover and allows for horizontal visibility. Old aspen within the TTS type possibly provides more nutritious food than other habitat types, because of the organic soils (Svoboda and Gullion 1972) characteristic of this site. Grouse avoidance of aspen within the AQV habitat type (Table 3.2) may be attributed to the dominance of bi g- tooth aspen on the dry, sandy soils characteristic of the AQV habitat type. Svoboda and Gullion (1972) showed a significant (p < 0.05) preference for catkins of quaking aspen over bigtooth aspen. In 2005, grouse demonstrated a high selection (w = 2.06) for non- aspen vegetation types within the AQV habitat type. Red oak occurs after aspen in the successional trajectory of vegetation within the AQV habitat type (Coffman et a1. 1981). Kubisiak et al. (1980) documented grouse use of oak in Wisconsin and (Cade and Sousa 1985) documented grouse use of oak in other areas within their range in North America where aspen is absent. The general trend in selection of young aspen and old aspen over medium-aged aspen may also be due to those age classes providing specific characteristics required for drumming behavior (young) and availability and preference of mature trees for feeding (old aspen). Management implications It is undisputed that aspen is the key to providing ruffed grouse habitat. Habitat management recommendations in the Upper Midwest historically were to cut aspen in blocks S 4 ha in size to create juxtaposition of stands < 10 years old for brood cover, 10— 25 years old for breeding and winter cover, and mature stands for food (Gullion 1977). McCaffery et a1. (1996) stated that while small-tract cutting is feasible on private lands or management areas where ruffed grouse is the primary objective, it is not practical on 204 large tracts of public land where timber is the primary objective. Management for timber (e.g., large aspen clearcuts), however, is compatible with ruffed grouse habitat management. McCaffery et al. (1996232), for instance, observed that a 243-acre aspen clearcut in Wisconsin produced “relatively high densities of drumming grouse during the first half of the aspen rotation.” Knowledge of how grouse select vegetation to meet specific life requisites can help managers identify and maintain areas highly selected by ruffed grouse (Figure 3.3) through timber harvesting. In habitat types where grouse showed greater selection for younger aspen age classes, timber harvest may be implemented at 50 years (i.e., the standard rotation age for aspen in the Lake States). In habitat types where grouse showed greater selection for older aspen (e.g., the TM and TTS habitat types), perhaps harvest age may be lengthened to 70 years in some areas to provide unique vegetation features important for grouse. For instance, to meet grouse management goals, managers may retain areas within the AVO and AOC habitat types to increase the availability of areas selected by drumming grouse. It is challenging for managers to meet multiple forest objectives, but increased knowledge of ecological relationships and wildlife habitat selection helps managers better understand potential outcomes of manipulation. This study expanded on existing knowledge of ruffed grouse drumming habitat selection within the Great Lakes region, and determined selection indices for different aspen age classes and habitat types. Although selection does not equate with preference, it does give us insight into how grouse use vegetation characteristics and distribute themselves to meet life requisites. The value of assessing grouse selection of habitat types for drumming is evident in 205 planning forest management activities and understanding how grouse may respond to temporal changes in vegetation types through natural succession or manipulation. Future research should address ruffed grouse drumming habitat selection and relate fitness to different aspen age classes and habitat types. Nonetheless, the results of this analysis can help managers understand the breadth of ruffed grouse ecology, establish realistic grouse management goals based on the range of ecological conditions throughout landscapes that support grouse and aspen, or determine how and where timber harvest may benefit grouse over the long term. 206 MODELING THE CUMULATIVE EFFECTS OF ASPEN MANAGEMENT PRACTICES ON TIMBER AND WILDLIFE AT MULTIPLE SPATIAL AND TEMPORAL SCALES VOLUME II By Alexandra B. Felix A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Fisheries and Wildlife 2008 CHAPTER 4 SIMULATIN G EFFECTS OF ASPEN HARVEST STRATEGIES ON WILDLIFE HABITAT AND TIMBER IN NORTHERN MICHIGAN Forest management involves designing and implementing strategies to meet the current and future objectives of landowners and society. Forest management decisions, therefore, are based on scientific understanding of forest ecology and perceptions of how activities will promote human values (Davis et al. 2001). The understanding of forest ecosystems and human values of forests has dramatically increased since the I960S when the forest management philosophy changed from maximizing timber products to sustaining forest ecosystems for multiple purposes (e. g., wildlife habitat, recreation, timber; [Kessler et a1. 1992, Nyland 2002]). This broader understanding, however, has resulted in identification of areas in which the scientific community lacks knowledge, and the subsequent demand for further knowledge in those areas (Htittl et al. 2000). Research has indicated several thematic priorities on which forest research should focus. These included effect of harvest regimes on forest structure, composition, and function, and resulting wildlife habitat (DeStefano 2002), product yields and sustainability, and ecological integrity (Kessler et al. 1992). Although there is still much information to be learned about all forest types, research should focus on those that are not well represented in the landscape either because geographic areas supporting them are limited or because they are exploited by humans. Aspen, for example, is commercially valuable and is exploited for numerous products. In the Great Lakes States, aspen comprises more than half the industrial products harvested annually (Piva 2003). To maximize its commercial value, most aspen 207 in the Great Lakes region is harvested on a 40—60-year rotation because older asan begins to show signs of deterioration and loses economic value (Graham 1963, Nyland 2002). As such, 40—60-year—old aspen may not be well represented in the landscape to meet diverse ecological objectives. For instance, currently approximately 8% of all aspen on state lands in the western Upper Peninsula, Michigan is 40—60 years old (Doepker et al. 2002). In 10 to 20 years, the amount of old aspen (2 50 years old) will be limited. The Michigan Department of Natural Resources (MDNR) has indicated a desire to maintain a relatively even-age class distribution to sustain the resources and maintain ecological integrity. Currently, in Michigan and in much of the Great Lakes region, aspen forests are managed similarly (i.e., 40—60-year rotation) regardless of habitat type. Because aspen can grow on several habitat types ranging from poorly-drained organic soils to excessively-drained sand, the structure and composition of aspen forests varies with habitat type and thus provides different wildlife habitat components and timber harvest potential (Chapters 1—3). Habitat types, however, have only recently begun to be considered in forest management (Kotar and Burger 2000). The low proportion of area in aspen 40—60 years old and the fact that habitat types have not typically been used in forest planning has implications for management including timber harvest sustainability and wildlife habitat suitability. For example, old (> 70 years) aspen forests in Michigan support relatively large-diameter (2 15 cm) aspen trees and provide snags necessary for cavity-nesting species, including the red-headed woodpecker, which is listed as a sensitive species by the Audubon Society (see Chapter 208 2). There is evidence that timber volume also differs among habitat types, even within the same age class (see Chapter 1). State and federal agencies and many timber companies have begun to implement forest management strategies that seek to balance maintaining ecological integrity while using forests to provide products and services valued by people (Kessler et al. 1992, Pedersen 2005) and thus have supported research to expand knowledge in the thematic priority areas. Specifically, in Michigan, the MDNR has supported forest-related research because as of 2004, Michigan law requires state forests to be certified as being sustainably managed. Thus, the MDNR has assumed a responsibility to clarify the range of choices and consequences that exist for state forests, and develop criteria for making specific choices (Pedersen 2005). Recently, tools have been developed to help managers understand consequences of management activities and plan future activities. For instance, HARVEST is a timber harvest model developed by Gustafson and Rasmussen (2002) as a strategic research and planning tool that allows assessment of the spatial pattern consequences of forest management activities. HARVEST can be used to project the cumulative effects of management over time and evaluate alternative scenarios by providing comparable projections about how the different scenarios affect landscape structure (Gustafson et a1. 2007). HARVEST, however, is not a tactical planning tool and therefore, other methods such as linear or goal programming should be used to determine the best harvesting schedule to meet specific goals (Buongiomo and Gilless 2003). Because aspen comprises the second largest single share of timber sales (volume) in Michigan (followed by upland hardwoods), and is expected to strongly influence 209 timber harvesting plans on state forests over the next 20—30 years (Pedersen 2005), the MDNR has made it a priority to understand the impacts of applied management practices on aspen communities and take a proactive approach in predicting spatial and temporal consequences of management strategies to establish and maintain an even-age class distribution of aspen (R. Doepker, MDNR, personal communication). The HARVEST simulator can be used, in part, to assess the cumulative effects of aspen management strategies on sustainability of the aspen resource, landscape patterns, and distribution of aspen patches within the landscape matrix. This knowledge can help managers develop realistic expectations regarding future timber harvests and create a harvesting schedule that is sustainable for timber volume and wildlife habitats. 210 OBJECTIVES Specific objectives for this chapter were to 1) 2) Model the effects of current aspen management strategies on timber harvest sustainability and wildlife response for the western Upper Peninsula, Michigan. Simulate spatial and temporal effects of aspen harvest strategies, and determine threshold levels for aspen harvest within different habitat types to meet Michigan Department of Natural Resources forest management goals. 21] METHODS To model the effects of current aspen management strategies on timber harvest sustainability and wildlife response, I developed a model to simulate the optimum aspen harvesting schedule over 10 decades for different management scenarios specified by the MDNR (M. Mackay, MDNR, personal communication). The harvest schedule was then used in a spatially-explicit simulation model (HARVEST v.6.1 [Gustafson and Rasmusson 2005]) to evaluate landscape-level effects of harvest on wildlife and spatial arrangement of vegetation types. Prior to any simulations, however, I needed to compile data and organize it into a format useful for modeling. Spatial data compilation For the harvest simulations, a subset (300 mi?) of the original 4-county study area was chosen (Figure 4.1). This area was chosen because it represented the most contiguous area of public land (for which forest management goals and activities are known), and fewer assumptions would have to be made in this area about activities occurring on private land. In the 4-county area, 20% comprised state-owned public land, whereas in the selected subset, 60% was public land. Several data sources were used to compile all the information necessary to simulate timber harvest (i.e., forest type, stand boundaries, age class, habitat type). Because a spatially-explicit stand-level dataset (i.e., OI data) was only available for 77% of state land in the study area, I used IFMAP data to determine cover types for the remaining 23% of state land. I used a 3 x 3 majority kernel filter in ArcInfo to reduce the number of single-cell patches and assumed the IFMAP dataset was an accurate 212 representation of actual stand cover types as determined in OI records. The IFMAP data, however, did not indicate stand age. Therefore, I randomly assigned age classes to aspen stands such that the resulting aspen age class distribution was consistent with that indicated by current MDNR records (Chapter 1). In the new study area, I determined the amount (ac) of aspen occurring in each age class and habitat type on state land. These values were used to initialize the timber harvest scheduling simulations. For industry-owned, and other privately owned land, I used IFMAP to determine cover types and stand boundaries. Age of aspen stands was probabilistically assigned based on the age distribution of aspen in that region according to United States Department of Agriculture Forest Service Forest Inventory and Analysis (FIA) data. Because I was not evaluating spatial distribution of age classes of non-aspen forest types following simulated harvests, ages for non-aspen stands were not determined. I assumed that areas of non-aspen would remain in the same vegetation type and would not be converted to other vegetation types. I recognize, however, that management goals may change on MDNR and private lands over 10 decades. 213 0 15 30 60 Kilometers Figure 4.1. Location of study area (approximately 300 miz) in the western Upper Peninsula, Michigan for timber harvest simulations. Black areas are public lands managed by the Michigan Department of Natural Resources, dark gray areas are industry-owned forests, and light gray areas are non-industrial private forests. 214 Development of a harvest scheduling model T o examine the implications of timber harvest scheduling strategies on timber yield and wildlife benefits, I constructed a model in Microsoft Excel and used Premium Solver, an optimization program developed by F rontline Systems, Inc. (Incline Village, NV) to produce a timber harvest schedule. The model uses goal programming to determine the best harvesting schedule over 10 decades that minimizes the deviation from meeting specified forest management goals. Specified MDNR goals for the aspen cover type included maintaining timber harvest, balancing distribution of aspen area over 8 age classes (0—9, 10—19, 20—29, 30—39, 40—49, 50—59, 60—69, and > 70 years old), and maintaining wildlife habitat. The main goal, however, was maintaining an even-age class distribution (i.e., 12.5% of aspen in each habitat type should be maintained in each age class) to yield long-tenn sustainable harvest and maintain representation of all aspen age classes within all habitat types. I used a goal-programming approach to allow flexibility in the goal of equal amount of aspen within each age class. A constraint that imposed strict constancy on that goal would be unrealistic and likely lead to infeasible solutions (Buongiomo and Gilless 2003). The general goal programming model is described as follows: Minimize the negative deviation (i.e., amount of area falling short of specified goal) from meeting an even-age class distribution By changing Harvest area for aspen on 40-, 50-, 60-, 70-, and 80-year rotations Amount of area (acres) falling short of goal 215 Amount of area (acres) exceeding goal Subject to, Non-negativity constraints Acreage constraints Objective function (Equation 4.1) Min Z = Z( Dijk) Goal variables Hijk (i = 4, 5, 6, 7, 8) Dijk + Dijk Subject to, Hij 5 AU (Equation 4.2) AJ- = A0 (Equation 4.3) Hijk Z 0 (Equation 4.4) ijk 2 0 (Equation 4.5) ng 2 0 (Equation 4.6) Aijk + Dijk - Dijk = GU (Equation 4.7) Where, D' = amount (acres) by which the resulting area falls short of the desired age class distribution. 216 D+ = amount (acres) by which the resulting area exceeds the desired age class distribution. i = age class 1—8 (0—9, 10—19, 20—29, 30—39, 40—49, 50—59, 60—69, and 70—79 years old). j = habitat type 1—9 k = decade 1—10 H = harvest (acres) A = area (acres) A0 = initial acres G = amount of area (acres) needed to establish and maintain desired age class distribution goal Model inputs and structure The model determined the amount of aspen harvested on 40-, 50—, 60-, 70-, and 80-year rotations within each habitat type and decade to minimize the amount of area falling short of the desired even-age class distribution goal within habitat types (Figure 4.2). The amount of area potentially available for harvest (i.e., the amount of area in each age class and habitat type during a particular decade) was a function of the amount of aspen harvested in the previous decade (Equations 4.8-4.10). Because the MDNR has expressed no intent to convert vegetation types (e.g., through succession or management) at this time (J. Ferris, MDNR, personal communication), I assumed that all aspen would be harvested before it succeeded to a different vegetation type. A harvest schedule was determined for aspen within 9 habitat types ranging from xeric to mesic soils: PV (which pooled PVD and PVC), QAE, AQV, TM, ATD, AVO, AOC, F (which pooled FI and FMC), TMC. Aspen occurring in hydric soils represents about 10% of aspen occurring in 217 the study area and is typically inaccessible for harvesting (J. Ferris, MDNR, personal communication). For i = 1 (0—9-year-old aspen), Aijk = Hj,k-1 (Equation 4.8) For i = 2, 3, 4 (10—39-year-old aspen) Aijk = AM J,“ (Equation 4.9) For i = 5, 6, 7 (40—69-year-old aspen), Aijk = Ai-lj,k-l — Hi-1,j, k4 (Equation 4.10) Where, i = age class 1—8 (0—9, 10—19, 20—29, 30—39, 40—49, 50—59, 60—69, and 70—79 years old) j = habitat type l~9 k = decade A = area (acres) H = area harvest (acres) The deviation from the desired age class distribution (i.e., even proportion of area within each aspen age class) was the sum of the amount of area that fell short or exceeded the specified area goal for each age class, habitat type, and within each decade. The negative deviation, or underachievement, was only considered in the minimization of the ob'ective firnction because 12.5% of as en area in each a c class re resented a minimum J P g P 218 threshold goal for the amount of aspen to be maintained in each age class. Therefore, the MDNR was concerned about deviations which fell short of that goal. The total yield per decade was determined from the product of area harvested and average volume per area (Table 4.1) for each habitat type and age class. Volume per area for age class within each habitat type was determined from a regression relationship between age class (x) and cords per acre (y) as determined from 01 data (Table 4.2). Yijk = Hijk * Vij (Equation 4.11) Where, i = age class 1—8 (0—9, 10—19, 20—29, 30—39, 40—49, 50—59, 60—69, and 70—79 years old) j = habitat type 1—9 k = decade H = area harvest (acres) V = average volume per area Y = total yield Initial acres for each habitat type were the following: PV = 2076 ac, QAE = 12983 ac, AQV = 22384 ac, TM = 20230 ac, ATD = 784 ac, AVO = 44184 ac, AOC = 21129 ac, F = 5248 ac, TMC = 11204 ac. Initial acres for each age class within each habitat type are indicated in Table 4.3. 219 Table 4.1. Average volume (cords) per acre of aspen for different rotation ages and habitat types within the western Upper Peninsula, Michigan. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F rarinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = QuercuS-Acer—Epigaea; TM = T suga-Maianthemum; TMC = T suga-Maianthemum- Coptis. Rotation age Soil moisture Habitat type 40 50 60 70 80 Xeric PV 12.00 16.12 19.32 21.93 24.13 QAE 8.25 12.98 16.65 19.65 22.19 AQV 10.01 15.28 19.36 22.70 25.52 Mesic TM 12.09 16.23 19.44 22.06 24.28 (dry-mesic) ATD 7.61 14.67 20.16 24.64 28.43 AVO 10.85 15.17 18.51 21.25 23.57 AOC 9.81 14.30 17.78 20.63 23.04 (wet-mesic) F 10.44 14.87 18.30 21.11 23.49 TMC 7.98 13.40 17.59 21.02 23.92 Table 4.2. Regression equations to predict volume (cords/acre) yields for aspen on 40-, 50-, 60-, 70-, and 80-year rotations in different habitat types in the western Upper Peninsula, Michigan. Volumes were determined from 2006 Operations Inventory data from the Michigan Department of Natural Resources. AOC = Acer-Osmorhiza- Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = T saga-Maianthemum- Coptis. Habitat type Regression Equation R2 PV y = 14.318 * ln(x) — 36.695 0.89 QAE y = 16.450 * ln(x) — 47.702 0.92 AQV y = 18.296 * ln(x) — 52.215 0.90 TM y = 14.378 * ln(x) — 36.808 0.82 ATD y = 24.572 * ln(x) — 75.969 0.98 AVO y = 15.015 * ln(x) — 40.222 0.89 AOC y = 15.622 * ln(x) — 43.327 0.91 F y = 15.402 * ln(x) — 41.949 0.86 TMC y = 18.815 * ln(x) — 56.010 0.89 220 Table 4.3. Number of acres of aspen within 8 age classes in 9 different habitat types used to initialize an aspen harvest scheduling simulation model for a 300 mi2 area in the western Upper Peninsula, Michigan. Data were obtained from 2006 Operations Inventory Data from the Michigan Department of Natural Resources. AOC = Acer- Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-Acer-Epigaea; TM = Tsuga-Maianthemum; TMC = T suga- Maianthemum-Coptis. Habitat type Age class PV QAE AQV TM ATD AVO AOC F TMC 0—9 164 1532 2687 2986 81 5039 2212 621 1347 10—19 414 2913 4674 6305 12 11284 4758 2810 2704 20—29 557 3338 5051 3967 66 10207 4358 1390 1948 30—39 428 2185 2642 2224 15 5744 2681 316 1208 40—49 200 277 1051 1194 6 3109 1888 188 241 50—59 25 224 330 703 24 1636 1112 52 286 60—69 42 290 1448 1045 202 2389 1242 168 386 >70 148 1869 3593 1758 376 4281 2696 488 3069 Totals 1978 12628 21476 20182 783 43689 20947 6034 11189 221 Area of aspen in previous decade - Area harvested fl Area of aspen in each Desired age class age class and habitat distribution type 11 fl Area harvested at 40, 50, 60, 70, or 80 years within * Volume/area habitat types Resulting age 1 class distribution MINIMIZE: amount of area by which Total volume 1 harvested g _ 7 _ . resulting age class distribution falls short ofdeshed. Figure 4.2. Schematic diagram of harvest scheduling model structure. Boxes in bold are model outputs. I used information from chapters 2 and 3 on the relationship between avian species and aspen age class/habitat type, and ruffed grouse and aspen age class/habitat type to evaluate the effect of simulated timber harvesting schedules on wildlife habitat. In Chapter 2, avian species were grouped into communities based on similar associations with a specific aspen age class and/or habitat type. Communities included species associated with 20—50-year-old aspen within the ATD and AVO habitat types, asan > 50 years old within the ATD and AVO habitat types, aspen > 50 years old regardless of habitat type, aspen in the 20-year age class within the TM type, aspen > 70 within the TM type, and within the TM habitat type regardless of age. The amount of habitat available for specific communities was assumed to increase or decrease in direct relation to the resulting increase or decrease in the area of their associated age class and/or habitat type. As such, populations of avian species may increase or decrease with changes in the amount of area available within different age classes and habitat types. I cannot, however, assume a linear relationship between changes in population sizes and changes in aspen area because some bird species may obtain life requisites fiom vegetation types other than aspen. Cavity-nesting birds, for instance, are not specifically associated with aspen. To evaluate the potential effects of timber harvest on drumming ruffed grouse, I used information on the relationship between preferred drumming habitat and aspen age class/habitat type (see Chapter 3). I assumed that drumming activity (i.e., number of grouse drumming) was directly related to the amount of preferred habitat available (Gullidn 1970). Ruffed grouse preferred aspen within the TM, ATD, AVO, and AOC habitat types for drumming (Table 4.4). Habitat types characterized by xeric and hydric 223 soils were generally avoided (see Chapter 3). The following equation quantified expected relative proportional changes in grouse drumming activity: 3 4 . Ck=[>: Z(Pijk-Pij,k-1)*Wij]* Alk i=1j=1 Aj.k-1 Where, C = expected percent change in ruffed grouse drumming activity. P = the percent of area. i = age class 1—3 (young = 0—29 years, medium = 30—49 years, old = 50— 79 years old). j = habitat type 1—4 (ATD, AVO, AOC, TM). k = time interval (decade). w = selection index (Table 4.4). A = area. Table 4.4. Selection indices (17v) for ruffed grouse drumming habitat in 3 habitat types (ATD, AVO, and TM) and 3 aspen age classes in the western Upper Peninsula, Michigan. AOC = Acer-Osmorhiza-Caulophyllum, ATD = Acer-Tsuga-Dryopteris, AVO = Acer- Viola-Ozmorlziza, TM = T suga-Maianthemum. Aspen age classes included young (< 30 years old), medium (30—49 years old), and old (2 50 years old). Data on ruffed grouse habitat selection were collected in the spring of 2005 and 2006 on state lands in Baraga, Dickinson, Iron, and Marquette counties. Selection indices were determined based on methods described by Manly et al. (2002). A w Soil moisture Habitat type young medium old Drier TM 2.06 1.43 2.28 ATD 1.14 0.36 0.08 AVO 1.99 1.03 1.54 Wetter AOC 2.3 0.85 1.98 224 The selection index (W) indicated. grouse selection or avoidance of aspen in a given age class and habitat type (Manly et a1. 2002). Selection indices > 1.0 indicated selection, and those < 1.0 indicated avoidance. Aspen age classes and habitat types found to be avoided by drumming grouse (i.e., medium and old aspen within the ATD habitat type) were not included to determine C because a decrease in the amount of avoided vegetation classes would not affect grouse populations. An increase in the amount of avoided habitat would be reflected in a subsequent decrease in preferred habitat, which would negatively affect grouse. [Note: Ruffed grouse have cyclic populations. The relationship between the relative proportion of area of preferred habitat availability and grouse drumming activity was based on data collected in 2005—2006. Thus, direct effects of changes in preferred grouse habitat can only reliably be deduced for grouse population sizes in the western Upper Peninsula that are equivalent to the size which occurred at the point in the population cycle in 2005—2006.] This method is useful, however, for predicting increases or decreases in drumming habitat selected by ruffed grouse. Actual changes in drumming activity would likely be confounded by changes in population size due to natural cyclic fluctuations (Gullion 1970). The product of the difference in the percent of aspen (Pit) between years and the selection index 8 indicated the relative contribution of each aspen age class i and habitat type j to grouse for drumming. In essence, w was a weight indicating importance of a specific age class and habitat type for drumming grouse. For instance, if w for vegetation class A was 2 and w for vegetation class B was 1, a 1% decrease in vegetation class A would likely affect the grouse population twice as much as a 1% decrease in vegetation class B. The sum of the product for each age class and habitat 225 type was multiplied by the ratio of aspen area in each habitat type to the aspen area in each habitat type within the previous decade. This indicated the relative degree of drumming activity for each decade. The underlying assumptions in this model were that preference can be determined from patterns of observed use, and species fitness is higher in highly selected areas (Garshelis 2000). Simulation descriptions Three timber harvest scenarios were simulated. The target goal for all scenarios, as specified by the DNR, was an even distribution of aspen area within 8 10-year age classes from 0—79 years. That is, 12.5% of the aspen area was the target goal for each age class. Criteria for feasible solutions included representation of aspen in all habitat types, and sustainable timber harvest. Harvest was considered sustainable once there were no net changes in yield between decades. Wildlife habitat should be maintained, but specific criteria were not indicated. The first scenario specified that the total area of aspen within the 3 habitat types of interest did not decrease, and all aspen was harvested by 80 years. The second scenario specified non-declining aspen area, and only aspen 260 years was harvested. Non-declining aspen area and a 40-year harvest rotation were parameters for the third scenario. HAR VEST simulations The program HARVEST was designed to assess how the spatial and temporal patterns of age classes and forest composition may change under specific management 226 scenarios (Gustafson 1997). The program requires 4 input maps to simulate harvest activity: forest age, forest type, management area, and stand identification. These maps must be in an ERDAS v7.4+, 8- or 16-bit GIS file. I created these files in ArcInfo by using the GRIDIMAGE command at the Arc: prompt on grid files. In the forest age map, there were 9 initial 10—year age classes specified only for aspen. Ten vegetation types were specified in the forest age map: aspen (1), birch (2), cedar (3), lowland swamp conifers (4), lowland hardwoods (5), mesic conifers (6), pine (7), oak (8), northern hardwoods (9), non-forested (10). Non-forested areas included grasslands, bogs, and open water. Management units (n = 30) were defined by an intersection of land ownership and 10 habitat types (Table 4.5). Separate management units were identified for public land, industrial forests, and non-industrial private forest because management practices differ among them. The stand identification map was created by passing the age map through the REGIONGROUP clumping algorithm in ArcInfo, which groups contiguous pixels with the same age class values. Initial landscape conditions were identical for each scenario. Once input maps were rendered in HARVEST, I initiated simulation by selecting the “Execute” option on the Model menu and specified appropriate harvest parameters. The following parameters were used: for all land ownerships, I selected the “Fill Stands” option to avoid partial cutting of stands. Stands were randomly selected for harvest using the “Dispersed” harvesting method. The age of all harvested stands was reset to 1 following harvest. Adj acency constraints and riparian buffers were not enforced. _ For MDNR lands, the amount of area to be harvested within aspen stands 2 40, within each habitat type, and during each decade was determined from the goal 227 _« . ..'-- ‘—____. 4 _.- programming harvesting schedule. The amount of area to be harvested on non-industrial private forests was set at 6% of all areas 2 40 years (as specified by Gustafson et al. 2007). On commercial lands, minimum age allowed for harvest was 50 years and 100% of aspen Z 50 was harvested each decade. Succession of aspen > 80 years old only occurred on non-industrial private forests and within hydric habitat types on state-owned and commercial forests. Aspen within the PV habitat type succeeded to pine; QAE and AWV succeeded to oak; TM succeeded to mesic conifers; ATD, AVO, AOC, F, and TMC succeeded to northern hardwoods; and aspen within hydric habitat types succeeded to lowland swamp conifers or lowland hardwoods. HARVEST program scripts were saved for each management scenario. (Due to the extensive length of these scripts, they were not included in this document. I can provide them on request). Simulations in HARVEST are stochastic. That is, the specific outcome of an event is determined by a random number seed. Consequently, repeated simulations result in differences in locations of harvest events. Therefore, each scenario was simulated 5 times. Landscape metrics The amount of aspen area within each age class and habitat type resulting from the spatial simulations was calculated for each scenario to evaluate the feasibility of the aspatial harvest scheduling model when applied spatially. Average area of forest edge and interior was also determined to compare landscape effects among scenarios and identify potential implications for edge and edge-sensitive wildlife species. I assumed 228 edge effects penetrated 295 ft (90 m) into the forest (Gustafson et al. 2007). The spatial distribution of habitat for different avian communities, and the spatial distribution of ruffed grouse drumming habitat was assessed for initial landscape conditions, decade 2, decade 5, and decade 10. 229 Table 4.5. Management unit codes for specific land ownership and habitat types (HT) used as inputs for the HARVEST simulation model. NIPF = non-industrial private forests. Code for Code for Code for state-owned NIPF industrial HT code HT 101 201 301 PV Pinus- Vaccinium 102 202 302 QAE Quercus-A cer—Epigaea 103 203 3 03 AQV Acer-Quercus- Vaccinium 104 204 3 04 TM T suga-Maianthemum 105 205 305 ATD A cer- T suga-Dryopteris 106 206 3 06 AVO Acer- Viola-Osmorhiza 107 207 3 07 AOC Acer-Osmorhiza—Caulophyllum 108 208 308 F Fraxinus 109 209 309 TMC T suga-Maianthemum-Coptis 110 210 310 hydric 230 RESULTS Scenario 1: Harvest all aspen by 80 years Timber harvest In the first scenario (i.e., all aspen harvested by 80 years), the harvest schedule produced an even-age class distribution in all habitat types by the eighth decade; however, an even-age class distribution was achieved by the fifth and sixth decades for aspen within the TM and AQV habitat types, respectively (Appendix 1). Results of the spatial harvesting simulations achieved similar results (Figures 4.3—4.5). Spatial constraints such as no partial cutting of stands made it infeasible to achieve exactly a 12.5% distribution of area within each age class, however the resulting distribution was within 1% of the 12.5% goal for each age class in all habitat types by the eighth decade (Figures 4.3—4.5). The total number of acres that fall short of the even-age class distribution in each habitat type declined sharply from > 5,000 ac in the AQV, TM, and AVO habitat types and > 2,000 ac in the QAE, AOC, F, and TMC types to < 500 acres in the fifth decade of the simulated harvest schedule (Figure 4.6). The number of deficient acres was 0 by the eighth simulated decade (Figure 4.6). The total number of aspen acres harvested remained relatively constant at approximately 17,800 ac after the third decade, but decreased from a harvest of 20,829 ac in the first decade (Figure 4.7a, Table 4.6). Total volume decreased from approximately 197,000 cords harvested in decade 1 to 131,000 cords by the third decade (Figure 4.7b). Volume fluctuated more than total acres because aspen may be harvested in different age classes and habitat types, which may yield different amounts of timber (Table 4.6). Total 231 volume increased to approximately 168,700 by decade 7 and remained stable for the duration of the simulated timeframe (Figure 4.7b, Table 4.7). Relatively few acres (approximately 4700 ac/decade) were harvested from the xeric habitat types (PV, QAE, AQV) because they were not as prevalent as mesic habitat types in the study area and thus supported less aspen. Harvest remained relatively consistent in those habitat types throughout the 10-decade scheduling simulation (Figure 4.8). Volume harvested from the AQV habitat type, however, declined substantially from 37,500 cords in decade 1 to 16,000 cords in decade 3 (Figure 4.9, Table 4.7), due to the increased number of acres harvested by 40 and 50 years (Table 4.7). Volume increased to roughly 29,000 cords by the sixth decade, where it remained for the duration of the simulation (Figure 4.9, Table 4.7). The number of acres harvested within the dry-mesic habitat types (TM, ATD, AVO) was relatively consistent throughout the scheduling simulation (Figure 4.8). Initial harvest was 8431 ac, which increased to 8574 ac by the fourth decade and remained at 8150 ac for the remainder of the harvest schedule (Table 4.6). Most (> 5000 ac) of the harvested acres came from the AVO habitat type, whereas the ATD habitat type contributed relatively little aspen (Figure 4.8, Table 4.6), due to its low availability in the study area. Volume harvested from the AVO habitat type was lowest (31,651 cords) in decade 2 due to the increased number of acres harvested by 40 years, but volume increased to approximately 78,000 cords by Decade 5 where it remained relatively consistent for the rest of the simulation. Cordage harvested from the TM habitat type declined from an initial yield of 23,086 cords to 12,382 cords in decade 3, but stabilized at 24,868 cords by decade 5 (Figure 4.9, Table 4.7). Aspen volume from the ATD habitat 232 type was about 4000 cords in the first decade, and then stabilized at 1128 cords by the eighth decade. I Total area harvested within the wet-mesic habitat types (AOC, F, TMC) was approximately 6500 ac initially, and remained at approximately 4700 ac by the fifth decade (Table 4.6). Most (3172 ac) of the acres harvested were from the TMC habitat type initially, but then harvest decreased to 1400 ac, and most harvested aspen came from the AOC habitat type (approximately 2600 ac; Figure 4.8, Table 4.6). Roughly 650 ac were harvested from the F habitat type within each decade. In decade 1, about 62,000 cords were harvested from the wet-mesic habitat types; 44,453 cords were harvested each decade by the seventh simulated time period. The most obvious fluctuations in yield occurred in decades 3 and 5 because of an increase in the amount of acres harvested by 40 years old in those decades (Table 4.6). Most of the yield came from the AOC habitat type, which stabilized at 24,645 cords/decade after decade 6, followed by the TMC habitat type, which stabilized at 13,567 cords/decade by the seventh decade (Figure 4.9, Table 4.7). Cordage from the F habitat type was stabilized at 6241 cords/decade by decade 9 (Figure 4.9, Table 4.7). Effects on wildlife Avian communities. The amount of habitat available for 5 of the 6 avian communities increased with the establishment and maintenance of an even-age class distribution for aspen (Figure 4.10). Habitat for avian species using aspen > 50 years old increased the most (from 19% to 37% by decade 4). Habitat availability for avian species using aspen > 50 years old within the ATD and AVO habitat types increased by 12% by 233 decade 4. Habitat for birds using aspen within the TM type increased from 17% to 23% by the second decade; subsequently, there was a 1% increase in habitat availability for birds using aspen > 70 years within the TM habitat type by the fourth decade. A 2% increase occurred in the amount of habitat available for avian communities using 20—50- year-old aspen within the ATD and AVO habitat types by decade 4, although habitat availability initially increased by 12% in the first decade. The only community in which a decrease in habitat availability occurred was the 20—year age class within the TM habitat type, which declined by approximately 2% by decade 3 (Figure 4.10). Ruffed grouse. Once an even-age class distribution was established (i.e., by decade 8), ruffed grouse drumming activity was not expected to change between successive years (Figure 4.11). Initially, however, to establish an even-age class distribution, the expected percent changes were substantial. For instance, drumming activity was expected to decrease by 15% in the first decade of simulated timber harvest from initial conditions due to a decline in. the amount of preferred drumming habitat within the TM, AVO, and AOC habitat types. Expected drumming activity further decreased by 7% in decade 2, but then increased in the next 2 decades by a total of 15%. Although drumming activity was not expected to change between successive years by decade 8, total drumming activity was lower than initial conditions due to a decrease in the amount of younger (< 30 years) age classes of aspen (Figure 4.11). 234 PV 0.30 ‘ “ . 0_9 0.25 +1 0-1 9 S 0.20 7 +2929 +3039 0 E 0'15 --....t__40-49 n. 0-10 ..... 4.-..-50-59 0,05 —1—60-69 0.00 """" >70 QAE 0.30 ...._-,_.....,.. . . . H _ ............. . .. . ,.....r . 0-9 0.25 , , +1049 8 0.20 +2049 g 0 15 —o—30-39 g ' at 40-49 a 0.10 _ . 4450-59 0.05 ——1-——60-69 0.00 ....... >70 Decade Figure 4.3. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). All aspen was harvested by 80 years old. AQV = Acer-Quercus- Vaccinium; Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea. 235 0.30 c 0.25 O E 0.20 8 e 0.15 n- 0.10 0.05 0.00 . 0.70 ...._ .... ........ . 0.60 g 0.50 ‘5 0.40 8 A 0.30 0.20 0.10 TM 1 0'30 .,.._._... .. 0.25 g 0.20 E. 0.15 g 0.10 0.05 . 0.00 Decade —o—1 0-1 9 +2029 1 —o—30-39 —— g.-....4o_49 -- 0» -~ 50-59 —o—60-69 ....... >70 —o—0—9 —a—1 0—1 9 +20—29 —o—30-39 arm-4049 "mo—- ---50-59 —+—60—69 “—H . . “>70 Figure 4.4. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (r1 = 5 simulations). All aspen was harvested by 80 years old. ATD = Acer-Tsuga- Dryopteris; AVO = Acer— V iola-Osmorhiza; TM = Tsuga-Maianthemum. 236 AOC 025 .. ....., .. . . . ..... . . . .. . . . . . . ...,.._.... .. .. .. . .._.......__ A __..._.....M. .._.“..- ——o—-0—9 0.20 ., —n—1 0-1 9 g 1 . —a—20—29 g 0' 5 ‘ —o—30-39 3- 0.10 « .,. .4049 a . ~ - 50-59 0-05 —+—-60-69 0.00 """" >70 TMC 0.30 ' “ ' 0.25 0.20 0.15 0.10 0.05 0.00 Proportion Decade Figure 4.5. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza- Caulophyllum; F = Fraxinus-spp.; TMC = Tsuga-Maianthemum-Coptis. 10000 * Soil 9000 moisture 8000 gradient 7000 6000 . —.—Pv é Xeric mm ;——Q% 4000 : ------on 3000.. “\\ ' 20001 \ f- g 1000 \\- , H i: l ‘ _ 1 1 Number of 6366161111 acres 100001 9000 _ . 8000 v 7000 ‘ 6000+ 50004 4000- 30004* 2000 1000. Number of deficient acres TM Dry-mesic — — ATD , Mesrc .~HWAVt> 1.52 3'“ 4 5 6 7 8' '9‘W10 6000 _ _._ AOC V — — F Mesic-wet n-.1inc _ Number of deficient acres | Figure 4.6. Number of deficient acres of aspen needed to establish and maintain an even- age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-Acer-Epigaea; TM = Tsuga-Maianthemum; TMC = Tsuga- Maianthemum-Coptis. 238 25000 .2N.U-.-u .., ”H. .,._ -_.-vx“s,--_,swss-s 1 20000 \ ~ - — we - M .. 1 15000 , - -— ~ _L -u_______._, 1 1 1 10000 1 5000 Total acres harvested 1 Decade 200000 ~77 1500007 100000 50000 7.- -..H. -,.-- Total volume harvested (cords) ’ 1 2 3 4 5 6 7 8 9 101 Decade Figure 4.7. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. All aspen was harvested by 80 years old. 239 7000 4000 3000 2000 Acres beast—ed— 6000* 5000‘" 1000t- -+-PV : — — QAE ‘ AQV \ \ -—— —_ —_ —— —__—_—- o—ofl—o—o-fifi—o ; 1.. 7000' 6000‘ 5000 Acreshawesgd 40001 30001- ZOOOT 1000 w 12 3 45167 8 910 Decade —_ _— —— —— ——————— 6000 4000‘ Acres harvested 2000« 10001 5000*7 3000s ‘ _W _,. r... .fi- 1 2 3 4 5 6 7 8 Decade 9 '10 ’— 1 12345678910 Decade Soil moisture ggadient Xeric 1 Dry-mesic Mesic 1 Mesic-wet Figure 4.8. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = Tsuga-Maianthemum; TMC = Tsuga-Maianthemum-Coptis. 240 70000 60000 1,.500004 1 m "e 1 8 40000 ‘ V d) E 30000 '9 20000 100001 0 70000 1 1 60000 1 50000- 1A 1CD i'E 1 8 40000. , : E 30000. .2 1g 20000 10000.- 0- 60000 50000 40000 0 30000 E 20000 1 1 I 1 um (cords) 10000 0,- *i -«*—Fu/' ’— I—- —— ——— A A A v __........-..,_ _-- , .-_. .33. _- 1 2 3 4 5 6 7 8 9 10 1 70000 A 3 4 5 Decade ..—— .-fl: ., -2 1.. _...___ Decade 6 7 8 9 1o Decade 241 5 6 7 8 9 10 : ""(1AE ”“”AQV : "°"A£X3H ' ”“““ITflC 1 1 Soil moisture ggadient Xeric v Dry-mesic Mesic +1 Mesic-wet Figure 4.9. Volume (cords) to be harvested in aspen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-A cer-Epigaea; TM = Tsuga-Maianthemum; TMC = T suga-Maianthemum-Coptis. £3 £3 £3 £3 £8 £3 38 was" 33 new... >3. .3.; 38 new new new 38 62.. .8. om .3... 82 cm o o o o o o 8.. o o c on o o o o o o o o o o 8 o o o o o o 85 2... . o o on o o o o o .35 N .82 ea. c 9. >0... «em. new. new. .5. mew. mew. ..R. new. new. 3.: ”.20 .35 .3. Km. mm. mam. mam. 42. RN 4mm com 8.: ow o o 6.. c o o as. o o o on o o o o o o o o o o 8 o o o o o o a... we... 86 o ..m o o o o c an 0 RN as o 9. .55 an as ta ta SN us an EA 6% SN >.. .56... SN SN .8. SN SN 3 com 8 Q. w... 2. o o o o o o m. o o o E o o .3 o o o o o o o 8 o o o o o o o S o o 8 o o mm o o o o 3. EN 8 e. >.. S a w e e m .. m N . 98.0 am... 6%. .956: 2.3.5 $3.8: some» ow .3 5%.. .7. avg: 8 33 $8.23. maumuimm 3%. .333 55.3 :oEEEmE mmfio smudge .8 5952». was 5.338 8 393% o. .95 cwwmaoaz 632.com .253 5883 05 E womb .353 Sesame 55.3 mucu— uogoofibm .8 88.6.3.8: meow mo .3855 253:8 meta... comma. 6.... 2an 242 .88 .88 .88 .88 .88 .88 .88 .88 88 .88 03. .588 .88 .88 88 .88 .88 .88 8.8 88. 8... .8.. 8 8 o o 8 o 8 8 8 8 8.. 8 8 o 8.. 8 8 8 o o o 8 8 o 8 o 8 8 8 88 88 8 8 8 8 8 8 o o 8 o 8.... .88.. 88 8.. o>< 8 8 8 8 8 8 8 8 8 .8 5.8. .38... 8 8 M. 8 8 8 8 ..8 8 88 8 8 8 8 8 8 8 8. 8 8 8. 8 8 8 8 8 8 8 o 8 8 8 8 8 8 8 8 8 8 o. 88 8 8 8 8 8 o 8 8 8 8 8 8 8 8.. E... 88 88 88 88 88 88 88 88 88 88 2.8 .888 88 88 88 88 88 88 8o. 8 8 8: 8 8 8 8 8 8 8 o 8 o 88 8 8 8 8 8 8 88 o 8 8 88 8 8 8 8 8 8 8 88 8 8 o 8 8 8 8 8 o 8 88.. 88 8... 8 8.. s: 8. 8 8 8 8 8 .. 8 N . 88.0 88... 88. .85.... «63.5 83.83 380. .8... 8.888 243 new: new: «1cm: new: new: men: an: new: 3%: maecu mihch 88. 88 88 88 88 88 88. 88 88 88 028 .3.; 8.... 88. 2.8. 88. 8.. o .8. o 88 88 8 o o .8 8 8 o 88. .8 88 o 8 o o o o 8 o o 8 o 8 8 o o o o 88 8 o 8... 8 8 8 o o 8 o 8 88. o 88. 88 8 8.. 028 ..8 ..8 ..8 ..8 ..8 ..8 88 ..8 ..8 ..8 m .588 ..8 ..8 .88 ..8 88 8.8 ..8 8 o 8.. 8 8 8 8 o 8 8 8 8 8 ..8. 8 8 8 8 o 8 8.. 8 8 8. . 8 8 8 8 8 8 8 8 8.. o 8 8.. 8 8 8 8. 8 o 8 8 ..8 88 o 8.. h. 8.8 8.8 8.8 8.8 8.8 8.8 88 88 8.8 88 we... .88.. 8.88 8.88 8.88 8.88 8.88 8.88 888. 8... 888. 8888 88 8 8 E 8 o 8 8 8 o o 8 8 o 8 8 o o o o o o 8 8 8 8 8 8 8 88. 8.. 88 8 8 8 8 8.. 8 8 8 8 88. ..8. 8 8.. co... 8. 8 8 8 8 8 .. m N . 88.0 88... 8.5 .958: 2.30m— .8382 3.83 6... 03.8... 244 88888 88888 88888 88888 88888 88888 88888 8888. 8888. 88.88 >98 .388 88888 88888 88888 88888 88888 8888. 8888. 8.88 888... 88.88 88 8 8 8 8 8 8 8888 8 8 8 88 8 8 8 8 8 8 8 8 8 8 88 8 8 8 8 8 8 888.. 8888 8 8 88 8 8 8 8 8 8888 8 8888 8.88 8 88 >08. 88.8. 88.8. 88.8. 88.8. 88.8. 8888. 8888 8888 .888 8888. 53.8.8.8 88.... 88.... 8888. 88.... 88.... 888.. 8888 8888 8888 8888. 88 8 8 888 8 8 8 888. 8 8 8 88 8 8 8 8 8 8 8 8 8 8 88 8 8 8 8 8 8 8888 8888 88.8 8 88 8 8 8 8 8 8.8 8 888 8888 8 88 8.80 8.88 8.88 88.8 8.88 8.88 8.88 8888 8.8. 888. .88. >8 .588 8.88 8.88 888. 8.88 8.88 8.88 888. .88 888 .8... 88 8 8 8 8 8 8 888 8 8 8 88 8 8 888 8 8 8 8 8 8 8 88 8 8 8 8 8 8 8 888 8 8 88 8 8 88. 8 8 8 8 8.8 888. .88 88 >8 8. 8 8 8 8 8 8 8 8 . 888.0 88% 888288888. 3:88: .8838: .8898 ow .88. comma :8 .8983: o. 883 38.8.8 $5883.83 .8898. .35.... 55.3 8.2.8558... 8838 88.8.55. .8 £88.38... .858 .888—£8.88 0. 8288028 o. .26 8.83.822 £33.58. 88.85 88.8883 8.... 8.. 8898. .858: .5888... GEE... 8.8.8. vogo-o.38 ..o 3.883.... 5%.. ..o 8.8.8 ..o .3832 .8... 83.8... 245 33m 53m 82m 33m 33m umfium have bvcmm wmmmm nmwav 08,888.82; Sam Ram No.88 Ram 12mm .29 mnemm $02 :88”... oowoq ow 8 8 8 8 8 8 8 8 8 8888 88 8 8 88.8 8 8 8 8 8 8 8 88 8 8 8 8 8 8 8888. 88.8 8 8 88 8 8 8 8 8 8 8 8888. 8888. 888. 88 o>< 88.. 88.. 888 888 .88 888 888 888 88.. 8888 8.8.8 388 88. . 88. . 88 8 8 8 8 888 88.. 8888 88 8 8 888 8 8 8 88. 88 8 888. 88 8 8 8 888 8 8 8 8 8 8 88 8 8 8 8 .88 8 88 888 8 8 88 8 8 8 8 8 888 888 8 8 8 88 8.8.. 88888 88888 88888 88888 88888 88888 8888. 8888. 8888. 88888 28.8.88 88888 88888 88888 88888 88888 888.8 8888. 8 8888 8888. 88 8 8 8 8 8 8 8 8 8 .88 88 8 8 8 8 8 .888 8 8 888 8888 88 8 8 8 8 8 8 8.88 8 8 8 88 8 8 8 8 8 8 8888 8888. 8888 8 88 8.8 8. 8 8 8 8 8 8 8 8 . 888.88 888. 2.829.888. oflaoon “mug: 8.88. .8... 8.888 246 88.88. 88.88. 88888. 88888. 88888. .888: 88888. .8888 8.888 .8888. 8.8.898 888. 888. 888. 888. 888. 888. 888. 888. 888. 8888 0.88 .388 888. 888. 888. 888. 88.. 8 .8. 8 888 8888 88 8 8 .8 8 8 8 888. .8 888 8 88 8 8 8 8 8 8 8 8 8 8 88 8 8 8 8 888 8 8 88 8 8 88 8 8 8 8 8 888. 8 888. 888 8 88 028 888 888 888 888 888 888 888 888 888 888 8.. .588 888 888 .88 888 888 8.8 88 8 8 888 88 8 8 8 8 8 8 8 8 8 88. 88 8 8 8 8 8 888 8 8 8.. 88 88 8 8 8 8 8 8 888 8 8 88 88 8 8 88. 8 8 8 8 888 888 8 88 8 8.88 8.88 8.88 8.88 8.88 8.88 8888 8.88 8.88 8888 Do... .388 8.88 8.88 8.88 8.88 8.88 8.88 888. 8... 888. 8888 88 8 8 88 8 8 8 8 8 8 8 88 8 8 8 8 8 8 8 8 8 8 88 8 8 8 8 8 8 888. 88.. 88 8 88 8 8 888 8 8 8 8 888. 8.8. 8 88 00.8 8. 8 8 8 8 8 8 8 8 . 888.0 888. 888.358.. 2:8qu— .8./88$ 3:00. .8.... 03.8% 247 45.0 40.0 35.0 30.0 -f , _. _. _. .___-__.____ 25.0 20.0 15.0 10.0 - 5.0 Percentof aspen In drfferent age classes and habitat types contributing to avian community habtiat 0.0 " , . , Start- 1 2 3 4 5 6 7 3 9 10 2006 Decade —o—20—50-yr-old ATD, AVO >50-yr-old ATD, AVO ....... >50-yr-old aspen 20-yr-old TM .._.-.. . .>70-yr-o|d TM _+__TM all ages Figure 4.10. Expected percent changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. All aspen was harvested by 80 years old. Percents may sum to > 100 because some communities overlap age classes or habitat types. ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with 20—50-year-old aspen in the ATD and AVO habitat types, aspen 2 50 years within the ATD and AVO types, aspen Z 50 years old, aspen in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type. 248 .E..Em£§..§<-sm..88 u E... 883.82.86-58; .88... u O>< firmiegémak -8888» u DEA. m88.838883898833586{8.888 n 00.... $883.23.... 88.5% .._ 882.832.... 0:98 .8....8: 52888 .9. 0.. 8.88.8... 3.8 .80... 888... 6.02. .20 8.898 3. .3 3.83.8: 8.83 82.88 E... d8wEouz .0 8.38558. .8853 on. E .8838... 888.88 3.8.38.8 w...30=o.. 88888.8 3 .0 £888 .0 E... 8.... .8 383.0230 2.8 8898. .938: 82.8.8.8. 8......» Augew 3.85836 .0 .8588... 8.3.38 3.88:... 888.0% 88.8.... E 8owa8no 88.3qu .. ..v 8.89... apnoea maxi s: .. 8.88- ..x_. w a m. . 1 . - - -, i- . i . : 1. r QON- D. W m m. w O m m. H. e -1 1 -w I _ w , 1 .i . -o.m_.- 1.. .2808 m m. .2 w 8.... o>< W N! w - ; . 8.88- m. m E... ....... , n m. i i m. w E ¢ - - - 1 r l 8.8- mm . .ON m. h 8. 8 8 8 8 8 8 .-m m 249 Scenario 2: Harvest aspen Z 60 years Timber harvest In the second scenario (i.e., harvest aspen Z 60 years), an even-age class distribution was only achieved in the PV and AQV habitat types by decades 10 and 7, respectively (Figure 4.12). Within the AVO habitat type, resulting area was within i 1% of an even-age class distribution (Figure 4.13), and within the AOC type, resulting area was within :1: 2% (Figure 4.14). The total number of acres that fall short of the even-age class distribution in each habitat type declined to < 700 ac by the eighth decade in all habitat types except for F (Figure 4.15). The total number of acres harvested remained relatively constant at approximately 18,000 ac/decade by decade 5 (Figure 4.16a, Table 4.8). Harvest decreased from an initial 20,706 ac in the first decade to roughly 16,000 ac in the second decade. Number of acres harvested per decade was similar to the harvest for the first scenario (harvest aspen by 80 years). Similar fluctuations occurred for total volume harvested per decade. Initial volume was approximately 198,700 cords in the first decade, 145,500 cords in decade 2, and between 166,000 and 172,300 cords for decades 6 through 10 (Figure 4.16b, Table 4.9). Trends in number of acres of aspen harvested were similar to those of scenario 1 for the xeric habitat types (Figure 4.17). Number of acres harvested within the QAE habitat type, however, demonstrated more fluctuation among decades until decade 8 when harvest appeared to stabilize at 1623 ac/decade (Figure 4.17, Table 4.8). Volume harvested from the xeric habitat types decreased from 57,000 cords in the first decade to about 45,000 cords/decade after decade 5 (Table 4.18). Volume harvested within the 250 QAE and AQV habitat types was lower in decadesl—3 when more acres were harvested from aspen in the 60 year age class rather than older aspen (Figure 4.18). Total acres harvested within the TM habitat type increased to a high of 3223 ac in decade 5 from an initial 2529 ac, and then stabilized at approximately 2500 ac in decades 6 through 8 (Figure 4.17, Table 4.8). Aspen acres harvested within the AVO habitat type remained relatively consistent throughout the 10-decade scheduling simulation (Figure 4.17). The ATD habitat type contributed < 300 ac of aspen after the first decade (Table 4.8), but the harvesting schedule never yielded an even-age class distribution. Fluctuations in volume were not as extreme as they were in Scenario 1 (Figure 4.18), although cordage was higher (Table 4.9). Volume from the AVO habitat type ranged from a low of 42,163 cords in decade 2 to a high of 52,720 cords in decade 5 (Figure 4.18, Table 4.9). Volume from the TM habitat type ranged from a low of 15,315 cords in decade 2 to 24,868 cords in decades 6 through 10 (Figure 4.18, Table 4.9). Aspen volume from the ATD habitat type was less than that harvested in Scenario 1, except in the first decade. Harvest stabilized at 1128 cords by decade 9, but did not achieve an even-age class distribution (Table 4.9). Total area harvested within the wet-mesic habitat types (AOC, F, TMC) was similar to acres harvested in Scenario 1 (Table 4.9). Although harvested acres were similar within each habitat type as they were in Scenario 1, number of acres fluctuated more during the course of the 10—decade simulated harvested schedule (Figure 4.17). Total volume for these habitat types was also similar to that in Scenario 1 (Table 4.9), but volume did not stabilize within any of the wet-mesic habitat types (Figure 4.18). 251 Effects on wildlife Avian communities. The amount of habitat available for 5 of the 6 avian communities also increased in Scenario 2, but later than in Scenario 1 for communities associated with 20—50-year old aspen within the ATD and AVO habitat types, and aspen Z 70 years old within the TM type. Habitat for avian species using aspen > 50 years old decreased by 6% in decade 1, but then increased by approximately 25% by decade 4 (Figure 4.19). Habitat availability for avian species using aspen > 50 years old within the ATD and AVO habitat types increased by 13% from initial conditions by decade 4. Habitat for birds using aspen within the TM type increased by 6% in decade 2, but habitat availability for birds associated with old aspen in the TM habitat type did not increase until decade 5. A 5% increase from initial conditions in habitat availability occurred for avian species using 20—5 0-year old aspen within the ATD and AVO habitat types by decade 4, but this followed a 26% decline from conditions in decades 1 and 2 (Figure 4.19). The only community in which a decrease in habitat availability occurred was associated with the 20-year age class within the TM habitat type, which declined 2% by decade 4 (Figure 4.19). Ruffed grouse. Expected ruffed grouse drumming activity at the end of the aspen harvesting simulation was approximately 2% higher in Scenario 2 than in Scenario 1. Initially, however, drumming activity was expected to decrease more than in Scenario 1 because the amount of young aspen in the TM and AVO habitat types decreased more under Scenario 2. For instance, activity was expected to decrease by 14% in the first decade and then by an additional 15% in the second decade (Figure 4.20). Changes in expected activity never stabilized, but fluctuated within 1% between successive years 252 after the fourth decade (Figure 4.20). Expected activity within the TM, ATD, and AOC habitat types increased from initial conditions, whereas activity within the AVO type decreased (Figure 4.20). 253 0.30 ..w .. . 0.25 - g 0.20 E 0.15. '2‘ a 0.10, 0.05 0.00 0.30 ' ‘ " 0.25 S 0.20 50.15 a. 0.10 0.05 0.00 . 0.20 8 £0.15 8 90.10 a. 0.05 0.00 - PV QAE AQV ' —o—0—9 —D— 1 0-19 +20-29 —°—30—39 ~ - w— ~40—49 ----- 0~~~-»5O-59 — —i—60—69 -—-~->7o —+——-0-9 —I—10—19 . —-A——20-29 —0—30-39 —~»~'fl~ —40-49 «ow—5059 -—-i—-60-69 - --—- - «>70 --_—“>70 Figure 4.12. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen Z 60 was harvested. AQV = Acer-Quercus- Vaccinium; Pinus- Vaccinium—spp.; QAE = Quercus-Acer-Epigaea. 254 TM 0.35 '* ~ 0.30 ‘ ‘E 0.20 a O 15 2 u “- 0.10 0.05 0.00. AVO 030 -W . , ., ,_ _. .. W. W WWWWWWW 9 0.25 i ‘ ' " "‘ +1049 8 0.20 +2o-29 E 0.15 —o—3a39 a. —+~—40—49 E 0-10 ~“Jess 0.05 ——+—-60-69 0.00, """" ’70 Figure 4.13. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen 2 60 was harvested. ATD = Acer—Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = Tsuga-Maianthemum. 255 —-l—1 0-19 —a—20—29 —°—30-39 ~+ - ~50-59 -—-o——60-69 ....... >70 2:: 9 on CD 0.30 .._., "...... . ..W ,. ...._........ ._ “WWWWHW... 0.25 ‘ ' _5 0.20 5 0.15 , S a 0.10. , 0.05 0.00. Start- 1 2 3 4 5 6 7 8 9 10 2006 Decade Figure 4.14. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Only aspen Z 60 was harvested. AOC = Acer-Osmorhiza— Caulophyllum; F = Fraxinus-spp.; TMC = Tsuga-Maianthemum-Coptis. 256 i . , 10000 ‘ - 9000 f I 8.011 8000, morsture 7000 gradient 6000 *_- ._._.Pv . . 5000 ~~ ; — _ QAE Xenc 4000 —— —\ ' " ""“AQV 3000 - \\ 2000 A \\ \ 10001- - \W.-. __ __ __; 0.____'_ ' .\§._ 4 A A x I Y ‘1 ‘ A Number of deficient acres — 7————v— — 7 —»-—-—.-— r W V‘ ' V' 1 2 3 4 5 6 7 8 9 1E?“ Decade 10000., 9000 8000 . 7000.- 6000 _ ~ +TM 1 Dry-mesic 5000 ~_ ‘ — — ATD Mesic 4000+ ' AVO 3000,-,, , \\ 2000 i 1000 " " .‘\\ Decade V Number of deficient acres 10000"" 9000* 8000“*' 7000 * 6000 * * —.—Aoc V 5000 -~~ i - - F : Mesic-wet 4000 ~ - ' "”‘TMC 3000 20004 1000— Number of deficient acres Decade Figure 4.15. Number of deficient acres of aspen needed to establish and maintain an even-age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. Only aspen Z 60 was harvested. AOC = Acer-Osmorhiza-C aulophyll um; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-A car-Epigaea; TM = Tsuga-Maianthemum; TMC = Tsuga- Maianthemum-Coptis. 257 A. 25000 . .. .. _fl j .3 20000 \/\ H i w _m y 7 ‘65 'é a 15000 ,, - ,, _,, #c. ,, ‘ .C l § 10000 a , N T g 5000 . _ , -- - , i , 1‘ ' 1— 0 e a 1 2 3 4 5 6 7 8 9 10 Decade 8' ... 250000 . ., 1. -___”fl 1 “’ 1 o 1 p 3 200000 - - , -4. i 1: _ i 3 l ‘— i 1 3 150000 7- can _ . E g 1 1 g 1 0 100000 , s, } i l E .1 9- 1 2 50000 f g 1 ,_ 0 1 2 3 4 5 6 7 8 9 10 Decade Figure 4.16. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. Only aspen _>_ 60 was harvested. _shargsted Decade h rvested 0'! O O c> j 1000» : o._;:_>._=_ . -_.——-«—""s———— l 1 2 3 4 5 6 7 8 9 10 Decade 7000 l 6000 i8 L‘JSSOOO “2’ ’«14000 ‘3 5(20001. A . -_ i / \ 1000 l\\ i,/ \,_‘_,__ i 1 2 3 4 5 6 7 8 9 10 Decade —o-—PV — — QAE ‘W“AQV —o—TM -— ATD ~~AVO —+— AOC ..-..- TMC Soil moisture gradient Xeric Dry-mesic Mesic 1i Mesic-wet Figure 4.17. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Only aspen 2 60 was harvested. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- T suga- Dryopteris; AVO = Acer- Viola-Osmorhiza; F = F raxinus-spp.; PV = Pinus- Vaccinium- spp.; QAE = Quercus-A cer-Epigaea; TM = T suga-Maianthemum; TMC = T suga- Maianthemum-Coptis. 259 i Decade i 60000 4 Volume #(cords) o o o O 10000 1 2 3 4 5 6 7 8 9 10 ‘ Decade 70000 i 60000e ’u? '2 50000 13: ' ”40000 E Figure 4.182 Volume (cords) to be harvestedina 1 2 3 4 5 6 7 8 9 Decade —o—-PV — — QAE —ww_on —+—TM -——-ATD -- AVO —+—AOC —-—F mwTMC Soil moisture gradient Xeric V Dry—mesic Mesic v Mesic-wet spen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Only aspen 2 60 was harvested. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = Tsuga-Maianthemum; TMC = Tsuga-Maianthemum-Coptis. 260 £8 £3 £3 £3 33 £8 £3 £3 £3 mama >3. .55 38 £8 £8 2: E: m: o m: we: .82 ow o o o o woo an o o a: o E o o c 8a m 32 £8 82 E: c 8 o o o o o o o o o o 3 o c o o o o o o o o ow >o< «5 #2 mi £2 $2 $2 mu: EB =3 3.: 55 .3.; 8s .8”: mm: £2 ”E as o O can as: ow o o a. o o o 8m o EN o E a: o o c o 22 8: £2 Rm o 8 o o o o o o o o o o cm o o o o o o o o o o ow mo an Eflfiaa Ea 5:388 8 863% 2 .8>o gwfiomz nagmfiaom Saab €8.83 05 E 893 “832 :58on £53» £53 3850.083 8% @835: modes no 53835 235:8 “meta: comm< .mé 2an 26] 3% 3% Gem 3% 3% «new $3. 3% $3. 6% 53.. .35. 3% 3% 3.8 8% 8% s o me. 88 a? ow o mm... N? o o o m: o o 3. E g o o o o 2% 8:. $3. 2: m m: _ 8 o o o o o o o o o o om o o o o o o . o o o o S o>< .3 .3 .3 an a a e z 2: man a: .32. Ma ”a o o o 2 e a we as ow o o o o S 8 o o o s E c o me an s o o o o o 8 o o o o o o o o o o 8 o o o o o o o o o o 8 a? 3m" 23 8mm QR S: 23 mg 23 E: 33 2e .33. as: mg. mg 23. 3: o c o :3 an: ow o o o o as o o 8N m2 $5 2 as o o o a; 3mm .83 4g 8” o 8 c o o o o o o o o o om o o o o o o o o o o 8 s: S a a s e m a m N _ was ow< 25 35$ 253: “main: 388 .3. 2.3 262 3:.— ewchu cma: bNNw— unmwm nmca— 3%: mg: :3: 3ch ijOH. 82 82 82 3.8 82 82 82 8a Na 88 Us; .33 Na 82 S: 84 88 o o o 88 88 8 c o a o as o o o 8N o 2 88 o o 22 o 82 82 82 SN o 8 o o o o o o o o o o 8 o o o o o o o o o o 8 03a. 8m 88 .8 E. «ms 83 EN 8m 3: 88 m .53. mm: 88 :8 EN 88 o o Nm 0 84 8 8m o o o a: o o o c we 8 o o o o o 22 EN 8: a: o 8 o o o o o o o o o o 8 o o o o o o o o o o 8 m ”SN 8NN EN «EN ”EN wSN ”SN NGN NGN 88N 09.. .85 EN owNN N_NN NGN MEN N8 N; Na: Nee 88N 8 o o 88 o mNN o 88 o o o 2 o c o o o 2M: 8: 82 85 o 8 c o o o o o o o o o 8 o o o o o o o o o o 8 02 S a a N 8 m v N N _ ago owe. 25 55$ 2.309 5295 388 .3. 0:3 263 NELN NvPN SEN 8va =88 88NN 88N EEN :8N 8.8 >o<§§ NENN NENN 38 89: N3: 88 o 8: 89: 88 8 o o o o 88 88 o o no: a 8 o o o NN: M: 882 82N 852 88 o 8 o o o o o o o o o o 8 o o o o o o o o o o 8 >o< 5.2 88— 8:: 8:; 8:; 85 8:: 888 8N8 33: 3.0.53 8: 88: 85 8:; 8:; 8_N o o 88 38_ 8 o o 8m o o o 88 o 8: o 8 28 o o o o 88 W8: 82: 88 o 8 o o c o o o o o o o 8 o o o o o o o o o o o8 mo SW 58:88 was 52388 3 868% 2 H96 sawEomE .33:th 6&3 5283 05 E 885 2352 Eofimfifi £515 mw§_ 3850.088 :o 880%: comma mo $88 Mo 53:52 dd 2an 264 weNfim mmSm mmb—m nm—NM Sam suave mmr—v mma—v Norm.“ vmrav O><=3°H mofiwv 85¢ Nofiwv _m_Nm imam 3% o oohv cmvmm acwov om o 88 88 o o o 88 o o 88 8 88:. o o o o 8888 888 8th 888 N88 8 o o o o o 8 o o o o 8 o o 8 o o o o o o o 8 o>< 8N: EN: 88 88 88 88 88 8N 8NN :88 a: .589 8N: 8N: o o o E 88 8N 88 88 8 o o o o 8: 88 o o 8 8 8 o o a8 88 888 o o o o o 8 o o o o o o o o o o 8 o o o o o o o o o o 88 m8< 888 8va 888 8va 888 N88 N83 88 88. _NSN 2238.8 888 888 888 8va 88: o o o 88 N8: 8 8 o o o 82 8 o o 88 88 88 8 88 o o o 88 N82 N82 N8: 888 o 8 o o o o o 8 o o o o 8 o o o o o o o o o o 8 3:. 8 8 8 8 8 8 8 m N _ 886 8< 8838888 25qu 8838: 388 .3 288 265 6253 gave“ ewavem mama 3:: Nae: mama vecauu 22.2 63mm— main—L0H. 88— 88. 888 88_ 882 88 88 38 88 888 0:? .3.; 88 882 88_ 88 88 o o o 88 88 8 o o 88 o 88 c o o 88 o 8 88 o o 82 _ o 88 88 38 a: o 8 o o o o o o o o o o 8 o c o o o o o o o o 8 88 88 88 :8 88 88 88: 88 88 82 88 m .3.; a: 88 :8 8: 88 o o 88 o 88 8 88 o o o 8:: o o o o 83 8 o o o o o 88: 88 88 82 o 8 o o o o o o o o o o 8 o o o o e o o o o o 8 m 888 888 888 888 8:8 888 888 8.28 :88 8.8 03. .3.; 888 888 :88 888 888 88 88 882 N8: 888 8 o o 88 o 88 o 88 o o o 8 o o o o o 882 882 882 88 o 8 o o o o o o o o o o 8 c o o o o o o o o o 8 oo< S a 8 8 8 m 8 8 N _ 82o 98. 25 88mm 2:.er 803mm 388 .3 28B 266 2 s g 2 50.0 3 8 a :3 .3 E cg 40.0 32 " a ‘E ">' 33 30.0 — 0 a ’6 a C :3 20.0 h: " o ‘5 a 10.0 1 '5', 8. g 3‘ E 0.0 - Start- 1 2 3 4 5 6 7 8 9 10 2006 Decade __._20—50-yr-old ATD, AVO >50-yr-old ATD. AVO ....... >50-yr-old aspen 20-yr-old TM .- _m >70-yr-old TM ——4—TM all ages Figure 4.19. Expected changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Only aspen Z 60 years old was harvested. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with 20—50-year-old aspen in the ATD and AVO habitat types, aspen Z 50 years within the ATD and AVO types, aspen 2 50 years old, aspen in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type. 267 8883586833 (38 u O>< Himiogémah 983‘ u E< ”532832330 -aNESEmQLon u UO< fimcouflsaoa 8:on E mcoumaosa 0:98 858: 53880 Ho: ow munch @8835 82: ”087a .3838; mm? 20 who» co N comma Eco dmeomE mo «Banana $85 05 5 “meta: hon—Eu 33:88 go 868% 3 @6523.“ 30382—8 98 395 83an cobomoa 55;» 35885.6 838m mo 85:53 53qu mafia—5.6 3:on cube 5 mowcmso unwounxm .omé 2:me 23000 mi 2.2 --Izl villi: 9mm- .m— o W, iiiiii j l v l l : l, lie. 0.8. M M n w l 1 l: o8- W. m m. m. ! - -. .- 1 ll: llllll odm- W m 8<|1o| _ , . - 1-.| - .. -IIIK .-. OAMFI W m. o>< m J a .m, 9< ....... o8- m m. SFIOI M.5 a m oh- M n e a m m w w W m. w l l l l - l , on sun M 0.3 268 Scenario 3: Intensive harvest of aspen (60% of all aspen harvested from the 40—year age class) Timber harvest In the third scenario, establishing an even-age class distribution under a more intensive management scheme (i.e., harvest most aspen on a 40-year rotation) was not feasible (Figures 4.21.—4.23). Representation of aspen in all age classes within all habitat types was not even feasible under this management scenario (Figures 4.21—4.23). The total number of acres that fell short of the even-age class distribution in each habitat type declined after the sixth decade in most habitat types, but was still > 1000 ac by the end of decade 10 in 7 habitat types (Figure 4.24). The total number of acres harvested fluctuated between roughly 22,000 and 30,000, but exhibited a downward trend over the 10 simulated decades (Figure 4.25a). Total volume harvested per decade also declined over the simulated timeframe. Initial volume was approximately 220,000,700 cords in the first decade, but declined to 127,000 cords in decade 10 (Figure 4.25b, Table 4.10). The number of acres of aspen harvested was not stable within the xeric habitat types (Figure 4.26). Total number of acres declined over time, except for a peak in harvest in decade 6 in all xeric habitat types (Figure 4.26, Table 4.10). Number of acres harvested declined by 2300 acres within the PV habitat type, 400 ac within the QAE habitat type, and 220 within the AQV type (Table 4.10). Volume harvested fi'om the xeric habitat types exhibited a similar pattern (Figure 4.27). Total acres harvested within the TM habitat type increased from an initial 3983 ac to a high of 7007 ac in decade 3, but declined in subsequent decades to 2523 ac in decade 269 10 (Figure 4.26, Table 4.10). Aspen acres harvested within the AVO habitat type increased throughout the simulation with 2 large peaks in decades 3 and 6. Harvest increased by over 7,000 ac from initial conditions in decades 3 and 6 (Figure 4.26). Final harvest was approximately 1700 ac higher than initial harvest (Figure 4.26). The ATD habitat type contributed 391 ac of aspen after the first decade, and declined to 171 ac by decade 10 (Table 4.10). Total volume declined in all dry-mesic habitat types over the simulated timeframe, although fluctuations in yield were evident (Figure 4.27; Table 4.11). Total area harvested within 2 wet-mesic habitat types (AOC, TMC) also exhibited a downward trend over the duration of the simulation (Figure 4.26, Table 4.10). Harvest from the AOC habitat type declined by approximately 2000 ac, and harvest declined by roughly 3000 ac from the TMC habitat type (Table 4.10). Although harvest from the F habitat type fluctuated, initial and final harvests were the same (804 ac). Total volume for these habitat types also decreased throughout time (Figure 4.27, Table 4.11). Effects on wildlife Avian communities. Habitat availability generally was lower for avian communities under Scenario 3 than the other 2 management scenarios. The amount of habitat available for 4 of the 6 avian communities increased in Scenario 3 from initial to final conditions, but was lower than in the other scenarios and exhibited dramatic fluctuations during the simulation (Figure 4.28). Available habitat for avian species using aspen > 50 years old within the ATD and AVO habitat types increased by 6% from initial conditions by decade 4, but then decreased by 16% by decade 6. Habitat 270 availability increased again in subsequent decades and was 5% higher in decade 10 than in initial conditions (Figure 4.28). Habitat availability for avian species using aspen > 50 years old increased by 4% between the first and last decades (Figure 4.28). Habitat for birds using aspen within the TM type decreased by 2% between the first and last decades. Habitat availability for avian communities using old aspen within the TM habitat declined to 0 in decades 4 through 9 (Figure 4.28). Ruffed grouse. Expected ruffed grouse drumming activity at the end of the aspen harvesting simulation was much higher in decade 3 than in Scenarios 1 and 2, but resulted in lower predictions at the end of decade 10 (Figure 4.29). Changes in expected activity never stabilized and fluctuated drastically, with an overall decline in activity. Expected activity within the TM and ATD habitat types was more stable than in other habitat type (Figure 4.29). 271 0.30 0.25 g 0.207 o 0.15 E 0.10 0.05 0.00 . 0.35 " '“ "‘ ‘ 0.30 c 0.25 5 0.20 0.15 n- 0.10 , 0.057 0.00 . Figure 4.21. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. AQV = Acer-Quercus- Vaccinium; Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea. 272 0.35 0.30 ° 0.25 0 0.20 0.10 0.05 0.00 0.60 0.50 _§_ 0.40 E 0.30 D. 2 0.20 , n. 0.10 0.00 _ 2 0.157 TM ATD Decade ' +1019 +2049 440-39 «4% M4049 -...,.,..-...5o-59 —+—-60-69 ------- >70 Figure 4.22. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. 0.30 ' 0.25 5 0.20 ' E 30.15 1 0.10 0.05' 0.00 0.45 0.40 0.35 5 0.307 'E' 0.25 8.0.204 7 2 0.15 0.10 , 0.05 0.00 TMC —o—o.9 —-—1 0-1 9 +2049 7 —0—30-39 .--a+--—40-49 -¢ -~-50-59 —o—60—69 ------- >70 —¢—1 0-19 +2029 — —o—-30-39 ———x_-40-49 ..... .,_5o.59 —+—60-69 ------- >70 Figure 4.23. Average proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest (n = 5 simulations). Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-('aulophyllum; F = Fraxinus-spp.; TMC = Tsuga-Maianthemum- Coptis. 274 l i s . I . . 200001 8011 8000 6000 — 4000 ~~ ‘ W- s j —0— PV . 2000 _ _ QAE X6110 0000+ . . on 8000» --i -. f g l moisture gadient tacres A—A—L—A—l Ien 4000 ax l 2000 “— "“ \ ,A/ ‘ \’ 5, Number of defic 200001' 1,, 18000” $16000 " w 1 a (U i E14000“ ~ '* w , . 12000 . . —°—TM D’y'mesm . ‘ ‘ — — ATD 1 MBSiC 123:3.» : --------- 6000“ 4000 2000* ___ ‘ l 0 1 2 3 4 “—5 6 7 i8 ‘ 9 10 Decade cue ' Number of defi 1 l, ,s 20000‘ {W 180001” '§ 16000 *5 14000 f}: 12000} 10000 3000* 6000‘ , 4000*‘ ”” 2000 0~ V Mesic-wet Number of de i Decade L __.,._ _ , , . Figure 4.24. Number of deficient acres of aspen needed to establish and maintain an even-age class distribution within different habitat types over 10 decades of simulated harvest in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer—Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = Tsuga-Maianthemum; TMC = T suga-Maianthemum-Coptis. 275 30000 . , , , , i 1 25000 1 20000 i 1 15000 i . . 3 --_- -- am 10000 I - ., , ,fi ,i ? Total acres harvested 5000 ,, ..- ,_,_ Decade 93 250000 l - ._AJ.H~ , a . ..-,Uw--fl,p3-mnmu,x V 1 200000 150000 100000 . . -fl__ns_cs_s_m.____li 50000 , ..- a, , , ,. .r Total volume harvested (cords) l 1 2 3 4 5 6 7 8 9 1o Decade Figure 4.25. Total acres of aspen (A) and total volume (cords; B) harvested in aspen stands in the Upper Peninsula of Michigan over 10 decades of simulated harvest. Most (60%) of the aspen was harvested on a 40-year rotation. 276 . 14000 - - - ‘ Soil 12000 . . moisture t, a, i gra lent E10000 1 PV 2 8000 .0 w - - QAE ' i 1:“ ~ ~AQV Xenc «at; 6000 ‘ 2 4000 » ,\ / '. (. 2000.e~\\‘_// V \;'/z"\, 0 _W_ 712345678910 Decade i 14000 1 12000 1 lg 10000 x v 2 80001e~w ‘ 2 +TM rY-mPSlC * a, . ' MeSlC 0) la < l i 1 l Decade U 2 . E ? y 8 ' MeSlC-WCt b < 1 ‘77—'73 4M5 6 7 8 9 10 l l Decade Figure 4.26. Acres of aspen to be harvested in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer-Tsuga-Dryopteris; AVO = Acer- Viola-Osmorhiza; F = Fraxinus—spp.; PV = Pinus-Vaccinium-spp.; QAE = Quercus-Acer-Epigaea; TM = T suga-Maianthemum; TMC = Tsuga-Maianthemum-Coptis. 277 Soil 1 700001 60000 —-- moisture ’9‘ 50000 . Meat 1 E —o—Pv ~ :8: 40000 _ f, 235 Xeric I § 30000 B __ . . > 20000 \ 7. / \.\ .. 5 2 \ / \ ,_ ,I " 10000 t_- \§./ _. WW \_ _ 1 2 3 4 5 6 7 8 9 10 Decade "A i if i i 1 l 100000 9 L ; 90000 , 7 i :E 70000 ~ - _ V 60000 — . TM Dry-mesnc 1 “5’ 50000 ’ " ‘ - — — ATD > ' Mesrc = ‘w., - ,6 40000 ‘ ' ”AVO" 1 > i ‘ 1 2 3 4 5 6 7 8 9 10 Decade 70000 600001 6‘ 1 E 500009 ; 1 ; 1 ‘§’400001 3 -—o—-ACK3 1 .r E 7 — - F 1 MeSlC-WCt % - , ~-----TMC > 1 Decade 1 Figure 4.27. Volume (cords) to be harvested in aspen stands in different habitat types over 10 decades of simulated harvest to establish and maintain an even-age class distribution within habitat types in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40-year rotation. AOC = Acer-Osmorhiza-Caulophyllum; AQV = Acer-Quercus- Vaccinium; ATD = Acer- Tsuga-Dryopteris; AVO = Acer- Viola- Osmorhiza; F = Fraxinus-spp.; PV = Pinus- Vaccinium-spp.; QAE = Quercus-Acer- Epigaea; TM = Tsuga-Maianthemum; TMC = Tsuga-Maianthemum-Coptis. 278 man 3?. 33 was 33 RS amen 25m 33 was >3. .3.; o o c o o o E: on 5.: 8mm ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o om 2mm :2 new 38 8% 2% new :8 3% $3 9. >3. 2.2 33 $2 32 «new 52 $2 at: 9.3 3.3 E5 3.; o 82 OS $2 we: as: RN «mm 8m as: ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o 8 $8 9.: we: cm 38 3: 82 mm: 8: Se 3. 95 Sn 3". ta ta am new 3». so. am Rm i _95 o o o o o 0 com mm Nv mi. cm 0 o o o o o o o o o on o o o o o o o o o o oo o o o o o o o o o o Om mmm mmm New New mom Km vow v 3 Em wmv ow >m S a w s e m 4 m N _ 920 ewe. 8.: 9%: 3:32— $3.82 .888% 8: Be 8.8 comma 38,—. 48:88 .39» ov a co comma __m we Xoc “mote: 8 33 .8886 @533: .893 “853 553, cousnwammc 320 03-5.6 ca 58558 98 52988 8 868% o— 53 gwfiomz .flsmfiaom 5&5 5833 05 5 393 “833 Bonfire ESE €52 8850-033 so,“ @885: meson Mo 59855 23028 “mote: commas .o fl .v 2%... 279 mm: Gem Svm am: wamfl mace Gem cNQN— mm: Gem O>< 33,—. o o Sow o 3.3 $3. ac; emf awmm 594 cm o o o o o o o o o o E o o o o o o o o c o 8 o o o o o o o o c o 8 mm: 6% EN mm: mm: 5% am as: 8:. 3: 9. o>< E as a 3 as an 5 3. as an a... .55 o o o o o o e 3. SN Em ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o om E mom a em 98 am a E 8 2 e. a: 23 83 33 ma... Nam 23». 3:. SE 2% 33. 2:95 o o o o o o a: as 2.2 mm: ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o c o o o o o em 32. 8mm 3mm as. 2% Sam 3% 88 $3 was 9. 2e 3 a a s e m .. m a _ 920 am... 833353 33qu $3.8m :88 .oE. 29¢ 280 an: 3.8: «mm: 2.3." 3.45. 3st SEN 8mm :8». came" 3.4.59 3.: «EN 82 32 2.: Ex .32 32 as: Eu. 0:; .38. s3 82 oi 32 as o :.N em as $8 ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o om NE 8: 82 NE NE gm 3: 82 m 2: 82 9. BE 4% ..ms ..ms 2.: $2 25 a; Sam $2 25 .._ .3.; o o o o o o a: mm ”2 we. 8 o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o 8 3w 5 5 8: 32 wow so 2% 8: 2m 2. a $3 $2 £3 EN». 2% 3mm 8:. 3% 8% 3mm .09. .35 o o o o o o as: 2 : $2 2.8 ow o o o o o o o o o o E o o o o o o o o o o 8 o o o o o o o o o o ow $8 MEN £8 3Q 8% saw 2mm 3:. mm? gm 2. co... 2 a a s e m .. m N _ 920 um... 899%: 3:83— smog: €08 2.... 29¢ 281 SEN EM: 3%: ..EE ESN ENmN SEN memNN S.E». .....Ee >0... .55 o o o o o o ENS 2: NE: 85 E o o o o o o o o o o E o c o o o o o o o o E o o o o o o o o o o E SE Em: SEE «ES GEN ENEN «ES Ea: EEN 2:: E >o< S.E ..Em. E3 NS: NMEN MNEN 5.8 EE Na.» ..NEN $5.38. 0 Rs: NBN moo: 2:: 3:: EN EE mEN EEE E o o o o o o o o o o E o o o o o o o o o o E o o o o o o o o o o E S.E RE E4 S S.E RE .39. E2. EE ENE. 2. mad 2: SS 82 NE_ NSN «EN EN EN an ER .3 .35 o o o o o e E2 :.N we. 5: E o o o o o o o o o o E c o o o o o o o o o E o o o o o o c o o o E 2: 32 NE_ NE_ NSN MEN wE NSN EN EN 2. Z S a w E e m .. m N _ 926 as... safes“: 2:309 “83mm 090.2030 8: Ba 83 comma 38. A8388 893:. a no comma 2e mo 338 “meg: 3 SB .3883 wcumotmm .889 533 £53, :23“:me mmflo owwégo Na 5859: use 5:383 8 330% 2 .350 33:32 ismficom 895 5883 on... E 895 “832 EchoecE £53» ".652 3850.086 :0 @8833 comma mo £88 mo 325.2 A 3. 033. 282 Em: EEN SEE ..EE Emma ES :2: ENE mEE ESE o>< 3.; o o EEN o E3 E5 ESE _Efl EE EEE E o o o o o o o o o o E o o o o o o o o o o E o c o o o o o o o o E Em: SEN _22 EE ESE WEE Em: EEE NEE E: E o>< Rm «E EN =: E» ES 2». E». E2 EE a: .55 o o o o o o E EN EN EE E o o o o o o o o o o E o o o o o o o o o o E o o o o o o o o o o E EE :5 EN 2 _ EE 82 E E WE E E DE EE ES Ems Ea: GEN E3 3% ER 2:3 233 SEQ; o o o o o 0 E: 2% ENE NE: E c o o o o o o o o o E o o o o o o o o o o E o o o o o o o o o o E Ems EMS EMS ESN SEN E2 E9: GEE WEE E? E s: E a E E E m a m N _ ago a? 255%: 2:“qu Evin: 288 ._ 3 03$ 283 «mug $9.2: mmvww NEE: hmmaa— enema cmavfl Samw~ ave: NMNVNN mibch $5 —mmm— mvrm :5: wmv: Ema age was 33 mmwmm USE. _Soh memo 3mm _ om: gmm fl Noam o wmmm comm mmhm Nmnmm ow o o o o o o o o o c on o o o o o o o o o o oo o o o o o o o o o o om comm mm? ooom Mme comm _omo ovum :Nv 3mm Sam ow 02H 3mm an; $2M mam name 3mm 33‘ ~32 F: 3% r.— 38,—. o o o o o o a: mow cog mvov 0w 0 o o o o o o o o c on o o o o o o o o o o oo o o o o o o o o o o ow 00mm on; 3:». flow 33 3mm Rom mww: Pwm «m2 ow m 5%: $2: 33: 532 $me an: maven chan 253 vcmmm UO< .3.; o o o o o 0 GS: 32: mam: mmfim ow o o o o o o o o o c on o o o o o c o c o o oo o o c o o o o o o o em 2%: $2: moi: :03 cmmmm 00mg wwmw 3%: 2m: 32: ow UO< 3 a w b e m w m N ~ EEO ow< cab “SEES 3:80: 63.83 388 ._ 3 2an 284 500 ,,._-~-c,i__m.__-. - -.__-__~ as- “an“ 50.0 40.0 30.0 20.0 10.0 0.0 Percent of aspen in different age classes and habitat types contributing to avnan community habitat Decade __._20-50-yr-old ATD, AVO >50-yr—old ATD, AVO ....... > 50-yr-old aspen 20-yr-old TM - w—w >70-yr—0ld TM —‘—TM all ages Figure 4.28. Expected changes in availability of different ages of aspen and habitat types that contribute to habitat for 6 avian communities using aspen during the breeding season following 10 decades of simulated timber harvest in the Upper Peninsula of Michigan. Most (60%) of the aspen was harvested on a 40—year rotation. ATD = Acer-Tsuga- Dryopteris; AVO = Acer- Viola-Osmorhiza; TM = T suga-Maianthemum. Avian communities included those associated with ZO—SO-year-old aspen in the ATD and AVO habitat types, aspen Z 50 years within the ATD and AVO types, aspen Z 50 years old, aspen in the 20-year age class within the TM habitat type, aspen > 70 in the TM type, and aspen of all ages within the TM habitat type. 285 .Easefieemuxxémfim u SF WRESEEQEBS teams u O>< ”utmiegémsmrm 13V n QH< mEaménfiacbéuEernQ{soV n 00¢. 6823888 838w 8 mcoumaosu 0:93 3.828 .8288 8: cc 385 380.88 805 682a 80:88 .8??? a no 888$: mm? 888 05 mo 3.68 322 8me0:2 mo «8883 $85 05 8 E252 838: 36886 mo 888% 3 836:8 $038230 88 $95 “838 895.88 85:3 68on 88883.8 Mo 89835 33qu 98882.8 0808 vote 8 mowgao wowoomxm dad 08me stucco r. ., Lint- xv.-.» -il.r.l1l tttttt a :1: r l. «i f- , l.-- fill- »yl.ll.l..o-!-lllll|l...\il liniiiii: itll. , l i --L i. .. , -I I. r. 1- {Tillilil .1123... 0.0Nl a M x w. w H m in . 11!; t ,t i , it tit ,‘ r i r . o.m—.l p o n w m . - R o o_. m. M. m a .59 m. m . - u OO<||¢| o m a” m O>< M... m. . M. e E.< ....... _ 0.0 .M m E IT I. u w oh 6 0. my. 0 m M 0.3 a a 0.2 286 Landscape metrics The amount of area classified as forest interior slightly increased from 411,889 ac to > 416,000 ac by decade 10 under scenarios 1 and 2, but remained similar to initial conditions under scenario 3 (Figure 4.30). The amount of forest interior under scenario 3 was always lower than that under scenarios 1 and 2 throughout the simulated timeframe (Figure 4.30). Fluctuations in the amount of forest interior between years differed among all scenarios, except in scenarios 1 and 2 after decade 8 when an even-age class distribution was achieved. The largest difference in the amount of interior area was observed in decade 3 between scenario 2 (418,384 ac) and scenario 3 (381,449 ac; Figure 4.31). The amount of edge declined under all management scenarios throught the simulation (Figure 4.30). 287 430000 ' ‘ H ' "“‘s . ,-"~ 1 Interior 420000 ~ g 1 4100000 1 400000 l_ .333 .,3. g t" ; —°—Harvest by 80 I 0 390000 .fi ; ‘.... ° - Harvest > 60 i ‘ 380000 _ . _ l "TT‘TJNEfiifiLaEYCStl JV // 3300004id-~~-~~u~ - - -—~ w - -~~ ~~fi/ 3200001 i Edge 310000 7" g 2000000 g 290000 _ i SUNT- 1 2 3 4 5 6 7 8 9 10 2006 Decade Figure 4.30. Amount of forested area classified as interior and edge afier 10 decades of simulated aspen harvest in the western Upper Peninsula, Michigan under 3 management scenarios: harvest all aspen by 80 years, harvest aspen > 60 years, and a more intensive harvest where 60% of all aspen harvested was in the 40-year age class. Edge area was defined as area within 295 feet (90 m) of a forest edge. .0300 00008 a no A8 0% t mmN 5503 «05.... >5» 60002300 003 035 A0202”: 002 me $20 0mm 39:; 05 82m 003 0000030: comma 05 .«o 008 N 000:3 98 €2.52: 088 and 0000030: 003 20 £00.» ow A comma :0 00:3 E030: comma 683:8? 09800.3 0 mo m 0.00000 waist camEomz .EBEEm “0&3 E0303 05 E 003 NE 8m 0 E @092? Ema 0MB 98 @0605 M53 0200:: 3000.“ ho was: 40v PEEK 00:: om 289 DISCUSSION Several concepts and techniques were integrated and synthesized in this chapter to demonstrate a process for identifying the effects of aspen management practices on wildlife and timber resources. For instance, the habitat-type classification system was used to predict vegetation development and successional change of aspen, and was linked with wildlife community and species associated with different successional stages of aspen in various habitat types. Then, availability and spatial and temporal distribution of aspen was manipulated through spatial simulations of harvest schedules developed through goal programming methods to establish and maintain an even-age class distribution of aspen within habitat types under 3 different management scenarios that MDNR managers and planners were most interested in to maintain forest and wildlife resources over time. The results of the analysis revealed the spatial feasibility of implementing harvest schedules, and the most pronounced effects of forest management on wildlife and timber under each scenario and throughout space and time. Feasibility of implementing harvest schedules developed from goal programming methods Timber harvesting schedules are obviously important to identify the best strategies for meeting specific objectives. Linear programming has typically been used to develop harvesting schedules that maximize values of a specific goal such as volume or revenue, or minimize costs, but may present problems when trying to find solutions to multiple goals. Goal pro gramming was derived from linear programming methods to 290 find solutions that minimize deviations from meeting specified goals (Buongiomo and Gilless 2003). Foresters have developed harvest scheduling plans using tools such as Spectrum (USDA Forest Service 1995) and FORPLAN (Johnson et al. 1996). Although those tools are useful for constructing linear-based programming models that can help evaluate alternative management strategies, they may be complicated to use. Excel Solver is a Microsoft Excel Add-in tool that can easily be used within a spreadsheet framework for constructing linear-based programming models. Solver has not typically been used to solve goal programming problems, but it has been shown to produce solutions consistent with other optimization software with the advantage of being easily displayed in a spreadsheet (Steiguer 2000). The Solver Add-in tool limits the size of the problems that can be solved by the maximum number of decision variables (n = 200) and constraints (11 = 100) that can be specified. I had 1350 decision variables (i.e., number of acres harvested in 5 age classes, 9 habitat types, and 10 decades); thus I used Premium Solver, which allows up to 2000 decision variables. Gustafson et al. (2006) demonstrated a process of linking linear programming models and spatial simulation models to predict spatial and temporal effects of forest management. In this chapter, I used a similar procedure and determined that the harvest schedule from the goal-programming model was spatially feasible, and that HARVEST allowed assessment of the variability associated with implementing the harvest schedule. This procedure can easily and cost-effectively be adopted by managers and planners who are interested in evaluating consequences of alternative management strategies. 291 Spatial and temporal dlfi'erences in availability and distribution of forest resources The most obvious effect of the forest management scenarios was the change in spatial and temporal availability and distribution of forest resources. Under all management scenarios, representation of area in 50—70-year-old aspen increased throughout space and time (Appendix 1). In management scenario 1 (i.e., harvest of all aspen by 80 years) and scenario 2 (i.e., harvest of all aspen 2 60 years), representation of aspen in the 50- and 60-year age classes reached 12.5% by the fifth decade in most habitat types. Management strategies under scenarios 1 and 2 were sustainable over the long-term. In the intensive aspen management scenario 3, however, representation of aspen 2 40 years old decreased and in some instances, became non-existent by decade 3 or 4. To maintain any harvest and have representation of aspen in all age classes under this management scenario, the amount of aspen in age classes < 40 would have to be greater than 12.5%, and the amount of aspen in age classes > 40 would have to be less than 12.5%. In management scenarios 1 and 2, representation of younger aspen age classes (< 30) decreased to allow for increased representation in older age classes; however, in scenario 3, representation of young aspen generally increased throughout time, except in decade 10. The main differences between scenarios 1 and 2 included the infeasibility of establishing a relatively even distribution of aspen area within all age classes of 6 habitat types under scenario 2, and the length of time required to establish a relatively even-age class distribution in habitat types where it was feasible (i.e., 5—8 decades under scenario 1, and 7—10 decades under scenario 2). 292 Comparison of effects of harvesting strategies on wildlife Changes in resource availability under scenarios 1 and 2 were expected to benefit 5 of 6 avian communities associated with the ATD, AVO, and TM habitat types, whereas under scenario 3 habitat availability increased for only 4 communities and the magnitude of the increases were lower (Figures 4.10, 4.19, 4,28). Scenarios 1 and 2 benefited avian communities more than scenario 3 because the establishment of an even-age class distribution increases diversity (i.e., one age class does not dominate the area). Under scenario 3, representation of some age classes within habitat types was lost. Thus, the amount of potential habitat area for avian species was also lost, unless the avian species could attain life requisites from other vegetation types available in the landscape. The association between avian communities and other vegetation types should be further investigated. For ruffed grouse, populations would likely benefit from conditions produced in decades 3—5 under scenario 3 when the availability of young aspen increased; however, overall availability of selected drumming habitat will likely decline by decade 5 under scenarios 1 and 2. These results demonstrate some of the trade-offs that managers may face when implementing activities to achieve an even-age class distribution within the aspen vegetation type. Evaluation of the spatial and temporal consequences of the different scenarios revealed areas that may be maintained for different avian communities or for habitat management for specific species. For instance, ruffed grouse is a popular game bird and has exhibited declines in population numbers due to declines the availability of aspen (Rusch et a1. 2000). Across the study area landscape, total aspen availability declined by 293 approximately 18% (i.e., 1.8% per decade) over the simulated timeframe due to successional conversions in hydric habitat types and in other habitat types on non- industrial private forests. Thus, most of the aspen declined on non-industrial private forests (Figure 4.32). This decline is less than trends in aspen decline observed in Michigan between 1980 and 1993 (i.e., 1% per year; Cleland et al. 2001), but declines were probably not evenly distributed throughout the state. Although total aspen area on state lands remained relatively constant under all scenarios, ruffed grouse and other species that depend on aspen may exhibit population declines because of aspen declines on adjacent land-ownerships. Thus, to meet stakeholder demands for ruffed grouse management on public lands, it may be beneficial to identify areas that might be most effective in benefitting ruffed grouse populations. Under all 3 management scenarios, an area in the central part of the study area consistently produced sites that were considered highly selected by ruffed grouse (Appendix 2) because they were located in habitat types that support aspen types providing the necessary structure and composition for drumming (see Chapter 3). If maintaining ruffed grouse habitat is a management priority for the MDNR, areas with high potential for providing habitat should be considered for management. Perhaps more intensive management could occur on those sites to increase representation of aspen age classes selected by ruffed grouse. Similarly, habitat availability for other species of concern can also be monitored and evaluated. For instance, the golden-winged warbler, pine siskin, red-headed woodpecker, and wood thrush are listed on the Audubon Society’s Watch List as sensitive species (Brewer et al. 2001, Audubon Society 2002). These species were 294 associated with aspen in the 50 year age class of ATD and AVO habitat types, 70-year class within the TM habitat type, 50-year age class in all habitat types, and 20-year class within the TM habitat type, respectively (see Chapter 2). Areas potentially providing habitat for the golden-winged warbler and the red-headed woodpecker are currently scattered throughout the study area and increase throughout time (Appendix 3). Areas potentially providing habitat for the pine siskin and wood thrush are located in the central part of the study area. Potential habitat for the pine siskin increased, while habitat decreased for the wood thrush throughout the simulation due to a decrease in the amount of aspen in the 20-year age class within the TM habitat type (Appendix 3). Adaptive management The results of this project can effectively be incorporated into an adaptive management framework with the recognition that adopted management approaches are working hypotheses and, therefore, should be continually monitored and evaluated (Lindenmayer and Franklin 2002) to assess the actual effectiveness and sustainability of the management program. Hypotheses are what we consider to be the most likely explanation of a particular event (Ratti and Garton 1996). The hypotheses generated from this project included: 1) it is feasible to establish and maintain an even-age class distribution within the aspen vegetation type under scenarios 1 and 2; 2) the harvesting schedule developed is spatially feasible given actual harvesting constraints such as not partially cutting existing stands; 3) the association between avian communities and different aspen age classes and/or habitat types is true; 4) ruffed grouse habitat selection relationships reflect how grouse are actually selecting drumming habitat; and 4) grouse 295 populations will respond to changes in habitat availability according to the model developed. Although evidence from this research project supports the stated relationships (with some ecological assumptions), there will always be some unpredictability and uncertainty in managed ecosystems that may cause slightly different outcomes (Davis et al. 2001) such as natural disturbances or changes in land ownership. Therefore, continued research is necessary to test these hypotheses. The success of management for sustainability will depend on continuing strategic research to build on current forest management strategies that are likely to achieve sustainability, such as harvesting aspen by 80 years old or harvesting all aspen > 60 years old, and establishing representation of aspen in several age classes and habitat types. 296 .300qu 8035 35305-8: 08 80am haw M80 05 .09: 0.00qu 320888 0a 0003 3mm Em: .65— 806 000065 00000 823 dined—0m 00gb E0603 05 E 00.8 m8 com 0 5 0000000 3 Us: 05 E commmooosm 3 03. no. on 8 @800on mm momma 20:3 9003 E BUSEEV 0002 .Nmé 0SmE 297 CHAPTER 5 MANAGEMENT IMPLICATIONS In forest management, current structure, composition, spatial arrangement, and availability of forest types in a given area is a consequence of all the preceding management actions occurring in that area. This is essentially referred to as cumulative effects of management. Cumulative effects of management should not be ignored because they explain the pattern, availability, and structure of forest types occuning in the landscape, the richness, diversity, and occurrence of species within communities, and the habitat conditions and viability of individual species. Assessing cumulative effects of management is critical for developing future plans to meet a variety of ecological, social, and economic objectives. Within many state and federal agencies, forest management goals have shifted from timber production toward sustainable management to meet those broad objectives (Kessler et al. 1992). Sustainable management is not a new concept. In fact, in the early part of the 20th century, Gifford Pinchot wrote that national forests were to be managed for the permanent good of all people and not for the temporary benefit of individuals or companies (Davis et al. 2001). This concept was put into practice for management of national forests by the Multiple Use-Sustainable Yield Act of 1960. Later state agencies began to adopt this concept and by the 19808 and early 1990s, the public had an increased awareness of the impacts of land use activities on wildlife. Thus, state and federal agencies and some private landowners began to develop plans to manage forests as 298 naturally functioning ecosystems to meet human and wildlife demands for forest resources (Davis et al. 2001). In the early 1990s, the MDNR undertook the task of initiating ecosystem management on state forest lands (Begalle 1991, as cited in Hanaburgh 2001), and in 2004, the Sustainable Forestry Act was signed by the Governor requiring that the MDNR shall seek and maintain forest certification. This research can be used, in part, to fulfill requirements for forest certification through the Forest Stewardship Council (F SC) and Sustainable Forestry Initiative (SFI). In 2005, the MDNR’s Forest Certification Implementation Team evaluated current plans and identified several shortcomings of meeting standards of the FSC and SFI. Areas needing emphasis included assessment of cover-type and wildlife habitat representation, strategic and long-term landscape-level planning, and development of programs to promote biodiversity conservation (Forest Certification Implementation Team 2005). Several aspects of this research project can help address those shortcomings or emphasis areas. In Chapter 1, structural and compositional characteristics of different age classes within aspen vegetation types were linked with an ecological classification system (i.e., habitat types) to understand potential spatial and temporal distributions of vegetation characteristics within the western Upper Peninsula landscape. I identified specific aspen age classes within habitat types that are currently not well represented in the landscape, thus potentially decreasing habitat availability for wildlife species or communities of species. For instance, 7% of all aspen is between 50 and 60 years old, and 50—60-year- old aspen within the ATD habitat type is virtually nonexistent. 299 Results of this project provided evidence that all aspen stands are not equivalent in terms of structure, composition, value to wildlife, and timber potential. The different abiotic conditions (e. g., soil moisture, texture, and nutrient regimes) associated with different habitat types are the underlying factors contributing to the variations in vegetation physiognomy associated with aspen vegetation types observed throughout the western region of the Upper Peninsula. Thus, aspen should not be managed as one vegetation type; habitat types should be considered in management. If aspen is not viewed as a diverse resource and is managed the same regardless of habitat type, plans for providing wildlife habitat may not be valid. For instance, in Chapter 2, results suggested that although several avian species use aspen, specific community associations were evident when habitat types were considered. If managers planned to provide aspen in a landscape, habitat requirements for some species may not be accounted for because life requisites occur in aspen associated with a particular habitat type. The pine siskin, for example, was not observed in aspen stands without a conifer component (see Chapter 2). The golden-winged warbler was not observed in aspen stands with a conifer component (see Chapter 2). It should be noted that the associations between various wildlife species and communities reported in this document were based on the availability and spatial arrangement of vegetation types that occurred in 2004—2006. These associations may change under different management scenarios if the availability or arrangement of vegetation types change. For example, suppose bird species X was observed in a particular habitat type and aspen age class only because its preferred vegetation type, was limiting or not available in the landscape. Once that preferred vegetation type becomes 300 available, its community association could change. Therefore, wildlife associations within various vegetation age classes and habitat types should be monitored throughout time and as vegetation distribution and availability changes. In Chapter 2, I identified avian community associations with specific aspen age classes and/or habitat types and described threshold conditions for vegetation structure and composition that were important for determining species presence. Because presence of avian species is now associated with aspen age and habitat type, managers can predict future changes in habitat availability for avian species or communities under different management activities that may change representation of aspen within specific age classes and habitat types. Increasing the representation of aspen in the 50-year age class in all habitat types will benefit avian communities. Results of Chapter 3 demonstrated that ruffed grouse respond differently to the various ecological conditions that support aspen, and thus aspen age classes and habitat types are important factors in grouse selection of drumming sites. Knowledge of this information will help managers identify and maintain areas highly selected by ruffed grouse and understand how grouse may respond to forest management activities or successional changes that may alter representation or spatial arrangement of highly selected vegetation types. For instance, grouse populations may likely increase if managers increased the amount of young (< 30-year-old) aspen within the TM, ATD, AVO, and AOC habitat types, the amount of aspen represented in all age classes within the TM type, or old aspen in the TM, AVO, and AOC habitat types . In Chapter 4, I integrated information from Chapters 1-3 to evaluate spatial and temporal outcomes of meeting specific MDNR forest management goals to maintain 301 wildlife habitat and timber production. The timber harvest scheduling model and results of the spatial simulations can help managers develop strategic long-term and landscape- level planning and predict spatial and temporal effects on long-term sustained timber yield, wildlife habitat availability, and potential wildlife population responses (e.g., for ruffed grouse). The modeling process has great value for evaluating trade-offs between maintaining habitat for individual wildlife species, communities, and timber yield under alternative scenarios (Table 5.1). For example, to maintain sustainable timber harvest, an even-age class distribution should be established, but a trade-off is a decrease in the amount of highly selected ruffed grouse drumming habitat. Scenario 1 (i.e., harvest all aspen by 80 years) appeared the best of the 3 scenarios described to provide a diversity of vegetation conditions for avian species and sustain timber. The process described in Chapter 4 should be used to evaluate consequences of other management scenarios that are also being considered. Further research should also quantify the magnitude of differences in trade-offs among alternative scenarios. The MDNR has adopted the practice of ecosystem management and forest certification because they are based on the premise that functioning, sustainable ecosystems will consequently conserve biodiversity (Grumbine 1994). While this project specifically evaluated responses of avian communities and ruffed grouse to vegetation differences within aspen age classes and habitat types, mechanisms illustrated in this project can be expanded to address specific measures of biodiversity conservation. For instance, an important aspect of biodiversity conservation is maintaining adequate ecological representation of cover types (Haufler et al. 1996). The MDNR’s goal of establishing and maintaining an even-age class distribution of aspen within all 302 represented habitat types is a start. Results of this project indicated this is a feasible goal, but it will take many decades (> 5) to establish an even-age class within all most habitat types. Identification of performance measures such as presence of sensitive species, species richness, species diversity, and landscape diversity can be used to monitor how well the plan to establish and maintain an even age class distribution of aspen is achieving biodiversity objectives. The maps resulting from timber harvest simulations can be used to evaluate representation of habitat for other wildlife species over time because they reveal changes in spatial arrangement and availability of forest types, which are important landscape-level variables for evaluating wildlife habitat. The modeling process can be used to simulate effects of other management scenarios in different forest types as well. The cumulative effects of management decisions are what determine how well long-term forest management goals are being met. As the public becomes more aware of environmental issues, and consequently desires agencies to become more accountable for sustainable management on public lands, it becomes increasingly important to be able to predict long-term outcomes of forest management decisions. Although there are limitations in use of any model because of uncertainty and variability in data or assumptions made, the modeling approach used in this project provides significant benefits. It provides evidence that aspen should not be managed as a single forest type, and it provides quantitative answers to questions such as how much aspen to cut, where can it be cut, and what are the effects on timber production and wildlife at multiple levels of biological organization. With increasing demands on natural resources, answers to 303 such questions are essential for agencies and other landowners to sustain timber and wildlife resources for current and future generations. 304 Table 5.1. Comparison of consequences of different aspen management strategies simulated for a 300 mi2 area in the western Upper Peninsula, Michigan. Strategies included: harvest all aspen by 80 years, harvest all aspen > 60 years, and intensively harvest aspen (i.e., > 60% of volume harvested on a 40-year rotation). Harvest by 80 Harvest > 60 Intensive Metric Timeline Consequence Consequence Consequence , _ . , Decade 2 -3% -l3% -8% Habrtat drversrty for avran Decade 5 10% 7% _5% commumnes Decade 10 10% 10% -9% Amount of highly selected Decade 2 41% 49% “10% ruffed grouse drumming Decade 5 ‘6% '4°/0 4% habitat Decade 10 -6% -6% -7% Decade 2 -32% -3 8% -8% Amount of young aspen (< 30) Decade 5 “36% 36% 10% Decade 10 -36% -38% -29% Decade 2 122% 160% 71% Amount of med-aged aspen Decade 5 41% 38% 28% (3°49) Decade 10 41% 43% 57% Decade 2 13% -33% -37% Amount of old aspen (> 50) Decade 5 61% 67% '53°/0 Decade 10 61% 70% 30% Decade 2 -30% -26% -24% Timber volume Decade 5 -23°/o -14% -29% Decade 10 -14% -16% -44% Decade 2 -l4% -l6% -7% Amount harvested DCCfldC 5 -]4% '12% -4% Decade 10 -l4% -l4% -28% Decade 2 2% 3% -4% Interior forest Decade 5 2% 4% 42% Decade 10 2% 2% 0% Decade 2 -2% -l% -3% Edge Decade 5 0% 0% -3% Decade 10 -2% -2% -3% 305 APPENDIX 1 The following figures display the proportion of area within 8 aspen age classes in 9 different habitat types resulting from harvest schedules from 3 different aspen management scenarios. The first scenario was to harvest all aspen by 80 years. Scenario 2 was to harvest only aspen > 60 years old, and the third scenario was to harvest aspen intensively (i.e., 60% of harvest came from aspen in the 40-year age class). The goal for all scenarios was to establish and maintain an even age class distribution of aspen in all habitat types. 306 —._.__-.--—-- chem\\\\x\\k\\\3 QAE burst __---.-_-.-_- IIIIIIIIIII-ll uyuyoyyuooaooyvy ...-lll-llllll- ...-...-.-.-... null-llllllllll annananononannnn ...—_________.__ 5¢€V\539Q<>¢Q 030 025 020 0.15 010 005 000 AQV uuvyyyvyv7/7vuuyuyv III-lllllll-llll-III .._.—.._.....__..._ .\\\\\\\\\ \\\\\\\\\\ __-.-.-.-—---_-_ IIIIIIIII-l-IIII- n997cco74974992 III-lllllI-llllll-l .............._.... \\\\\\\<\\\\V\\\\\\ 020 3 cow—89E Decade la of rnsu All aspen within different habitat types in the Upper Pen' Michigan at the beginning of 10 decades follo was harvested by 80 years old. AQV QAE Proportion of aspen Figure A] lated aspen harvest. wrng srmu A cer-Quercus— Vacc ’ imum-spp. Pinus- Vacc 9 mzum igaea. Quercus-A cer-Ep 307 TM 0.35 \\\ x. .\ AVO 0.30 .._-.222... \\\\\\\\\\\\\\\ -:-=----. Ill. 1’!///////////1 l I>7D 00.0000 00000.0 no.5 ..o :oanoE Decade ula of Michigan at the beginning of 10 decades following simulated aspen harvest. All aspen Figure A.2. Proportion of aspen within different habitat types in the Upper Pe Viola- Acer- Acer-Tsuga-Dryopteris; AVO Tsuga-Maianthemum. was harvested by 80 years old. ATD Osmorhiza; TM 308 , .IE 11.5: _ .a 7///////30(V \\\\\\\\\\\\\\\\\\\\ aunl-III-uuuun! =--=---==- llllIllllllIllII le/IllrI/I/rllzau 0. 0 0.25 0.0.000 0 «Eu be 55.30.... «20 we :oEonoE SPP-; QAE mtum— the Upper Peninsula of m ing of 10 decades following simulated aspen harvest. Only aspen Pinus- Vacc Decade in different habitat types inium 310 Acer-Quercus- Vacc ron of aspen wi Proport' 4 Michigan at the beginn 60 was harvested. AQV Quercus-A cer-Epigaea. Figure A Z TM .\\\\\\\\\\\\\.\\\\\N :-:-==-=E l I>70 10 TM Acer- Viola-Osmorhiza; AVO Decade Acer-Tsuga-Dryopteris; AVO 31 1 Start- 0.30 Figure A.5. Proportion of aspen within different habitat types in the Upper Peninsula of Tsuga-Maianthemum. Michigan at the beginning of 10 decades following simulated aspen harvest. Only aspen > 60 was harvested. ATD AOC TMC ___-I-uunnnn-IE lung-II //////I/////. I/. Decade Figure A.6. Proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest. Only aspen Fraxinus-spp.; TMC F: -Caulophyllum; Acer-Osmorhiza Tsuga-Maianthemum-Coptis. 60 was harvested. AOC Z 312 Decade in the Upper Peninsula of itat types Figure A.7. Proportion of aspen within different hab Michigan at the beginning of 10 decades following simulated aspen harvest. Most (60%) of the aspen was harvested on a 40-year rotation. AQV Pinus- Vaccinium Quercus- Vaccinium; Acer- Quercus-A cer-Epigaea. ,QAE= -5pp-' 313 TM 3 130:9 III/Itl/lef/lzfle. I AT D 100977 ' lune ‘ tum-29} 14049; ‘asose‘ «960691 ‘ I>70 0.60 77% ‘5“ E 7 , wmmm 0000 mm 00 020 “_o :oEonEm Decade Figure A.8. Proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest. Most (60%) of the aspen was harvested on a 40-year rotation. ATD Tsuga-Dryopteris; AVO Acer- TM = Tsuga-Maianthemum. Acer- Viola-Osmorhiza; 314 AOC Decade Acer-Osmorhiza- Tsuga—Maianthemum-Coptis. spp.; TMC = F raxinus- F: 315 000.0000. 000000 0 0 0 0 0 0 0 0 0 0. 0 0 0 020 .0 5.0.0005 «20 .0 005.302.... 09.0 M.0 005.0005 Figure A.9. Proportion of aspen within different habitat types in the Upper Peninsula of Michigan at the beginning of 10 decades following simulated aspen harvest. Most (60%) of the aspen was harvested on a 40-year rotation. AOC C aulophyllum; APPENDIX 2 The following appendix shows the spatial and temporal changes in distribution of highly selected and avoided ruffed grouse drumming habitat under different simulated management scenarios for a 300 mi2 area in the western Upper Peninsula, Michigan. Harvest was simulated for 10 decades. Selection values > 1 indicate selection, 1 = neither selection or avoidance, and < 1 indicates avoidance. For all figures, the following key applies: Highly selected (2.30) Neutral 1 .00 Avoided 0.00 316 Initial conditions Scenario 1: harvest all aspen by 80 years Decade 2 ,. - J 317 Scenario 1: harvest all aspen by 80 years Decade 5 Scenario 1: harvest all aspen by 80 years Decade 10 wig: ““I. l‘ ’d 23": v. ..,~ 318 Scenario 2: harvest all aspen > 60 years Decade 2 Scenario 2: harvest all aspen > 60 years Decade 5 319 Scenario 2: harvest all aspen > 60 years Decade lO Scenario 3: intensive aspen management. At least 60% harvested from 40-year age class. Decade 2 320 Scenario 3: intensive aspen management. At least 60% harvested from 40-year age class. Decade 5 Scenario 3: intensive aspen management. At least 60% harvested from 40-year age class. Decade 10 321 APPENDIX 3 The following appendix shows the spatial and temporal changes in avian habitat distribution under different simulated management scenarios for a 300 mi2 area in the western Upper Peninsula, Michigan. Harvest was simulated for 10 decades. For all maps, the following color scheme applies: Color code Community association aspen in the 20—50—year age class in ATD and AVO aspen in the 50-year age class old (> 60 years) aspen in ATD and AVO 0. 4’5 1.1 aspen in the 50-year age class within TM old (> 70) aspen in within TM aspen in the 20-year age class within TM TIN] Black-throated blue warbler Blue jay Chestnut-sided warbler Red-breasted nuthatch Eastem-wild turkey Red-eyed vireo Northern flicker Red-headed woodpecker Rose-breasted grosbeak White-breasted nuthatch American redstart Eastern peewee Golden-winged warbler Scarlet tanager Veery _ Black-and-white warbler Wood thrush Brown creeper Yellow-bellied sapsucker Pine siskin A Yellow-rumped warbler _ Golden-crowned kinglet White—throated sparrow , Winter wren 322 Location of study area within Baraga, Dickinson, Iron, and Marquette Counties. 323 Scenario 1: Harvest all aspen by 80 years old to establish and maintain an even age class distribution. A) Initial conditions J 324 C) Decade 5 D) Decade 10 Scenario 2: Harvest all aspen > 60 years old to establish and maintain an even age-class distribution. L E) Initial conditions —I F) Decade 2 L G) Decade 5 Scenario 3: Harvest 60% of the aspen from the 40-year age class. I) Initial conditions J) Decade 2 328 T L K) Decade 5 329 REFERENCES Audubon Society. 2002. Audubon watch list 2002. Audubon Society. http://www.audubon.org/bird/watch. Accessed August 25, 2007. Bertin, R. I. 1977. Breeding habitats of the wood thrush and veery. Condor 79: 303—311. Blake, J. G., J. M. Hanowski, G. J. Niemi, and P. T. Collins. 1994. Annual variation in bird populations of mixed conifer—northern hardwood forests. The Condor 96:3 8 1—3 99. Bissonette, J. and I. Storch. 2007. Temporal dimensions of landscape ecology: wildlife responses to variable resources. Springer, New York, New York, USA. Bond, R. R. 1957. Ecological distribution of breeding birds in upland forests of southern Wisconsin. Ecological Monographs 27: 351—384. Brinkman, K. A., and E. I. Roe. 1975. Quaking aspen: silvics and management in the Lake States. US. Department of Agriculture Handbook 486. Washington, D. C. 52pp. Brewer, R., G. A. McPeek, and R. J. Adams, Jr. 1991. The atlas of breeding birds of Michigan. Michigan State University Press, East Lansing, Michigan, USA. Bump, G. R., W. Darrow, F. C Edminster, and W. F. Crissey. 1947. The ruffed grouse: life history, propagation, management. New York State Conservation Department, Holling Press, Buffalo, New York, USA. Bumgiomo, J. and J. K. Gilless. 2003. Decision methods for forest resource management. Academic Press, Burlington, Massachusetts, USA. Cade, B. S. and P. J. Sousa. 1985. Habitat suitability index models: Ruffed grouse. US. Fish and Wildlife Service Biological Report 82 (10.86). Campa, H, 111., J. B. Haufler, and D. E. Beyer, Jr. 1992. Effects of simulated ungulate browsing on aspen characteristics and nutritional qualities. Journal of Wildlife Management 56: 158—164. Canfield, R. H. 1941. Application of the line intercept method in sampling range vegetation. Journal of Forestry 39:388—394. Carey, A. B. and R. 0. Curtis. 1996. Conservation of biodiversity: a useful paradigm for forest ecosystem management. Wildlife Society Bulletin 24: 610—620. 330 Cleland, D. T., L. A. Leefers, and D. I. Dickmann. 2001. Ecology and management of aspen: a Lake States perspective. Pages 81—99 in USDA Forest Service Proceedings RMRS-P-18. Coffrnan, M. S., E. Alyanak, and J. Ferris. 1980. Field guide to habitat type classification system for Upper Peninsula of Michigan. School of Forestry and Wood Products, Michigan Technological University, Houghton, Michigan, USA. Confer, J. L. and K. Knapp. 1981. Golden-winged warblers and blue-winged warblers: the relative success of a habitat specialist and a generalist. Auk 98: 108—114. Crawford, B. T. 1950. Some specific relationships between soils and wildlife. Journal of Wildlife Management 14: 115—121. Crow, T. R., and E. J. Gustafson. 1997. Ecosystem management: managing natural resources in time and space. Pages 424—450 In K. A. Kilm and J. F. Franklin (eds.), Creating a forestry for the twenty-first century, Island Press, Washington, D. C., USA. Daubenmire, R. 1966. Vegetation: Identification of typal communities. Science 151:291—298. Davis, L. S., K. N. Johnson, P. S. Bettinger, and T. E. Howard. 2001. Forest Management, Fourth Edition. McGraw Hill, New York, New York, USA. DeStefano, S. 2002. Regional and national issues for forest wildlife research and management. Forest Science 48:181—189. DeStefano, S. and D. H. Rusch. 1984. Characteristics of ruffed grouse (Bonasa umbellus) drumming sites in northeastern Wisconsin, USA. Transactions of the Wisconsin Academy of Sciences, Arts, and Letters 72: 177-182. Dettrners, R., D. A. Buehler, and K. E. Franzreb. 2002. Testing habitat-relationship models for forest birds of the southeastern United States. Journal of Wildlife Management 66: 717—424. Doepker, R., L. Evison, D. Wagner, and D. Wilcox. 2002. Aspen management guidelines for Dickinson and Iron counties. Michigan Department of Natural Resources, Wildlife Division, unpublished report on file at the Norway Field Office, Norway, Michigan, USA. Evers, D. C. 1991. Michigan bird survey, summer 1990. Jack Pine Warbler 1: 21—26. Felix, A. B. 2003. Development of landscape-scale models to describe habitat potential of white-tailed deer (Odocoileus virginianus) in Michigan. M. S. Thesis. 331 Department of Fisheries and Wildlife, Michigan State University, East Lansing, USA. Felix, A. B., D. W. Linden, and H. Campa, III. 2007. Building and using habitat models for assessing temporal changes in forest ecosystems. Pages 195—217 In J. A. Bissonette and I. Storch, eds. Temporal dimensions of landscape ecology: wildlife responses to variable resources. Springer, New York, New York, USA. Forest Certification Implementation Team. 2005. Michigan Department of Natural Resources Forest Certification Initiative. Action plan to address identified major non-conformances with standards of the Forest Stewardship Council (FSC) and Sustainable Forestry Initiative (SFI). On file with the Michigan Department of Natural Resources, Lansing, Michigan, USA. Forest Stewardship Counci. 2005. Revised final regional forest stewardship standards for the Lake States-Central Hardwood region, USA. Forest Stwwardship Council Version LS V 3.0. Available at: lrfittp://www.fscus.org/images/documents/Z006 standards/ls__3.0 NTC._pdf Franzreb, K. E. 1985. Foraging ecology of brown creepers in a mixed coniferous forest. Journal of Field Ornithology 56: 9—16. Frawley, V. R., J. Kleitch, and C. A. Stewart. 2006. Ruffed grouse and American woodcock status in Michigan, 2005. Michigan Department of Natural Resources Wildlfie Report No. 3453, Lansing, Michigan, USA. Frelich, L. E. and C. G. Lorimer. 1991. Natural disturbance regimes in hemlock- hardwood forests of the upper Great Lakes region. Ecological Monographs 61: 145—164. Garshelis, D. L. 2000. Delusions in habitat evaluation: measuring use, selection, and importance. Pages 111—153 In L. Boitani and T. K. Fuller (Eds). Research techniques in animal ecology. Controversies and consequences. Columbia University press, New York, New York, USA. Graham, S. A., R. P. Harrison, Jr., and C. E. Westell, Jr. 1963. Aspens, phoenix trees of the Great Lakes Region. University of Michigan Press, Ann Arbor, Michigan, USA. Gullion, G. W. 1970. Factors influencing ruffed grouse populations. Transactions of the North American Wildlife and Natural Resources Conferences 35: 93-105. Gullion, G. W. 1977. Forest manipulation for ruffed grouse. Transactions of the North American Wildlife and Natural Resources Conference 42: 449—458. 332 Gullion, G. W. and F. J. Svoboda. 1972. The basic habitat resource for ruffed grouse. Pages 113—1 19 in Aspen Symposium Proceedings. US Department of Agriculture, forest Service General Technical Report NC-l. US Department of Agriculture, Forest Service, North Central Research Station, St. Paul, Minnesota, USA. Gustafson, E. J. 1997. HARVEST: a timber harvest allocation model for simulation of management alternatives. Pages 277—282 in Proceedings of the 1996 Society of American Foresters National Convention, Albuquerque, NM, USA. Gustafson, E. J ., D. E. Lytle, R. Swaty, C. Loehle. 2007. Simulating the cumulative effects of multiple forest management strategies on landscape measures of forest sustainability. Landscape Ecology 22: 141—156. Gustafson, E. J. and C. V. Rasmussen. 2002. HARVEST for Windows. Timber harvest simulation model version 6.0. US Department of Agriculture, Forest Service, North Central Research Station, Rhinelander, Wisconsin, USA. Gustafson, E. J. and C. V. Rasmussen. 2005. HARVEST for Windows v6.1: User’s guide. Published on the Internet by the USDA Forest Service, North Central Research Station, St. Paul, MN. Available at http://www.ncrs.fs.fed.us/4153/harvest/v61/documentation/default.asp Gustafson, E. J ., L. J. Roberts, and L. A. Leefers. 2006. Linking linear programming and spatial simulation models to predict landscape effects of forest management alternatives. Journal of Environmental Management. 81: 339—350. Hagan, J. M. and A. M. Meehan. 2002. The effectiveness of stand-level and landscape- level variables for explaining bird occurrence in an industrial forest. Forest Science 48: 231—242. Hagan, J. M., P. S. McKinley, A. L. Meehan, and S. L. Grove. 1997. Diversity and abundance of land birds in a northeastern industrial forest. Journal of Wildlife Management 61: 718—735. Hamelin, P. L. 1992. Wildlife response to whole tree harvesting of aspen: a ten year analysis. M. S. Thesis, Michigan State University, Department of Fisheries and Wildlife, East Lansing, Michigan, USA. Hammil, J. H. and R. J. Moran. 1986. A habitat model for ruffed grouse in Michigan. Pages 15-18 in J. Vemer, M. L. Morrison, and C. J. Ralph (eds). Wildlife 2000: Modeling habitat relationships of terrestrial vertebrate. The University of Wisconsin Press, Madison, Wisconsin, USA. Hanaburgh, C. 2001. Modeling the effects of management approaches on forest and wildlife resources in northern hardwood forests. Ph.D. Dissertation. Department 333 of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, USA Haufler, J. B. 1990. Static management of forest ecosystems. Pages 123—130 in J. M. Sweeney, ed. Management of dynamic ecosystems. Proceedings of a Symposium held at the 51St Midwest Fish and Wildlife Conference, North Central Section, The Wildlife Society, West Lafayette, IN, USA. Haufler, J. B., R. K. Baydack, H. Campa, III, B. J. Kemohan, C. Miller, L. J. O’Neil, and L. Waits. 2002. Performance measures for ecosystem management and ecological sustainability. Wildlife Society Technical Review. 02—1. 33pp. Haufler, J. B., C. A. Mehl, and G. J. Roloff. 1996. Using a coarse-filter approach with species assessment for ecosystem management. Wildlife Society Bulletin 24: 200—208. Heinen, J. T. and T. L. Sharik. 1989. The influence of mammalian browsing on tree growth and mortality in Pigeon River State Forest, Michigan. American Midland Naturalist 123: 202—206. Hobson, K. A. and E. Bayne. 2000a. The effects of stand age on avian communities in aspen-dominated forests of central Saskatchewan, Canada. Forest Ecology and Management 136: 121—134. Hobson, K. A. and E. Bayne. 2000b. Breeding bird communities in boreal forest of western Canada: consequences of “unmixing” the mixed woods. The Condor 102: 759—769. Hunter, M. L., Jr. 1990. Wildlife, forests, and forestry: principles for managing forests for biological diversity. Prentice Hall, Engelwood Cliffs, NJ, USA. Hiittl, R. F., B. U. Schneider, and E. P. Farrell. 1000. Forests of the temperate region: gaps in knowledge and research needs. Forest Ecology and Management 132: 83— 96. J enness, J. 2005. Distance/Azimuth Tools (dist_az_tools.avx) extension for ArcView 3.x, v. 1.6. Jenness Enterprises. Available at: http://www.jennessent.com/arcview/distance__azimuth.htm. Johnson, K. N., T. W. Stuart, and S. A. Crim. 1986. FORPLAN Version 2: an overview. USDA Forest Service, Land Management Planning Systems, Washington, D. C., USA. Kessler, W. B., H. Salwasser, C. W. Cartwright, Jr., and J. A. Caplan. 1992. New perspectives for sustainable natural resources management. Ecological Applications 2: 221—225. - 334 Kidd, R. P. and M. Koelling. 1988. Aspen management in Michigan. Extension Bulletin E-1552, Michigan State University, Cooperative Extension, East Lansing, Michigan, USA. Kilham, L. 1964. The relations of breeding yellow-bellied sapsuckers to wounded birches and other trees. Auk 81: 520— 527. Koerper, G. J. and C. J. Richardson. 1979. Biomass and net annual primary production regression for Populus grandidentata on three sites in northern Lower Michigan. Canadian Journal of Forest Research 10: 92—101. Kohn, B. E. and J. J. Mooty. 1971. Summer habitat of white—tailed deer in north-central Minnesota. Journal of Wildlife Management 476—487. Kotar, J. and T. L. Burger. 2000. Field guide to forest habitat type classification for North Central Minnesota. Terra Silva Consultants, Madison, Wisconsin, USA. Kubisiak, J. F., J. C. Moulton, and K. R. McCaffery. 1980. Ruffed grouse (Bonasa umbellus) density and habitat relationships in Wisconsin, USA. Wisconsin Department of Natural Resources Technical Bulletin 117: 1—15. Leatherberry, E. C. 1994. Forest statistics for Michigan’s western Upper Peninsula unit, 1993. Resource Bulletin NC-153. US Department of Agriculture, Forest Service, North Central Research Station, St. Paul, Minnesota, USA. Lindenmayer, D. B. 1999. Future directions for biodiversity conservation in managed forests: indicator species, impact studies, and monitoring programs. Forest Ecology and Managemetn. 115: 277-287. Lindenmayer, D. B. and J. F. Franklin. 2002. Conserving forest biodiversity; a comprehensive multiscale approach. Island Press, Washington, D. C., USA. Manly, B. F. J ., L. L. McDonald, D. L. Thomas, T. L. McDonald, and W. P. Erickson. 2002. Resource Selection by Animals. Kluwer Academic Publishers, Dordrecht, The Netherlands. Maurer, B. A. 1993. Biological diversity, ecological integrity, and neotropical migrants: new perspectives for wildlife management. Pages 24—31 in D. M. Finch and P. W. Stangel, (eds.). Status and management of neotropical migratory birds. General Technical Report RM-229, Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, Fort Collins, CO, USA. McCaffery, K. R., J. E. Ashbrenner, W. A. Creed, and B. E. Kohn. 1996. Integrating forest and ruffed grouse management: A case study at the Stone Lake area. Wisconsin Department of Natural Resources Technical Bulletin 189: 1-39. \ 335 McGarigal, K., S. A. Cushman, M. C. Neel, and E. Ene. 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Computer software program produced by the authors at the University of Massachusetts, Amherst. Available at the following web site: www.umass.edu/landeco/research/fragstats/fragstatshtml Michigan Department of Natural Resources. 2007. Ecosystem-based management. Available online at www.michigan.gov/dnr/0,l 607,7-153-10366 11865-24426-- 00.html. Accessed August 26, 2007. Michigan Department of Natural Resources. 2003. IFMAP/GAP Upper Peninsula land cover [map]. 1997—2001. 1:64,000; created by Forest, Mineral, and Fire Management Division, MDNR using ERDAS Imagine 8.5. http://www.dnr.state.mi.us/spatialdatalibrary/sdlZ/land use cover/2001/IF MAP u p_landcover.htm ( 1 April 2003). Michigan Forest Products Council. 2006. About Michigan’s sustainable forestry initiative program. Michigan Forest Products Council, Lansing, Michigan, USA. Available at http://www.michiganforest.com/index.php‘?pid=43. Michigan Sustainable Forestry Initiative. 2004. Sustainable Forestry Initiative: 2005— 2009 standard. Sustainable Forestry Initiative, Lansing, Michigan, USA. Available at http://www.aboutsfb.org. Miller, R. O. 1998. Growing aspen fiber in Michigan: a rationale for producing this material in short rotation, intensively cultured plantations on abandoned agricultural land in Michigan’s Upper Peninsula. Abstract. Upper Peninsula Tree Improvement Center, Escanaba, Michigan, USA. Minnis, R., and J. B. Haufler. 1994. Habitat model-American redstart. Pages 177—188 in G. J. Roloff. Using an ecological classification system and wildlife habitat models in forest planning. Dissertation, Michigan State University, East Lansing, Michigan, USA. Morrison, D. F. 1990. Multivariate statistical methods. Third edition. McGraw-Hill, New York, New York, USA. Negri, S. J. 1995. Analysis of habitat suitability models for primary cavity-nesting birds in Michigan’s Upper Peninsula. Thesis, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, USA. Noss, R. F. 1987. From plant communities to landscapes in conservation inventories: A look at the Nature Conservancy (USA). Biological Conservation 41: 11—37. Noss, R. F. 1999. Assessing and monitoring forest biodiversity: a suggested framework and indicators. Forest Ecology and Management 115: 135—146. 336 Nyland, R. D. 2002. Silviculture concepts and applications, 2nd ed. McGraw-Hill, New York, New York, USA. Palmer, W. L. 1963. Ruffed grouse drumming sites in northern Michigan. Journal of Wildlife Management 27: 656-663. Parsons, S., K. J. Lewis, J. M. Psyllakis. 2003. Relationships between roosting habitat of bats and decay of aspen in the sub-boreal forests of British Columbia. Forest Ecology and Management 177: 559—570. Pederson, L. 2005. Michigan state forest timber harvest trends. A review of recent harvest levels and factors influencing future levels. Michigan Department of Natural Resources, Forest Management Division, Lansing, Michigan, USA. Petraborg, W. H., E. G. Wellein, and V. E. Gunvalson. 1953. Roadside drumming counts - a spring census method for ruffed grouse. Journal of Wildlife Management 17: 292—295. Pettingill, O.S., Jr. 1974. Ornithology at the University of Michigan biological station and the birds of the region. Kalamazoo Nature Center, Kalamazoo, Michigan, USA. Piva, R. J. 2003. Pulpwood production in the north-central region, 2000. Resource Bulletin NC-221. US Department of Agriculture, Forest Service, North Central Research Station. St. Paul, Minnesota, USA. Pojar, R. A. 1995. Breeding bird communities in aspen forests of the sub-boreal spruce (dk subzone) in the Prince Rupert Forest Region. Forest Management Handbook, Chapter 33. Province of British Columbia, Ministry of Forest Research Program, Victoria, B. C., Canada. Pravosudov, V., and T. Grubb. 1993. White-breasted nuthatch (Sitta carolinensis). No. 54 In The birds of North America (A. Poole and F. Gill, eds.). Academy of Natural Sciences, Philadelphia, Pennsylvania and the American Omithologists Union, Washington, D. C., USA. Ratti, J. T. and E. O Garton. 1996. Research and experimental design. Pages 1—23 in T. A. Bookhout, ed. Research and management techniques for wildlife and habitats. Fifth edition, revised. The Wildlife Society, Bethesda, Maryland. Richardson, M. and D. W. Brawning. 1995. Chestnut-sided warbler (Dendroica pennsylvanica). No. 190 In The birds of North America (A. Poole and F. Gill, eds.). Academy of Natural Sciences, Philadelphia, Pennsylvania and the American Omithologists Union, Washington, D. C., USA. 337 Robbins, C. S., J. R. Sauer, R. S. Greenberg, and S. Droege. 1989. Population declines in North American birds that migrate to the Neotropics. Proceedings of the National Academy of Sciences (USA) 86:7658-7662. Rodewald, P. G., M. J. Santiago, and A. D. Rodewald. 2005. Habitat use of breeding red- headed woodpeckers on golf courses in Ohio. Wildlife Society Bulletin 33: 448— 453. Roloff, G. J. 1994. Using an ecological classification system and wildlife habitat models in forest planning. Dissertation, Michigan State University, East Lansing, Michigan, USA. ‘1 Rusch, D. H., S. DeStefano, M. C. Reynolds, and D. Lauten. 2000. Ruffed grouse (Bonasa umbellus). The birds of North America online. A. Poole and F. Gilll (Eds). Available at http://bna.birds.comell.edu/BNA/account/ruffed_grouse/ Salt, W. R., and J. r> Salt. 1976. Birds of Alberta. Hurting Publishing, Edmonton, ‘ 4 Alberta, Canada. " xi: Schroeder, R. L. 1983a. HSI models: Pileated Woodpecker. U. S. Department of the Interior, Fish and Wildlife Service. F WS/OBS-82/ 10.39. Schroeder, R. L. 1983b. HSI models: Black-capped chickadee. U. S. Department of the Interior, Fish and Wildlife Service. FWS/OBS-82/ 10.37. Siegel, S. and R. Castellan. 1988. Nonparametric statistics for the behavioral sciences. McGraw Hill, New York, New York, USA. Siepielski, A. M., A. D. Rodewald, and R. H. Yahner. 2001. Nest-site selection and nesting success of the red-eyed vireo in central Pennsylvania. Wilson Bulletin 113: 302—307. Sommers, L. M. (editor). 1977. Atlas of Michigan, Michigan State University Press, East Lansing, Michigan, USA. Sousa, P. J. 1982. Habitat suitability index models: veery. U. S. Department of the Interior, Fish and Wildlife Service. FWS/OBS-821/ 10.22. Steams, F. 1990. Forest history and management in the northern Midwest. Pages 107— 122 in J. M. Sweeney, ed. Management of dynamic ecosystems. Proceeding sofa Symposium held at the 51St Midwest Fish and Wildlife Conference, North Central Section, The Wildlife Society, West Lafayette, Indiana, USA. Steiguer, J. E. 2000. Applying EXCEL Solver to a watershed management goal programming problem. Pages 325—329 in Proceedings of Land Stewardship in the 338 21St Century: The contributions of watershed management. USDA Forest Service, Rocky Mountain Research Station. RMRS-P-13. Fort Collins, Colorado, USA. Stelfox, J. B. (editor). 1995. Relationships between stand age, stand structure, and biodiversity in aspen mixed wood forest in Alberta. Jointly published by Alberta Environmental Center and Canadian Forest Service. Edmonton, Alberta, Canada. 308pp. Svoboda, F. J. and G. W. Gullion. 1972. Preferential use of aspen by ruffed grouse in northern Minnesota. Journal of Wildlife Management 36: 1166-1180. SYSTAT. 2002. SYSTAT for Windows, Version 10.2. SYSTAT Software, Inc. Richmond, California, USA. Titterington, R. W., H. S. Crawford, and B. N. Burgason. 1979. Songbird responses to commercial clearcutting in Maine spruce-fir forests. Journal of Wildlife Management 43: 602—609. Trani, M. K., R. T. Brooks, T. L. Schmidt, V. A. Rudis, and C. M. Gabbard. 2001. Patterns and trends of early successional forests in the eastern United States. Wildlife Society Bulletin 29: 413—424. USDA Forest Service. 1995. Spectrum User’s Guide. Ecosystem Management Analysis Center, USDA Forest Service, Fort Collins, Colorado, USA. Walkinshaw, L. H. and W. A. Dyer. 1953. Nesting of the black-throated-blue warbler in Michigan. Jack Pine Warbler 23:2—9. Westworth & Associates. 1984. Impact on wildlife of short-rotation management of boreal aspen stands. ENFOR Project P-203. D. A. Westworth & Associates Ltd., Edmonton Alberta, Canada. 148pp. Winnett-Murray, K. 1991. Veery. Pages 350-351 in R. Brewer, G. A. McPeek, and R. J. Adams, Jr. The atlas of breeding birds of Michigan. Michigan State University Press, East Lansing, Michigan, USA. Yamasaki, M., R. M. DeGraaf, and J. W. Lanier. 1999. Wildlife habitat associations in eastern hemlock-birds, smaller mammals, and forest carnivores. Proceedings: Symposium on sustainable management of hemlock ecosystems in Eastern North America. GTR-NE-267. USDA Forest Service, Durham, New Hampshire, USA. Zhang, Q., K. S. Pregitzer, and D. D. Reed. 2000. Historical changes in the Luce District of the Upper Peninsula of Michigan. American Midland Naturalist 143: 94—110. 339 Zimmerman, G. 2006. Habitat and population regulation of ruffed grouse in Minnesota. Ph.D. Dissertation, University of Minnesota, Minneapolis, Minnesota, USA. 340