LIBRARY Michigan State 4 University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. f/T DATE DUE DATE DUE DATE DUE JAN 0 7 2002 011202 ”’1 L.__E MSU Is An AIflrmetive Action/Equal Opportunity Institution encircmplnnd AN ECONOMIC COMPARISON OF BVEN- AND UNEVEN-AGED MANAGEMENT IN LAKE STATES NORTHERN HRRDWOODS 3! Jeffrey Neel Nieee A DISSERTATION Submitted to Hichigan state University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1993 Dr. Karen Potter—Witter, Major Professor ABSTRACT AN ICONONIC COMPARISON OF IVRN- AND UNEVEN-AGED MANAGEMENT IN LAKE STATES NORTHERN HARDWOODS BY Jeffrey Neal liese Current economic information is inadequate to compare the financial productivity of even- and uneven-aged management in Lake States northern hardwood forests. Forest managers need more detailed economic information on northern hardwood cutting methods from long-term replicated studies. This study summarises the results of a 40-year management study in Wisconsin and gives economic returns for 9 even- and uneven-aged northern hardwood cutting methods on good northern hardwood sites. It also considers the effects of varying cutting method on the diversity of northern hardwood tree species and future tree quality. Under most economic conditions tested, the medium to heavy basal area cuts produced the highest total economic returns; a heavy diameter- limit treatment had the highest returns when discount rates were high and price expectations were low. Even-aged management methods yielded the highest regeneration diversity, but economic returns were low to moderate. The heavy selection treatment had the highest net present values of total lumber yields and harvests. However, the medium selection treatment gave the best combination of total economic return, improved tree quality, and regeneration diversity. Forest managers have a variety of treatment options that can yield good economic returns. However, this study confirms that, in the long- term, medium or heavy basal area treatments will provide superior economic returns in northern hardwood management. DEDICATION This dissertation is dedicated to my wife, Rebecca Neal Niece, who provided me with the encouragement and support to complete this work. Her patience and love kept me going through the most difficult momenta, and helped me keep my perspective on life throughout the last four years. I also dedicate this work to my children--Andrew, Timothy, and Emily Joy--who have patiently accepted the many hours when Dad couldn‘t always be with them, even though he wanted to be. Finally, I dedicate this thesis to my own mother and father, Joan and Ben, who taught me to dream big dreams and to pursue excellence in all things. iii ACKNOWLEDGMENTS Many people have helped make the completion of this project a reality. Dr. Karen Potter-Witter, my major professor, gave me expert guidance on writing, editing and economic methodology. She was always generous with her time and resources, and her empathetic ear and expertise helped me negotiate many academic and personal hurdles. I also would like to thank my other committee members, Dr. Robert Manthy and Dr. Alan Schmid, for their helpful comments and encouragement. I am particularly indebted to the 0.8. Forest Service, North Central Forest Experiment Station for providing the funding to support my research. The Forest Service's contribution, though, extends beyond funding to include many people. Special thanks go to Dr. Nike Vasievich, Economics Project Leader at East Lansing, who gave liberally of his time and resources, providing the perfect mixture of admonition and affirmation. Terry Strong served as both research partner and friend, providing data, helpful comments, and many ”reality checks" to make sure this work applied to the real world of forest management. Thanks also to Ed Hansen for his cooperation throughout this project. Dr. David Baumgartner deserves special tribute: he was always willing to listen, guide, and give helpful yet gentle criticism. Sharon Hobrla and Bill Main helped with statistical questions and, along with Nike, were instrumental in introducing me to the world of computers and their use in forestry research. Particular thanks are due Sharon for untangling the difficult quality data set and making it ready for computer analysis. Donna Paananen assisted with editing and formatting, and provided a liberal arts perspective in the midst of a technical analysis. iv Carol Hyldahl deserves special praise for her insightful comments on the economic analysis, and the shared struggle as a graduate student. Noel Bennett provided expert secretarial support, a listening ear, and dry wit to help weather the storms. She was especially helpful and patient in finalizing the manuscript. Thanks are due Bobbe McDaniel and Michelle Laing for data entry, word processing, and organizing of early drafts. Jaci Hamberg cheerfully helped in the final process of putting the thesis together. I am grateful to Carl Reidel, Gordon Guyer, Larry Tombaugh, and Larry Leefers for their strong encouragement to pursue this degree, and faith in my ability to complete it. A final thanks goes to the many foresters and economists who have shared thoughts on this topic. Their helpful comments have sharpened my thinking and improved this manuscript. TABLE OF CONTENTS L1.t Of Twl.’ O O O O O O O O O O O O O O O O O O O O O O O L1.t Of ’19“:.. O C O O O O O O O O O O O O O O O O O O O 0 Chapter 1: Chapter 2: Chapter 3: The Problem . . . . . . . . . . . . . . . . . . Literature Review . . . . . . . . . . . . . . . History . . . . . . . . . . . . . . . . . . . . Uneven-aged Management Methods . . . . . . . . Even-aged Management Methods . . . . . . . . . Economics of Cutting Methods . . . . . . . . . Quality Considerations in Economic Management of Northern Hardwoods . . . . . . . . . . . . R‘..arCh “ethOd. O O O O O O O O O O O O O O O 0 Long-term Silvicultural Trials-~Argonne Experimental Forest . . . . . . . . . . . Data . . . . . . . . . . . . . . . . . . . . . Description of Silvicultural Cutting Treatments Plot Measurements . . . . . . . . . . . . . . . Modelling Growth and Yield . . . . . . . . . . Economic Evaluation . . . . . . . . . . . . . . Hardwood Management Costs . . . . . . . . . . . Analysis of Hardwood Price Data . . . . . . . . Quality Change and Economic Return . . . . . . Application of Hanks' (1976) Tree Grade Lumber Yield Equations . . . . . . . . . . . . . Valuation of Lumber Yields . . . . . . . . . . Overall Economic Evaluation . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . vi Chapter 4: Chapter 5: Hardwood Price Expectations . . . . . . . . . Introduction . . . . . . . . . . . . . . . . Historical Price Trends . . . . . . . . . . . Hardwood Markets and Prices--National Demand and Price Trends . . . . . . . . . . . . Export Markets . . . . . . . . . . . . . Hood Household Furniture . . . . . . . . Hardwood Price Analysis . . . . . . . . . . . Data . . . . . . . . . . . . . . . . . . Methods To Estimate Stumpage Value . . . Derivation of Estimated Price Trends . . Model Justification . . . . . . . . . . Results and Discussion, by Species . . . . . Summary . . . . . . . . . . . . . . . . . . . Northern Hardwood Management Costs . . . . . Introduction . . . . . . . . . . . . . . . . Literature Review . . . . . . . . . . . . . . Characteristics of Hardwood Management Costs Methods . . . . . . . . . . . . . . . . . . . Cost Assumptions . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . Timber Stand Improvement . . . . . . . . Post-harvest Removal of Non-commercial Trees Timber Cruise and Marking . . . . . . . Thinning Costs . . . . . . . . . . . . . Clearcut Layout . . . . . . . . . . . . Cost Variation Between Even- and Uneven-aged Methods. Application of Cost Estimates to Argonne Cutting Methods . . . . . . . . . . . . Application of Thinning and Marking Costs Application of Clearcut Layout Costs . . Application of Road Maintenance Costs . vii 43 43 44 46 47 48 50 SO 52 53 54 S7 72 73 73 73 74 76 76 77 77 77 79 80 81 82 82 82 83 83 Chapter 6: Chapter 7: Chapter 8: Sensitivity Analysis: Selection Marking, Sale Layout, and T81 Costs . . . . . . . . . . . Opportunity Costs . . . . . . . . . . . . . . . . . . Social Costs of Timber Management . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . Growth and Yield Predictions . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . Characteristics of THIGS . . . . . . . . . . . . . . Validation of Diameter-growth Predictions . . . . . . Possible Reasons for Underpredictions . . . . . . . . Diameter-growth Adjustment Methodology . . . . . . . Stand Growth Parameters for Running TWIGS . . . . . . Results After Forty Years of Management . . . . . . Results of Management Including TWIGS Projections . . Summary . . . . . . . . . . . . . . . . . . . . . . . Analysis of Trade-offs Between Tree Species Diversity and Economic Returns . . . . . . . . . . . . . . Intranet ion 0 O O O O O O O O O O O O O O O O O O 0 Diversity Measurement . . . . . . . . . . . . . . . . Results of Management Effects on Stand Diversity . . Results of Management Effects on Regeneration Div.r.ity O O O O O O O O O O O O O O O O O O 0 Economic Analysis Methods . . . . . . . . . . . . . . Economic Results . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . Uneven-aged Management . . . . . . . . . . . . . Even-aged Management . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . Hardwood Sawlog Quality: Economic Evaluation of Cutting Method Impacts . . . . . . . . . . . . . Literature Review . . . . . . . . . . . . . . . . . . D‘t‘ O O O O O O O O O O O 0 O O O O O O O O O O O 0 viii 85 85 87 88 89 89 90 90 92 93 9‘ 98 100 103 104 105 105 106 107 108 113 115 116 116 118 119 121 121 125 Chapter 9: ‘conmic mthOd. O O O O O O O O O O O O O O O O Silvicultural Results: Changes in Tree Quality and Grade . . . . . . . . . . . . . . . . . Results: Regression Analysis of Lumber Quality Ch.nq.. O O O O O O O O O O O O O O O O O O Economic Results . . . . . . . . . . . . . . . . Harvested Lumber Volumes and Cash Flows . . R‘.1dua1 v.1ue' O O O O O O O O O O O O O O Marginal Benefit/Cost Analysis of Treatment Lumber Values . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . Changes in Tree Quality . . . . . . . . . . Economic Returns to Tree Quality Changes . . Economic Analysis, Results, and Discussion . . . Results of Analysis for 1952-1992 . . . . . . . . Base Case . . . . . . . . . . . . . . . . . Sensitivity to Discount Rate . . . . . . . . Sensitivity to Changes in Marking Costs . . Results of Analysis for 1952-2022 with TWIGS Growth Projection . . . . . . . . . . . . . 3.“. ca" O O O O O O O O O O O O O O O O O Sensitivity to Discount Rate . . . . . . . . Sensitivity to Price Changes . . . . . . . . Sensitivity to Both Price Variation and Different Discount Rates . . . . . . . Sensitivity to Road Maintenance Costs . . . Sensitivity to Individual Tree Marking and Sale Layout Costs . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . Discussion: 1952-1992 Analysis . . . . . . . Discussion: Extended (1952-2022) Analysis . Discussion: Discount Rate Changes, Extended Analysis . . . . . . . . . . . sumry O O O O O O O O O O O O O O O O O O O O O ix 126 128 132 135 135 136 137 137 140 140 141 142 142 142 142 144 145 145 145 146 148 149 150 150 151 153 155 157 Chapter 10: Conclusion and Recommendations . . . . . . . . . . . 159 Future Research Needs . . . . . . . . . . . . . . . . 163 Appendix A: Summaries of TWIGS Growth Projections, by Treatment and Replication . . . . . . . . . . . . 165 Appendix B: Regression Equations Used to Calculate Lumber Yields from Graded Hardwood Logs . . . . . . . 192 Bibliogr.Phy O O O O O O O O O O O O O O O O O O O O O O O O O O O 196 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 3-1e ‘-1e 5-1. 5-2 e 5-3 e 6.10 6-2 s 6-3 e 6.4 e 6-Se 7-2e LIST OF TABLES Area of northern hardwood forest types in the Eastern United States in millions at .cr.’ O O O O O O O O O O O O O O O O O O O O Description and stand basal area of treatments in 1951 and in 1990 . . . . . . . . . Regression equations used to estimate Michigan Upper Peninsula sawtimber stumpage pric. tr.nd. O O O O O O O O O O O O O O O O O O Regression equations used to estimate Wisconsin sawtimber stumpage price trends. . . . Stumpage price scenarios used in analysis, expressed in average real price increases (decreases) per year . . . . . . . . . . . . . . Summary of cost survey results: variable costs . Average nominal road maintenance costs used in th. ‘n.ly.i. O O O O O O O O O O O O O O O O Data used to test sensitivity of treatment results to changes in marking, T81 and sale layout c°.ta O O O O O O O O O O O O O O O O O Comparison of growth predictions for three species before adjustment . . . . . . . . . . . Basswood and white ash coefficient adjustments by dObOhO c1... O O O O O O O O O O O O O O O O Comparison of THIGS diameter adjustment coefficients before and after final adjustments Sawtimber yields by treatment and year, in Scribner board feet/acre . . . . . . . . . . Pulpwood yields by treatment and year, in cords per acre . . . . . . . . . . . . . . . Species composition by percentage of total in each of three replicates prior to initial M"..t. in 1951 O O O O O O O O O O O O O O O O Shannon's diversity values and changes for trees 4.6“ d.b.h. and larger in 1951 (before and after cut) and in 1990 . . . . . . . xi Page O 2 32 55 56 S7 78 85 86 93 96 97 101 102 110 111 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 7.3 e 7-4e 8-1e 8-2. 8-3 . 8“ e 8-7e 8-8. 8.9 e 9-1e 9.2 e 9.3 e 9‘4e 9-5e 9.6 e 9.7 e Shannon's diversity values of saplings 2.0- to 4.5- d.b.h. by treatment in 1957, and in 1990 for all trees and commercial trees . . . . . . Sapling numbers in trees/acre and frequency in percent species and Lumber prices by grade lumber yields . . . . Average tree grades by treatments in 1990 . . . . . . used to evaluate treatment in 1971 and 1991, and grade increase from 1971 to 1991 . . . . . . Percent of trees by grade, 1991 . . . . . . . . . . . Regression statistics for plot analysis of variation in #1 common and better lumber . . . . . . . Analysis of variance for #1 common and better lumber . . . . Increase in #1 common and better lumber, by plot and treatment (1951-92) . . . . . . . . . . . Net harvested lumber volumes and total cash flows, 1971-92 . Residual values of lumber yields, 1992 . . . . . . . . Comparison of net present value of the differences between the control and cutting treatments for tot‘l 1m: v.1“. yi.ld O O O O O O O O O O O O O O O Net present values per acre (1990 S) of differences between treatments and control for 1952-1992 analysis Net present values per between treatments and using a 4 percent real Net present values per between treatments and using a 2 percent real Net present values per between treatments and using a 6 percent real acre (1990 S) of differences control for 1952-2022 analysis, discount rate . . . . . . . . . acre (1990 S) of differences control for 1952-2022, discount rate . . . . . . . . . acre (1990 5) of differences control for 1952-2022 analysis, discount rate . . . . . . . . . Ranking of treatment net present values using a 4 percent real discount rate (1952-2022) . . . . . . . Ranking of treatment net present values using a 6 percent real discount rate (1952-2022) . . . . . . . Ranking of treatment net present values using a 2 percent real discount rate {1952-2022) . . . . . . . xii 112 114 129 130 131 133 134 135 136 136 137 143 145 147 147 155 156 157 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 2-1. 3.3 e 3-4e 4-1e 4-2. 4-3 e 4-4 e 4-6. ‘-7e 4-8 e 4-9 e ‘-10e 4-11e LIST OF FIGURES Comparison between Arbogast and other Lake States cutting guides for recommended reserve growing stock . . . . . . . . . . . . . . Location of Argonne Experimental Forest within the Nicolet National Forest, Wisconsin . . Location of long-term silvicultural trials in northern and Appalachian hardwood timber types in the Eastern United States . . . . . . . . . . Location of the cutting methods study within the Iron River-Wabeno (IWa and IWb) soil mapping units on the Argonne Experimental Forest Sampling design for replication 1 (40 acres), Argonne cutting methods study (reproduced from an original drawing) . . . . . . . . . . . . . . . . Real Michigan Upper Peninsula sugar maple sawtimber stumpage prices, 1954-91 . . . . . . . Real Wisconsin sugar maple sawtimber stumpage prices, 1950-90 . . . . . . . . . . . . Real Michigan Upper Peninsula basswood sawtimber stumpage prices, 1954-91 . . . . . . . Real Wisconsin basswood sawtimber stumpage DIICOO, 1950-90 e e e e e e e e e e e e e e e e e Real Michigan Upper Peninsula white ash sawtimber stumpage prices, 1978-91 . . . . . . . Real Wisconsin white ash sawtimber stumpage priCOO ' 1950-90 e e e e e e e e e e e e e e e e 0 Real Michigan Upper Peninsula yellow birch sawtimber stumpage prices, 1954-91 . . . . . . . Real Wisconsin yellow birch sawtimber .tmp‘g. ptiCOO ' 1950-90 e e e e e e e e e e e e Real Michigan Upper Peninsula red maple sawtimber stumpage prices, 1954-91 . . . . . . . Real Wisconsin red maple sawtimber stumpage pric.‘ ' 1950-90 O O O O O O O O O O O O O O O O O Real Wisconsin elm sawtimber stumpage ptiCOO, 1950-90 s e e e e e e e e e e e e e e e e xiii Page 12 26 28 29 34 58 58 60 60 62 62 64 64 66 66 68 Figure Figure Figure Figure Figure Figure Figure Figure Figure 4-12 e 4-13e 4-14. 4-15. ‘-16e 7-1e 7-2 s 8-1e 9-1e Real Wisconsin hemlock sawtimber stumpage FIICOO, 1950-90 e e e e e e e e e e e e e e e e e e e 68 Real Michigan Upper Peninsula mixed hardwood sawtimber stumpage prices, 1972-91 . . . . . . . . . 70 Real Wisconsin mixed hardwood sawtimber stumpage prices, 1972-90 . . . . . . . . . . . . . . 70 Real Michigan Upper Peninsula hardwood pulpwood stumpage prices, 1954-91 . . . . . . . . . . 71 Real Wisconsin mixed hardwood pulpwood stumpage prices, 1972-1990 . . . . . . . . . . . . . 71 Comparison of 1990 overstory basal area and Shannon's index of saplings 2.0 to 4.5” d.b.h. for uneven-aged treatments . . . . . . . . . 109 Comparison of net present value and Shannon's diversity index of saplings 2.0 to 4.5” d.b.h. NPV's are based on the differences between the control and cutting treatments . . . . . . . . . . 117 Comparison of harvested, residual, and total lumber values for four tree quality study treatments. Values are for differences between treatments and control . . . . . . . . . . . . . . . . . . . . 138 Comparative economic returns, 40-year and 70-year analyses. NPV's reflect differences between control and cutting treatments with medium prices . . . . . 154 xiv CHAPTER 18 THE PROBLEM Northern hardwood forests cover about 40 million acres in the Eastern United States. These forests stretch from Minnesota, across southern Canada and New England as far as central Maine, and south to West Virginia and southern Indiana. These forest types are among the most valuable and diverse hardwood forests in North America. Sugar maple (Acer saccharum Marsh.) and red maple (Acer rubrum L.) dominate most northern hardwood forests. Other primary species include basswood (Tilia americana L.), American beech (Fagus grandifolia Ehrh.), yellow birch (Betula alleghaniensis Britten.), and white ash (Fraxinus americana L.). Most commercial northern hardwood stands are in five States: Michigan, New York, Wisconsin, Maine, and Pennsylvania (Table 1-1). In the Lake States (Minnesota, Wisconsin, and Michigan) northern hardwoods cover over 11 million acres. They represent 25 percent of the commercial forestland in these states, yet they also provide wildlife habitat and a vacation setting for millions of visitors each year. Because these hardwood forests respond to a variety of silvicultural techniques, forest managers have many options in managing them for a variety of resource values (Erdmann 1986). Both even- and uneven-aged management systems are used to manage northern hardwood forests. However, there is considerable debate within the forestry community concerning the most reliable and profitable cutting and regeneration methods for the various northern hardwood types. Although commonly called "methods," even- and uneven-aged management are actually management goals that forest managers work toward achieving by applying specific cutting methods (Godman 1985). In even-aged management the forest has a definite beginning and ending point. Trees are harvested and regenerated at nearly the same time. 2 TABLE 1-1. Area of northern hardwood forest types in the Eastern United States in millions of acres.’ , _____I , I State Herthern Sugar Maple I Hardwood Types Types lNew York (1982) 9.3 4.4 I Michigan (1983) 6.2 5.0 I Maine (1984) 5.0 3.1 I Pennsylvania (1982) 4.3 3.5 Wisconsin (1985) 3.5 3.0 West Virginia (1978) 3.0 2.7 Vermont (1985) 2.7 2.0 New Hampshire (1985) 2.0 1.1 Minnesota (1990) 1.6 1.4 Ohio (1981) 1.5 0.6 Indiana (1988) 1.0 0.7 ‘ Total 40.0 ll 27.4 TOTALS: 40 million acres; about two-thirds in sugar maple types. ' SOURCE: USDA Forest Service, Forest Survey reports for each of the States listed above. Dates of publication are listed in parentheses. ' ”Sugar maple types“ represent an aggregate of acreage of the following forest types recognized by the Society of American Foresters: 1) SAF Type 25 - sugar maple/beech/yellow birch: 2) SAF Type 26 - sugar maple/ basswood; 3) SAF Type 27 - sugar maple. In a few States, all sugar maple types are lumped into SAF Type 25 above. NOTE: Tubbs (1977b) said that northern hardwoods, broadly defined, cover as much as 100 million acres in the eastern United States and Canada. The above figures came from Forest Survey data and included the major commercial area for northern hardwoods. Sturos and Thompson (1986) estimated that northern hardwoods, more narrowly defined, occupied 32 million acres in the eastern United States. Their definition probably incorporates only stands where sugar maple is a prominent species. Northern hardwood stands where red maple, beech, cherry, hemlock, or red oak are dominant are not necessarily ncluded. 3 The type of cutting method used at the end of the forest‘s rotation (for example, clearcutting or shelterwood cutting) defines the silvicultural method used to achieve even-aged management. The choice of even-aged management involves managing the entire forest as a single crop. The usual emphasis in even-aged northern hardwood management is on fiber or pulp production with sawlogs as a by-product. An uneven-aged forest, unlike an even-aged forest, cannot be treated like an agricultural crop to be planted and harvested all at once. Rather, the uneven-aged forest has no identifiable beginning or and (Murphy and Guldin 1987). Uneven-aged management of northern hardwoods involves removing a few of the larger trees at each thinning, with thinnings occurring on a cutting cycle of 10-20 years. The goal is to create a variety of sizes of trees, called stand structure. Eventually, most growth can be harvested as sawtimber, yielding fewer but larger and more valuable trees. Selective cutting is the method used to achieve this management goal. Economically, the choice of uneven-aged management involves managing the forest for those individual trees that can attain the highest value in the shortest possible time. The emphasis in uneven- aged management, then, is on sawlog production, with fiber as a by- product from improvement cuttings. Most previous studies of northern hardwood management strategies have made recommendations based primarily on silvicultural grounds. However, there is now a critical lack of economic information that compares the profitability of even- and uneven-aged forest management strategies in Lake States northern hardwood stands. Today's forest managers and planners need economic information that considers financial returns from silvicultural methods practiced on a variety of northern hardwood site and stand conditions. In short, economic management guides for northern hardwoods need improvement. 4 Several problems must be overcome to achieve higher economic returns in northern hardwood forests. First, the overall quality of northern hardwoods must be improved. Improvements in quality are needed because most stands originated following large-scale commercial clearcutting, and many have suffered from periodic high-grading, leaving forests with many poor quality trees. Second, species composition of northern hardwood forests should be modified to bring higher economic returns. Shifts in species composition hold the promise of improving economic performance, reducing insect and disease risks, and creating more diverse forest habitats for non-timber uses. Third, new markets and improved utilization for the maples are required, because they will remain the dominant northern hardwood species in the Lake States. Both red and sugar maple are increasing significantly in area and volume (Spencer et a1. 1988, Spencer and Hahn 1984). Finally, there is a need to identify those northern hardwood forests and types that are most productive and concentrate investments on these sites. Underlying the realisation of all these goals is the need for improved knowledge of the economic response of northern hardwoods to the various cutting methods used in their management. Some current issues underscore the need for improved decision- making on even- and uneven-aged management of northern hardwoods. Many special interest groups now Oppose clearcutting. Their negative stance has pushed agencies such as the USDA Forest Service and State natural resources departments to look more closely at the potential for wider application of uneven-aged management of their hardwood forests. This is particularly true within and near environmentally sensitive areas such as river corridors, natural areas, and recreation areas. Recent concern about the negative effects of clearcutting has caused the USDA Forest Service to change its policy on the use of clearcutting on the National Forests. Though acknowledged as a proven forest management 5 tool, clearcutting will no longer be the standard method of harvesting national forest timberh Some private landowners, particularly large individual and corporate owners, remain interested in conditions where even-aged management is feasible in their forests. These private owners have a renewed interest in even-aged management because of the potential it has for increasing species diversity, wildlife habitat, and profits. No matter the type of ownership, all these clients need better information on the economic trade-offs involved in managing their northern hardwoods with either even- or uneven-aged methods. Although there is a body of theoretical knowledge about the economics of even- and uneven-aged management in hardwood forests, most of these studies have been based on limited data sets or hypothetical site conditions. A few studies have compared the economics of cutting methods in northern hardwoods. But generally they suffer from short- term data bases or a lack of replication. At this writing no comparative analysis, based on long-term field data from a replicated study, has been completed on the economics of even- and uneven-aged management in northern hardwoods. By answering forest managers' questions concerning the economic productivity of even- and uneven-aged management in northern hardwood forests, this study will help fill this research gap. This inquiry encompasses three major areas of concern to forest managers: 1) an analysis of economic returns from nine northern hardwood cutting methods: 2) an analysis of the contribution of tree quality improvements to the economic performance of specific cutting methods; and, 3) an analysis of the trade-offs between achieving 'Robertson, F.D. “Ecosystem Management of The National Forests and Grasslands." Letter from Dale Robertson, Chief, USDA Forest Service, to Regional Foresters and Station Directors. Dated June 4, 1992. 3 p. plus enclosures. 6 economic returns and improvements in tree species diversity in northern hardwood management. CHAPTER 2: LITERATURE REVIEW Sugar maple usually comprises a plurality of the stocking in northern hardwood forest types: often it makes up 70-80 percent or more of the stems or basal area. Researchers and managers are concerned that the dominance of sugar maple is increasing in these forests (Stearns 1986), both for financial reasons and for potential losses to forest health and ecosystem diversity. The Society of American Foresters recognizes six major cover types where sugar maple is abundant. They are: SAF Type 25--Sugar Maple/Beech/Yellow Birch; SAF Type 26--Sugar Maple/Basswood; SAF Type 27--Sugar Maple: SAF Type 28--Black Cherry/Sugar Maple; SAF Type 31--Red Spruce/Sugar Maple/Beech: and SAF Type 60--Beech/Sugar Maple (Eyre 1980). A recently established type (Eyre 1980), SAF Type 108--Red Maple, recognises the growing dominance of red maple in disturbed and wet-site transitional northern hardwood stands. American beech is a common associate in eastern Lake States and in Northeastern hardwood types. Basswood is a common associate throughout the Lake States, while American elm (Ulmus americana L.) occurs on the more mesic northern hardwood sites. However, usually the scarcer and less tolerant northern hardwood species command higher economic returns and have greater wildlife benefits in northern hardwood forests (Marquis 1967, Della-Bianca and Beck 1985). These species include yellow birch, white ash, basswood, and northern red oak (Quercus rubra L.). In the southern areas of northern hardwood types, other high-value species such as black cherry (Prunes serotina Ehrh.) and black walnut (Juglans nigra L.) may be present in diverse stands. Intolerant pioneer species that occur in the early stages of northern hardwood stand development may include: quaking aspen (Pcpulus tremuloides Michx.), bigtooth aspen (Populus grandidentata Michx.), paper birch (Betula papyrifera Marsh.), and pin cherry (Prunes pensylvanica L.f.). History Northern hardwood forests have regenerated throughout the North following large-scale cutting that occurred in the mid-to-late nineteenth century, and early in the twentieth century. Many of these forests have been high-graded repeatedly, particularly in the Northeast, and more recently in parts of the Lake States. Frequently these forests have regenerated on sites that were formerly occupied by pine and other softwoods. Poor quality trees often dominate these second-growth forests as a result of multiple entries and logging practices that took only the best trees (i.e., high-grading). New markets for lower grade trees hold the promise of fostering regeneration on these hardwood forests to increase their value to society (Webster 1986). However, there has been a long-running debate within the forestry community concerning the most reliable and profitable regeneration methods for the various northern hardwood types. Most early cutting of northern hardwoods, from Maine to Minnesota, was 'logger's choice." Commercial clearcutting of northern hardwoods began anywhere from 10-50 years after the big old-growth white pine were removed from mixed stands. Application of commercial clearcutting meant that the best trees were cut first, followed by cutting down to a diameter-limit that could range from 5-10 inches, depending on local markets. In a few areas such as western Pennsylvania, there were chemical-wood (charcoal) markets, so these northern hardwood stands received the equivalent of a complete clearcutting, resulting in stands dominated by high-value, intolerant trees such as black cherry. By the time the hardwood industry reached the northern Lake States in the early 1920's, it was clear that old-growth hardwood resources were running out. Many managers in the industry began to look at “selective cutting" as a way to "stretch out“ a dwindling old-growth northern hardwood resource (Eyre and Zillgitt 1953). Selective logging not only prolonged the lumber industry's exploitation of old-growth 9 northern hardwood forests, but it also provided a method to harvest only the largest and best trees that were the only marketable ones in the region's depressed lumber markets. Truck logging, which had begun to replace railroad logging in the early 1920's, made selective logging technically feasible in old-growth stands. Early research by the USDA Forest Service on the economics of selective logging (Son and Garver 1930, Buell 1937) gave credence to its practice. Unfortunately, the rise of selective logging (partial cutting) created confusion with the selection system of silviculture. This confusion persists in the minds of the public and some in the forestry community to this day (Nelson et a1. 1975). Selective cutting refers to a harvest operation that takes only a portion of the sawtimber-sized trees in a stand; this selective cutting often involves "high-grading," or removal of the best trees and leaving of the worst. By contrast, the selection system of silviculture involves all the management steps necessary in scientific uneven-aged management, of which a selection cut every 10-20 years is only one step. As forest managers in the Northeast observed the differences in stand quality between the 'logger's choice“ cuts and the less common complete clearcuts, it became obvious that even-aged management by complete clearcutting could result in stands of higher value than indiscriminate partial cutting. Thus, a trend toward scientific even- aged management began in the valuable cherry-maple forests of the Allegheny Plateau and spread throughout the Northeast (Dana 1930). In the Lake States, however, forest managers had greater difficulty applying clearcutting successfully, due both to regeneration problems and the lack of pulpwood and chemical-wood markets. Research in the Lake States concentrated on partial cutting methods in northern hardwood stands, with the emergence of the selection system of silviculture in the 1950's (Arbogast 1957) and a science-based uneven-aged management. 10 uneven-aged Management Methods Various studies have been conducted on the management and regeneration of northern hardwood timber types. Most of these studies have been based on silvicultural objectives. In the Lake States, many of these studies have concentrated primarily on uneven-aged management. As early as the 1920's researchers began to compare cutting methods in old-growth Lake States northern hardwoods. Eyre and zillgitt (1953) summarised the results of an old-growth cutting methods study that compared diameter-limit cuts of 5, 12, and 22 inches with light, medium and heavy selection cuts, and a control. They reported a maximum net growth rate of 295 board-feet per acre per year (gross growth minus mortality without deduction for cull) over a 15-year period, with a residual volume of 6,300 board-feet per acre. They defined optimum stocking for best growth between 4000 and 7000 board feet or 50 to 75 square feet per acre, on a 15-year cutting cycle. They found that light selection cuttings were best for maintaining quality, and heavy cuttings were best for growth. Moderate cuttings reduced tree quality only slightly while increasing growth rate. This intermediate treatment yielded premiums for larger quantities of sugar maple whitewood (sapwood). One treatment in particular (removal of overmature and defective stems) had the highest average net board foot growth (252 ft./ac./yr.), suggesting the importance of cutting out defective and cull trees. Another result of Eyre and Zillgitt's (1953) study was a call for similar cutting methods research in second-growth northern hardwood stands, which were rapidly replacing the old-growth in the Lake States. Much of this research began at the Argonne Experimental Forest in Wisconsin in the early 1950's. Arbogast (1957) developed a silvicultural cutting guide for northern hardwoods that has been widely applied in northern hardwood stands, especially in the Lake States. His guide called for cutting 11 northern hardwood stands dominated by sugar maple to a residual stocking of 92 square feet per acre, based upon the following decision criteria, in their general order of importance: 1) risk: 2) cull; 3) quality; 4) vigor: 5) species: 6) crown position; and, 7) size. Maximum tree size recommended by Arbogast was 24 inches, though he suggested that most northern hardwoods became financially mature in the 20-24 inch size range (Figure 2-1). Godman and Erdmann (1981) recommended uneven-aged management for tolerant hardwoods such as sugar maple because this system produced greater lumber yields, product grades, merchantable heights and diameters. Erdmann (1986) cited sugar maple as a species producing larger yields of high quality material under uneven-aged management systems, stressing that sugar maple cover types were particularly well adapted to intensive cultural treatments. Godman and Erdmann (1981) emphasised that management of northern hardwoods on higher quality sites should be for producing sustained yields of high grade veneer and sawlogs, using uneven-aged management. Partly because of the above research findings, Lake States foresters often manage northern hardwood stands with single tree selection (Eyre and Zillgitt 1953, Arbogast 1957). This method develops well-regulated stands dominated by sugar maple and other tolerant species such as American beech, red maple, and Eastern hemlock (Tubbs 1977b, Erdmann 1986). Most of the high-value but less tolerant species (the oaks, cherry, yellow birch, white ash) tend to be lost from these managed stands (Zillgitt and Eyre 1945). However, Erdmann's (1987) work has shown that smaller (less than 15 percent of each species) components of non-maple species can be maintained and managed in selection stands. 12 Trees Per Acre 140 120—— ------------------------------------------ 100— ---------------------------------------- 60— ———————————————————————————————————————— 40— --------- — ———————————————————————————— 20— ———————————————————————————————————————— 011111 '11. 2 4 681012141618202224 Diameter class (inches) *Arbogast +Adams 8. ER '3‘? Martin "' Bars 8. Opalach * Lyon 8. Reed Figure 2-1. Comparison between Arbogast and other Lake States uneven-aged cutting guides for recommended reserve growing stock. 13 Managers using the single-tree selection method must be concerned with three factors that impinge on silvicultural development and economic yields. They are: stand diameter distribution (the 'q' factor--the average quotient between number of trees in successively smaller diameter classes, Sammi 1961), residual basal area, length of cutting cycle, and the largest diameter tree (LDT) that will be allowed to develOp before cutting. The existing diameter distribution in unmanaged second-growth stands depends primarily on how the original old-growth stand was cut. Early logging operations that took only the best grades have left many high-graded stands with an uneven-aged character. Their diameter distributions have the typical inverse J- shaped curve associated with uneven-aged stands. Recommended uneven- aged stand structure (Erdmann 1987, Arbogast 1957) can be achieved in these stands within two cutting cycles if they are cut to residual basal areas of 90, 75, or 60 square feet per acre (Reed of a1. 1986). Schultz2 studied management opportunities for a high-quality sugar maple stand partially within a national scenic river management area of northern Wisconsin. He found that uneven-aged management by single tree selection would yield the highest economic returns. Also, this cutting protected the sensitive visual resources of the area and critical nesting habitat of the bald eagle. Schultz's study emphasised the need for further research on: 1) the trade-offs in timber quality and value between even- and all-aged management: and, 2) the need to determine the economic trade-offs of management at different "q" levels. The group selection method of cutting is an uneven-aged technique that calls for the removal of small groups of trees one-quarter to two acres in sise (Smith 1962). This cutting method has met with mixed success in northern hardwoods because it makes regulation of the cut and management 2Schultz, J.R. Silvicultural prescription: stand number 2, Namekagon River Tract. Unpublished paper for silvicultural certification. 1981. Milwaukee, WI: 0.8. Department of Agriculture, Forest Service, Region 9, Chequamegon National Forest. 81 p. 14 for sustained yield very difficult, if not impossible, to achieve (Roach 1974, Tubbs 1977a). Studies in New England demonstrated the effectiveness of group selection in regenerating white ash and yellow birch (Leak and Filip 1977). Group selection, however, has typically regenerated undesirable tolerant species such as ironwood (Ostrya virginiana Mill.) x. Roch) in the Lake States (Erdmann 1986). Therefore, it is not a viable cutting method in this region. Diameter-limit cuttings are attractive to forest managers because they offer high initial financial returns and are easily applied to northern hardwood stands (Crew 1985). However, stand development, net growth and long-term financial returns are often less than with other cutting methods (Erdmann 1986, Niese and Strong 1992). An 8-inch diameter-limit cut applied to a productive (Site Index 50 c 65) Wisconsin northern hardwood stand resulted in fewer than 25 potential crop trees per acre after 35 years and no merchantable timber until 40 years after application (Erdmann 1986). Other studies (Eyre and Sillgitt 1953, Reed et a1. 1986) have reported similar results. When diameter-limit cuts of any diameter are strictly followed, stand quality will gradually decline because small, low-quality, and defective trees are not cut (Crew 1985). Erdmann (1986) recommended that if they must be used, diameter-limits should be set at 16 inches or higher so that good quality sawlogs can be produced. In Appalachian mixed hardwoods containing sugar maple, Smith and Miller (1987) recommend use of a flexible diameter-limit cut, but they also cautioned against using diameter-limit cuts with lower than a 16-inch limit. Even-aged Management Methods Northern hardwood stands also can be managed successfully using even-aged methods. However, there are varying degrees of success with even-aged methods, especially clearcutting, depending on geographic location, stand condition, and regeneration conditions. 15 Old-growth stands that were commercially clearcut, a more common treatment than above, are mostly even-aged stands today. Their diameter distributions tend to follow the normal bell-shaped frequency curve associated with even-aged stands. Sometimes these hardwood stands, though even-aged, are a mixture of fast- and slow-growing species, creating an even-aged stand with an uneven-aged diameter distribution. In either case, these stands can be managed as even-aged (when the manager desires to maintain valuable intolerant species), or converted to uneven-aged management with three selection cuts applied at about 10- year intervals (Erdmann 1987). Research has shown that clearcutting Lake States northern hardwoods is desirable only for reproducing stands of sugar maple when adequate advanced regeneration is already present (Godman 1985). These types of cuts are sometimes called "one-cut shelterwoods' to prevent misuse of clearcutting. One study on clearcutting northern hardwoods in Wisconsin resulted in shrub and grass-dominated communities, despite abundant small advance reproduction present before cutting (Metsger and Tubbs 1971). Clearcutting without adequate established regeneration present can increase northern hardwood rotation length by 10-15 years (Metsger and Tubbs 1971). Yet, in the Northeast, species such as white ash, black cherry, and other intolerants regenerate successfully under clearcutting, with a reduction of shade-tolerant beech and sugar maple (Walters and Nyland 1989). Contrary to conventional forestry wisdom, Godman and Erdmann (1981) found that even-aged management could be applied to all hardwood species in Lake States northern hardwood types. They reported, however, that growing high quality holes with even-aged methods was more difficult than with uneven-aged methods. Their work suggested a one- half to one leg loss in merchantable height was possible when using these methods in the Lake States. A later report (Godman 1985) showed that the two-cut shelterwood silvicultural system was the most reliable 16 way to regenerate even-aged Lake States northern hardwoods, especially small-seeded species such as yellow and paper birch and hemlock. Economic benefits of shelterwood management can also be important. Niese and Strong (1992) reported a two-cut shelterwood treatment as having the highest net present values of three even-aged methods in a Wisconsin cutting methods study. With the increased interest in even-aged hardwood management, shelterwood methods of regeneration have received closer attention, particularly in the Lake States (Tubbs 1987). In the last decade researchers have learned that shelterwood and clearcutting methods can sometimes increase both the value and diversity of hardwood resources (Erdmann 1983, Niese and Strong 1992). Hannah (1988,1991) also found renewed interest in shelterwood regeneration methods because selection methods often do not achieve regeneration and stand structure objectives, and because clearcutting has proven aesthetically unacceptable, especially in hardwoods. Higher costs of implementation, lack of experience among foresters, and private landowner reluctance were limiting factors in the application of shelterwood in hardwoods. Research (Helty and Nyland 1981) also has shown that, in some areas, heavy deer browsing and undesirable beech regeneration must be overcome before shelterwood cutting can be used successfully. Many of the finest hardwood stands are even-aged, such as the cherry/maple forests of the Allegheny plateau. Smith (1987) cited this historical precedent as a compelling reason for even-aged hardwood management. He described most northeastern hardwood stands as vertically-stratified, even-age mixtures often misdiagnosed as uneven-age stands, a portrayal confirmed by Oliver (1980). Smith found that in most of these hardwood mixtures, moisture supplies were sufficient to manage the upper-stratum less tolerant species separately from the tolerant lower-stratum species, using a modified shelterwood 17 system. Such a two-tiered management regime may be more appropriate in the Northeast than in the droughtier Lake States (Smith 1987). Leak and Gottsacker (1985) reported that new approaches were making even-aged management systems more feasible in New England. They showed that second-growth hardwood stands with lower proportions of sawtimber produced better net growth responses than stands with higher percentages of sawtimber. BlashockJ recomended even-aged management for maximising financial returns and providing a mix of other benefits in a maple- basswood stand in northern Michigan. He cited the need for research on artificial regeneration of oak in northern hardwood types. He suggested that managers could use oak enrichment plantings as economic options to improve species diversity and grow larger diameter sawtimber trees. A study by Oliver (1978) found that in mixed, even-aged central New England hardwood stands regeneration of maples and birches greatly outnumbered and outgrew oaks during the first 20 years. Eventually, however, the oaks came to dominate the stand. Many of these mixed, vertically stratified yet even-aged hardwood stands have been managed as though they were uneven-aged stands (Oliver 1980). This improper assessment has led to a variety of management problems. By careful management of openings in these mixed, even-aged stands oak species can be regenerated to create improved stand diversity and potential value. Forest managers and researchers have long been aware of the fact that regeneration requirements of the various northern hardwood species vary greatly. Soil temperature and light availability at germination, amount of scarification, insect pests of seeds and seedlings, and predation by rodents and browsing by herbivores are particularly important factors governing future stand composition (Godman 1985). ’Blashock, L.F. Mackenzie National Ski Trail--silvicultural options. Unpublished paper for silvicultural certification. 1983. Milwaukee, WI: 0.8. Department of Agriculture, Forest Service, Region 9, Huron-Manistee National Forest. 59 p. 18 In northern hardwoods, the dominance of sugar maple is another factor mitigating against managers' efforts to increase tree diversity through careful regeneration practices. Fre-settlement stands in the Lake States had a greater species diversity than managers find in today's stands (Stearns 1986), which are often 70-90 percent sugar maple. Past high-grading, protection from fire, a greatly increased deer herd, and succession to tolerant species have accelerated the trend toward lower species diversity in Lake States hardwoods (Stearns 1986). The increased dominance of sugar maple under these conditions has made it difficult for forest managers to maintain the less tolerant species in Lake States northern hardwood stands. Maintaining or enhancing species diversity in northern hardwood stands is often dependent upon meeting the specific regeneration requirements of the less common associated species. Varying the timing and intensity of cutting can improve the manager's chances of reaching diversity goals in stand management. Both even- and uneven-aged management can be used to regenerate northern hardwood stands (Godman 1985). Conventional wisdom holds that even-aged management produces more diverse stands than uneven-aged management. However, either system can increase or reduce species diversity under varying site conditions (Stearns 1986). Jacobs (1974) evaluated a two-cut shelterwood applied to an uneven-aged northern hardwood stand in northeastern Wisconsin. He found species composition unchanged three years later, with sugar maple comprising 87 percent of the understory stems and white ash 8 percent. However, shelterwood cutting alone does not necessarily increase species diversity. A shelterwood with scarification increased species diversity in a similar Wisconsin study (Godman 1985), particularly of light-seeded species such as hemlock, white birch, and yellow birch. Group selection cuts of 0.1 to 0.2 acres in size have been successful in regenerating white ash and yellow birch in New England 19 (Leak and Filip 1977). However, group selection has regenerated undesirable tolerant species such as ironwood, and few desirable species, in the Lake States (Erdmann 1986). Basel area treatments in northern hardwoods typically are cut to residual stand densities of 60-90 ftnz/acre. As partial cuttings geared to sawlog production, these often favor the development of sugar maple and other shade-tolerant species (Tubbs 1977b). However, where site conditions are favorable for other species, more aggressive basal area cuts can favor the establishment of less tolerant species (Tubbs 1977a). Economics of Cutting Methods Relatively few studies have considered the economic returns of various treatments applied to hardwood stands. Eyre and Zillgitt's (1953) classic cutting methods study of old-growth northern hardwoods in Michigan's Upper Peninsula found silviculturally optimum residual stocking for maximum growth occurred at 5,500 to 6,000 board feet per acre (BA 60-65) in sawtimber-sized trees. However, the economically optimal stocking was much lower, 3,500 board feet per acre, yielding a 4.7 percent real rate of return on a 15-year cutting cycle. Eyre and Sillgitt (1953) suggested that forest managers compromise with residual stocking of 4,500 to 5,000 feet, which would approach optimum growth while yielding a 3-4 percent real rate of return. A 22-year financial evaluation of cutting methods in Michigan northern hardwoods (Reed et a1. 1986) found that a 16-inch diameter- limit out had the greatest financial return among nine different treatments. A lack of data on quality ingrowth in the basal area treatments was a possible factor causing results to favor the diameter- limit treatments. The authors said that there would be a loss of quality from diameter-limit cutting, and improvements in quality from the basal area cuts that would change the financial returns. Also, this 20 study only considered periodic harvest revenues, not total stand value, in making its economic comparisons. A recent evaluation of the same study with ten more years of management (Erickson et a1. 1990) confirms this observation. Though the diameter-limit out still had the highest financial ranking, the light improvement treatment had improved its ranking from sixth to third of the eight treatments studied. A 34-year Appalachian cutting methods study in a sugar maple/yellow poplar/black locust stand (Smith and Miller 1987) compared the silvicultural and economic performance of four methods--a 16-inch diameter-limit, commercial clearcutting, and two selection treatments. The diameter-limit cut had the greatest value of periodic revenues, but total treatment value, including residual stand value, was comparable between the diameter-limit and the selection cuts for sawtimber and pole-sized trees. A few studies of northern hardwoods management have considered the rates of return from various cutting treatments. The premise of these studies is that an individual tree should earn the discount rate or greater to be left in the stand in a selection cut. If the tree's rate of value growth (Gansner and Herrick 1987) falls below the discount rate, it should be cut (Murphy and Guldin 1987, Duerr et a1. 1956). Marquis, et a1. (1984) noted that there were no analyses on the trade-offs between diameter growth and quality sufficient to suggest that investments in precommercial thinning of northern hardwoods are economically feasible. Several recent studies show that commercial thinnings in northern hardwood stands can provide some very attractive rates of return. Johnson (1986) found that thinning overstocked sugar maple pole stands in Michigan resulted in an average 45 percent reduction of time needed to reach maturity. Thiede (1986) found rates of return of 6.4-25 percent for hardwood thinnings on State lands in Michigan, assuming that timber did not have to cover land purchase 21 costs. Webster (1986) also noted real rates of return of 9-12 percent for northern hardwood management. Large acreage of northern hardwoods with commercial thinning investment opportunities have been partially responsible for prompting Michigan to initiate a Forest Development Fund. The State expects to generate sufficient capital to pay for these forest improvements that yield competitive rates of return (Murray 1989). Martin (1982) developed a set of investment-efficient stocking guides (Figure 2-1) for northern hardwood forests based on prior work done by Adams and Ek (1974) and Adams (1976). Because of the opportunity costs of holding growing stock between harvests, Martin recommended that managers should leave larger numbers of small diameter trees (6-14 inches) and fewer large diameter trees (16-24 inches) than was recommended in previous guides published by Arbogast (1957) and Tubbs (1977b). However, Martin's investment-efficient stocking guides make no assumptions about the higher stumpage prices obtainable for large sawtimber trees. The lack of information on the economics of quality differences between various northern hardwood stands is a critical factor in evaluating returns to management. Martin recognised this need and recommended that future research was needed to quantify the economic effects of these differences in quality. Haya and Buongiorno (1987, 1989) developed a Markovian cutting guide for northern hardwoods that considers the uncertain growth of forest stands and the variation in product prices. Their guide differs from the Arbogast guide and deterministic economic guides because instead of prescribing a fixed residual stand and a constant cutting cycle, the Markovian guide suggests a decision that depends on the state of the stand and on product price when the stand is examined. Significantly, Raya and Buongiorno (1989) found that northern hardwood stands could be managed more economically by concentrating growth and harvest in the larger sawtimber classes, while minimizing the cutting of 22 pulpwood, which they found uneconomical to harvest from even-aged stands. However, the limiting of harvest treatments to saleable sawtimber classes will not maintain a stable stand structure (Roach 1974, Erdmann 1986). Without cutting of smaller pole-sized trees, crowding will develop in the middle diameter classes. This form of diameter-limit management will effectively reduce future returns to the landowner, and appears to perpetuate stand high-grading. As in Martin's investment-efficient stocking guide, Maya and Buongiorno's work does not reflect the potential impact that cutting of hardwood pulpwood and cull has on quality growth in current and future sawtimber classes. Haya and Buongiorno's (1989) work raises an interesting question, one debated in many forestry circles today. Should forest economics be based upon scientific silvicultural decision-making, or should silvicultural options be based upon economic considerations, and in what time frame? Whose time preferences should count, and when, and on what lands? Current debates make it clear that economic management regimes for northern hardwoods must be based on biologically sustainable systems, yet silvicultural efforts must also be responsive to economic priorities. The two disciplines must inform each other; indeed, they must be accountable to one another in forest management. Quality Considerations in Economic Management of Northern Hardwoods Several past studies (Erdmann 1986, Nyland 1986, Tubbs 1977b, Eyre and Sillgitt 1953) cite management of northern hardwoods for the sustained production of high grade veneer and sawlogs as an important objective. However, there is scant economic evidence to support these assertions. Most studies on tree quality have been theoretical or silvicultural in nature (McCay and DeBald 1973, Brisbin and Dale 1987). Economic evaluation of long-term (i.e., greater than 20 years) quality response to treatment has not been possible previously due to lack of data. However, surveys of research priorities for Eastern hardwoods 23 (McLintock 1987) have shown that improved prediction of tree quality changes resulting from silvicultural practices is one of the greatest research needs. A few hardwood studies have found that aggressive timber stand improvement practices result in favorable changes in tree grade and species composition, with favorable economic results. In an uneven-aged mixed hardwood stand in Illinois, White and Campbell (1988) found these changes were sufficient to increase the economically optimal largest diameter tree (LDT) from 16 to 22 inches with an average price scenario, to as high as 26 inches under an enhanced price scenario. Their work confirms the need for an active timber stand improvement (TSI) effort that assures uneven-aged crop trees are of the highest quality and the most marketable species. In medium- to high-value hardwoods, research has confirmed that quality change over time is at least as important, if not more important, than diameter and height growth in contributing to timber stand value (Rast et a1. 1987, Sonderman 1986). In Lake States northern hardwoods, increases in tree grade over a 10-year period consistently added greater value than either gains in merchantable height or diameter growth (Godman and Mendel 1978). Previous work in Ohio oak stands by McCay and DeBald (1973) also confirmed the financial superiority of grade increases over height or diameter gains alone; their study showed a doubling in rate of value increase for 18-inch red oak trees that changed from a butt-log grade 2 to grade 1, versus those with no change. Mautiyal (1983) has proposed a theoretical model for managing uneven-aged selection hardwood forests based on the theory of financial maturity. Using tree vigor as a criterion, he derived optimal rotations from crops of trees that have increasing stumpage price functions, thus effectively ”safeguarding" the penultimate (next-to-last) crop before financial maturity. Increases in tree grade or size caused most of these increasing stumpage price functions. Nautiyal's work underscores 24 the importance of including hardwood tree quality considerations in economic evaluation of hardwood management, because tree grade improvements tend to lengthen financial maturity (Nautiyal 1982, Godman and Mendel 1978). Miller (1991) used a linear programming model similar to one devised by Buongiorno and Michie (1980) to create optimal stand structure goals for single tree selection management in an Appalachian hardwood stand. In his work, Miller compared grade yields and financial profitability between diameter-limit and selection cutting regimes. He calculated a maximum net present value of $433 per acre (for periodic revenues) with a 22-inch diameter limit cut, while the best single-tree selection management model produced a NPV of $404 per acre using a q- value of 1.2, largest diameter tree of 20 inches, and a residual basal area of 75 square feet per acre. However, he also found that the selection management treatments experienced distinct improvements in residual tree quality over the 30-year period, while the diameter-limit and control treatments did not. Based on these results and the similar economic returns between the optimal diameter-limit and selection practices, Miller (1991) recommended that managers consider single tree selection management to promote higher quality and total stand value. In the Lake States, no definitive work has been done to compare the economic profitability of even- and uneven-aged cutting methods. Many cutting guides and models have been developed, but no research has yet confirmed the economic superiority of one or more cutting methods in northern hardwoods, under various conditions. The lack of data on the long-term effects of cutting methods on tree quality has increased the difficulty of achieving accurate estimates of economic returns under the various methods. However, current work holds the promise of overcoming the above obstacles, and should give forest managers much improved data for decision-making in northern hardwoods management. CHAPTER 3: RESEARCH METHODS This study has three objectives that parallel the research components (page 6) described earlier: 1. Determine which cutting methods are profitable and maintain or enhance tree species diversity. 2. Estimate the economic returns that occur because of changes in tree quality due to cutting method. 3. Determine total economic returns from nine cutting methods over both 40- and 70-year investment periods. Three research hypotheses were tested in this study. They are: 1. IL: Diversity of tree regeneration is not affected by cutting method. 2. 15: ‘There is no difference among economic returns for northern hardwood stands under different management methods. 3. 1%: ‘Total economic return is not affected by varying even- and uneven-aged cutting methods. This paper addressed the above research objectives and hypotheses generally, and specifically in Chapters 7, 8, and 9, respectively. The study is based on data collected over a 40-year period on the Argonne Experimental Forest in north central Wisconsin (Figure 3-1). The data originate from a northern hardwoods cutting methods study sponsored by the USDA Forest Service, North Central Forest Experiment Station (NH-25). Previous work on the study centered around silvicultural options for managers of northern hardwoods. Various aspects of this silvicultural work were summarized by Stoecklerh Arbogast (1957), Godman and Books (1971), Metzger and Tubbs (1971), |Stoeckler, J.H. Establishment report for cutting methods study (A-3) in pulpwood-sized northern hardwoods. Unpublished manuscript (dated January 1955). St. Paul, MN: U.S. Department of Agriculture, Forest Service, Lake States Forest Experiment Station, Argonne Experimental Forest. 60 p. 25 26 Argonne Experimental Forest Nicolet National Forest Figure 3-1. Location of the Argonne Experimental Forest within the Nicolet National Forest, Wisconsin. 27 Erdmann and Oberg (1973), Jacobs (1974), Tubbs (1977a, 1977b), Godman (1985), Erdmann (1986), and Niese and Strong (1992). Prior research (Niese and Vasievich 1990) identified five information needs that must be addressed successfully to compare the economic returns of even- and uneven-aged management. These research needs are listed in order below: 1. Long-term silvicultural trials. 2. Definition of silvicultural strategies. 3. Stand growth responses to management and harvesting treatments, especially effects on timber quality. 4. Future timber demands and prices. 5. Costs of different forest management practices. Long-term Silvicultural Trials--Argonne Experimental Forest The economic analysis of a cutting methods study must begin with a well-defined and documented biological production function. Long-term silvicultural trials are necessary to produce such a function. There are at least seven long-term silvicultural trials in northern and Appalachian hardwoods in the Eastern United States. The Argonne cutting methods study is the only replicated study (Figure 3-2). In the current study, biological data were first summarised from the three managed northern hardwood stands that compose the cutting methods study (NH-25) in the Argonne Experimental Forest. The original study had been established in 1951, to assess the effects of nine different cutting methods on tree and stand development in a 45-year old, even-aged, pole-sised northern hardwood stand. Early researchers laid out three 40-acre treatment replicates along a low ridge north of Route 32 near Hiles, Wisconsin (Figure 3-3). The study plots were second-growth northern hardwood stands typical of those found throughout the northern Lake States. They had developed from a period of cuts in mixed, old-growth stands in the early 1900's: first removing white pine (Pinus strobus L.), then removing hemlock (Tsuga canadensis L.), and finally removing the hardwoods with commercial clearcutting. 28 WV; /m Figure 3-2. Location of long-term silvicultural trials in northern and Appalachian hardwood timber types in the Eastern United States. 29 FR 2183 ............ I State Hwy 32 Figure 3-3. Location of the cutting methods study within the Iron-River Wabeno (Ma and Mia) soil mapping units on the Argonne Experimental Forest. 30 By 1950, regeneration from this series of cuts had developed into fully stocked pole-sized stands, with the following characteristics: 0 Average diameter: 6.0 to 9.2 in. 0 Trees per acre: 237 (z 4.6 in. d.b.h.). o Basal area: 85-114 square feet/acre. 0 Average age: 45-50 years. a Site index, sugar maple: 65 feet (50 years). 0 Species composition: 63% sugar maple, 4-10t each of basswood, white ash, yellow birch, Eastern hemlock, and red maple. Northern red oak, black cherry, paper birch, aspen, black ash (Fraxinus nigra), and American elm (Ulmus americana) were also present, but in insignificant (0.1-1.8 percent) quantities at the time. Scattered in the stands were occasional residual trees that had grown into large-diameter, heavy-branched, poor-quality trees. The Acer- Viola-Osmorbisa habitat type (Coffman et a1. 1984) and Iron-River Wabeno stony loam soils occurred over most (90 percent) of the study sites (Figure 3-3). Data Most of the study data were in hand-written, tabular or card format when received in 1990. Included in the data were: 0 Individual tree records for approximately 3000 trees in six uneven-aged treatments and three replications. The data included measurements taken every five growing seasons since 1951 on d.b.h., form class, crown class, number of logs and/or bolts, merchantable and total heights, mortality, ingrowth, reproduction, and tree grade if harvested. 0 Tree quality data that included log and/or tree grade and defect data from about 400 individual trees dating back to 1951. 0 Individual tree records for a total of 272 crop trees from three replications of a crop tree treatment installed in 1951. Researchers remeasured these trees in 1953, 1958, 1961, and 1966. (This study did not include these data due to difficulty in interpreting the changing definition for crop trees). 0 Cutting history and financial records for all treatments, both even- and uneven-aged. (The economic data were so site-specific that they were not useful for this analysis). 31 0 Regeneration, reproduction, and regrowth data for the even-aged treatments, beginning in 1958, including clearcut and shelterwood blocks. 0 Data on the species diversity of both woody and non-woody species. Data from a tree quality study, superimposed on the original cutting methods study in 1971, was received in 1991. About the same time, additional plot data on the even-aged treatments became available, and new regeneration measurements were taken for a tree diversity study. Description of Silvicultural Cutting Treatments This study evaluated nine treatments including the control. They are: 1. Control (no cutting) 2. Selection to residual basal area of 90 square feet/acre 3. Selection to basal area 75 4. Selection to basal area 60 5. Crop tree selection to basal area 60 6. Eight-in. stump diameter limit cut 7. Two-cut shelterwood with residual basal area 55 (after first cut) 8. Block clearcut with five-in. limit 9. Block clearcut with one-in. limit In treatments 1-7, each block treated was 2.5 acres; while in treatments 8 and 9, each block treated was 5 acres in size. Initial stand basal areas ranged from 85-114 square feet per acre (Table 3-1), but were not significantly different. The three basal area treatments were cut to light, medium, and heavy selection levels. These treatments left 90, 75, and 60 square feet of basal area, respectively, in trees 4.6 inches d.b.h. (diameter breast height) and larger. The cuttings favored commercial hardwood species while discriminating against balsam fir, ironwood, hemlock, and all cull trees (Stoecklerfi. 2Stoeckler, J.H. Cutting methods in pulpwood-sized northern hardwoods (A-3). Unpublished work plan (dated October 19, 1951). St. Paul, MN: U.S. Department of Agriculture, Forest Service, Lake States Forest Experiment Station. 18 p. 32 TABLE 3-1. Description and stand basal area of treatments (ft’lac) in 1951 and in 1990. 1951 1951 Treatment Description Before out After out 1990 Control Uncut through entire period 94 94 147 Light Cut to residual selection basal area of 90 ft’lacre 104 as 106 Medium Cut to residual selection basal area of 75 ftz/acre 98 77 87 Heavy Cut to residual selection basal area of 60 ft’lacre 85 62 73 Crop tree Crown-released 30-50 crop trees/acre: residual basal area was between 60-75 ft’lacre 87 63 87 Diameter Cut all trees limit with stump diameters of 8 inches and larger 84 23 124 Shelterwood* Cut to a residual basal area of 55 ftz/acre (88% crown cover), 114 55 30 Commercial Cut all trees clearcut 4.6” d.b.h. and larger 89 14 115 Complete Cut all trees clearcut 1.0" d.b.h. and L—i larger 89 0 93 1 I * First cut in shelterwood occurred in 1958. 33 In the crop tree treatment, marking crews left 25-40 crop trees per acre. Then, a crown thinning to a residual basal area of 60-75 square feet per acre allowed these dominant or co-dominant trees room to grow. Small trees were left adjacent to the crop trees as trainers where needed (Stoecklerfi. The diameter-limit cut was to an 8-in. stump diameter, taken at a 12-in. stump height, a common commercial treatment at the time the Argonne study began. The clearcutting treatments removed the stand in blocks of 5 acres. Cuttings left trees 5 in. and smaller in the commercial treatment and 1 in. or less in the silvicultural treatment. Foresters prefer the silvicultural clearcut because it does not allow poorly formed non-commercial trees to become dominant over the more desirable regeneration (Marquis 1967). The study added a shelterwood treatment in 1957, six years after the other treatments. Researchers prescribed a two-cut shelterwood system (Godman 1985) with 55 square feet of basal area left in the original cut (1958). Removal cuts occurred in 1965 and 1975, once the mixed hardwood regeneration was 4 to 8 feet high. Plot Measurements Each treatment block, 2.5 acres in size, was sampled by five 0.1 acre circular sample plots (Figure 3-4). The 5-acre clearcut blocks were each sampled with two 0.1 acre plots. Research workers made a 100 percent inventory of each 0.1 acre growth plot before and after cutting in the winter of 1951-1952, and estimated cull and saw log grade recovery from the volume cut in 1951-1952. End of the growing season inventories were repeated in 1957, 1961 before and after out, 1966, 1971 before and after cut, 1977, 1981 before and after cut, and 1986. All trees 4.6 in. d.b.h. and larger were measured to the nearest 0.1 in. and tallied by species. ’Stoeckler, J.H. Establishment report for cutting methods study (A-3) in pulpwood-sized northern hardwoods, p. 6. 34 / L C . Q9 00 . n» F O . o m JG Block 1 K‘) C709 Clearcuts 521232. \2 “3° Commerczal O Silvrcultural --.---“ I l \ Sheiterwood Figure 3-4. Sampling design for replication 1 (40 acres), Argonne. cutting methods study. Each of the smaller treatment blocks is 2.5 ac., with five 0.1 ac. sample plots (reproduced from an original drawing). 35 Trees were felled and sectioned in 1952 and 1969 for local volume table construction. Measurements for cubic foot volume included diameters inside bark beginning at a 1-foot stump to the nearest 0.1 in. at 8.3 feet intervals to a 4.0-in. top and total tree heights to the nearest foot. In addition, measurements on saw log trees (9.6 in. d.b.h. and larger) included Scribner board foot volumes to a variable merchantable top diameter that was limited by branches, defect, or other deformity, but not less than an 8.0-in. diameter inside bark (d.i.b.) top. Total cubic foot volumes inside bark were computed for each tree over 4.6 in. d.b.h. using Smalian's formula for the following tree portions: stump, stem in 8.3 foot lengths to a 4.0-in. d.i.b. top, the tip, and branches to 4.0-in. d.i.b. tops. Then, following Spurr's (1954) volume line method, workers constructed separate local cubic foot volume tables for sugar maple, basswood, yellow birch, and white ash (Erdmann and Oberg 1973). The study used sugar maple volume-line equations for calculating cubic foot volumes of other less commonly associated species such as red maple and American elm. Merchantable board feet and cordwood volumes were derived from the total cubic-volume estimates by use of appropriate converting factors for different mean stand diameters (Spurr 1954). The study defined a cord to contain 79 cubic feet of solid wood (Erdmann and Oberg 1973). A lack of data on merchantable heights made the use of the volume line equation method (Erdmann and Oberg 1973) described above necessary to determine stand volumes. However, beginning in 1972, the study included calculations of individual tree volumes for all harvests and residual stands, using merchantable height measurements, diameters, and appropriate deductions for cull. Modelling Growth and Yield (1992-2022) This study used the Lake States version of TWIGS (Belcher 1982, Shifley 1987) to model future growth and yield. Because of poor predictability for white ash and basswood, the author revised the TWIGS 36 equations for these species following the additive method developed by Holdaway (1985). The methods used for this adjustment are described elsewhere in this report (Chapter 6, pp. 92-98). Economic Evaluation For each of the above treatments, the author evaluated growth data to determine economic returns resulting from the various stand manipulations. The economic model estimated returns to treatment, given initial stand conditions, cost and price expectations, and changes in tree quality over time. This model was based on regression analysis of the variables that affected both residual stand value in 1992 after 40 years of treatment, and total stand value in 1992 (cut plus residual). In addition to treatment effects, the author tested the following variables for their effects on 1992 lumber value: number of white ash stems in initial stands, trees per acre in 1951, basal area of white ash at the end of the study, and the average tree size in 1951 for each treatment replicate. A benefit/cost analysis framework was used to compare economics of cutting strategies. A marginal economic analysis was conducted using Quick-Silver version 2.0 (Vasievich et a1. 1984), a forestry investment software program. All costs, prices, volumes, and stand conditions were entered into Quick-Silver version 2.0 (Vasievich et a1. 1984). Sensitivity analyses in Quick-Silver determined the response of each treatment to changes in discount rate, costs, and price expectations. Hardwood Management Costs Only direct or variable costs of management were used to compare, economic performance of the cutting methods. Since the fixed costs of the operation as a whole did not affect relative choice, they were excluded in this comparative analysis (Davis 1966). Thus, the financial performance of each treatment was compared with the control or no management treatment. 37 The study used cost survey data specifically from Lake States northern hardwood management to construct estimated costs of treatments. Variable costs included in the analysis were: cruising and marking for partial cuts: clearcut sale layout: post-harvest removal of non- commercial trees; periodic road maintenance; and annual administrative costs. Sensitivity to cost variation was tested with the following low, medium, and high cost assumptions, on a per acre basis (indicated in parentheses, respectively): cruise/mark for partial cut ($16,20,24): clearcut sale layout ($12,26,40); and periodic road maintenance (53,5,7). Most estimates of variable management costs for the nine treatments in the study were obtained from a recent Lake States study (Vasievich, Potter-Witter”. These costs were compared with other cost data from the Nicolet National Forest in Wisconsin, and cost data from previous Lake States studies (Winebar and Gunter 1984, Olson et a1. 1978, Hilliker et a1. 1969). The author also checked these costs for accuracy with costs incurred by consulting foresters in the northern Lake States. Administrative costs ($0.50/ac in 1952) were also assigned to each treatment and applied annually with a 3 percent annual increase. All nominal costs were adjusted to constant 1990 values using the Producer Price Index for all commodities (1990 a 100). Therefore, all economic analyses were done in real or constant dollar terms. Income and property taxes and land costs were not included in the analysis. Analysis of Hardwood Price Data Prices for northern hardwoods stumpage were obtained from 1954-91 reported prices from Michigan's Upper Peninsula (Michigan Department of ‘Vasievich, J.H.: Potter-Witter E. Costs of timber management practices in the Lake States, 1987-88. Unpublished draft manuscript (dated 8/16/90) of a cooperative study. East Lansing, MI: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station and Michigan State University, Department of Forestry: 16-19. (In author's personal file.) 38 Natural Resources 1954-91) and the Wisconsin Forest Products Price Review (Peterson, various dates). Of particular concern were the five major Argonne species--sugar maple, white ash, basswood, yellow birch, and red maple--although price series from Wisconsin (Peterson, various dates) were also analysed for Eastern hemlock, American elm, and mixed hardwoods. These prices were compared with transaction evidence from local sales in the Eagle River Ranger District of the Nicolet National Forest to assure comparability. The Producer Price Index (1990 - 100) was used to adjust these nominal Wisconsin stumpage prices for inflation. This study used transaction evidence because it is the method that most closely approximates market value. FORSight (Vasievich et a1. 1988), a timber price analysis software program, analyzed the long, medium, and short-term price trends for eight northern hardwood species. Each price series was analyzed using a natural log transformation model of the form: ln(P year) s a + b (Year) Where ln(P) - natural log of the reported stumpage price, and the corresponding year is the dependent variable. The b parameter is interpreted as the annual rate of compound price change. The author evaluated each cutting treatment using these estimated prices for the four major hardwood species and Eastern hemlock present on the study sites. Other hardwood sawtimber‘ comprised less than 6 percent of stand volume. Also, hemlock sawtimber from the early years (1951-72) of the study was priced at softwood pulpwood stumpage prices. These hardwood price trends were considered in the context of exogenous factors influencing hardwood prices, derived from the Hardwood Assessment Market Model (HAMM, Cardellichio and Binkley 1984), RPA ’Other sawtimber included small (<1.5t of basal area) amounts of American elm (Ulmus Americana L.), white birch (Betula papyrifera Marsh.), black ash (Fraxinus Nigra Marsh.), black cherry (Prunus serotina Ehrh.), and northern red oak (Quercus rubra L.). 39 timber assessment reports (Haynes 1990), and other sources. Finally, three hardwood price scenarios were chosen. These scenarios were the ones considered most appropriate for the five major species in northern Lake States hardwood markets (Chapter 4, p. 57). Then, the inflation- adjusted price estimates were applied by product to the cutting treatment yields. Sensitivity to price changes was incorporated in the analytical model with these low, medium, and high real price change assumptions: sugar maple (0, 1.8, 4 percent): basswood (0, 1.5, 4 percent); white ash (2, 3.8, 6 percent): yellow birch (-l, 0, 2 percent), red maple and mixed hardwoods (0, 2, 4 percent); mixed hardwood pulpwood (0, 1, 3 percent): and softwood pulpwood (-2, 0, 1 percent). Quality Change and Economic Return The financial value attributable to tree quality changes has two dimensions: value from trees already harvested, and the value of residual trees. In valuation of residual trees, each tree left has an opportunity cost, and must earn more than the discount rate to justify being retained (Murphy and Guldin 1987). This study evaluated both components of tree quality value for each cutting method. The average tree grades (Hanks 1976) of both harvested and residual sawlog trees, by treatment, were calculated for 1971, 1981, and 1991. Analysis of variance (ANOVA) tested which treatments had significant changes in quality due to silvicultural treatment. When these grade improvements occur, sawlog trees often have high value increases, and the decision to leave them for future growth can be justified (Nautiyal 1983). The economic analysis of tree quality changes included five uneven-aged treatments plus the control, but did not include the shelterwood, or either of the clearcut treatments, because quality data was not available for them. The framework for quality analysis considered the trees before quality treatments (1971) and after treatment (1992), after three 40 10-year periodic harvests had been made. Regression equations for factory lumber yields from graded hardwood trees, first developed by Hanks (1976), were used. In Hanks' (1976) equations, the dependent variable was lumber grade, and the independent variables were a constant term, (d.b.h.)’, merchantable height, and [(d.b.h.)’ x merchantable height). These regression equations are shown in Appendix 8. Hanks (1976) did not derive lumber yield equations for white ash: however, because of similarity between basswood, elm, and white ash form class and taper functions (unpublished data on file at the Argonne Experimental Forest!) the basswood lumber yield equations were also used for white ash and American elm. Lumber volumes were calculated for the following grades: FAS, FAS- 1F, Selects, #1 Common, #2A, #28, #2 Common, #3A, #38, and #3 Common. However, not all grades apply to all species. For example, in this study, the #2A and #28 grades applied only to red maple; and the #3A and #38 grades applied only to sugar and red maple, yellow birch, and paper birch. Because of very small variations in value, the grades FAS, FAS- lF, and Selects were combined into one grade (total FAS) for economic evaluation. Similarly, the #2A and #28 red maple grades were combined into the #2 Common grade for pricing outputs. Six grades were used in the final economic evaluation of hardwood tree quality--Total FAS, #1 Common, #2 Common, #3A, #38, and #3 Common. Application of Hanks'(1976) Tree Grade Lumber Yield Equations Cut and residual tree files were exported from Paradox. v.3.5 to Lotus‘”. The data included tree grade, d.b.h., merchantable length, and species. Gross board feet volumes (Scribner) for each tree were derived using the following equation: ‘Strong, Terry F.) Erdmann, Gayne G. Unpublished data on file at the Argonne Experimental Forest, Rhinelander, Wisconsin. iMention of trade names does not constitute endorsement by the U.S. Department of Agriculture. 41 (l/merch. length) + 0.723 (109» merch. length)] Het board foot volumes were then calculated by multiplying gross volume by percentage of cull in each tree in the spreadsheet. All data in the spreadsheet were sorted by replicate, species, and plot. Then, the Hanks (1976) equations for each lumber grade and species were entered into a Microsoft Excel. spreadsheet and volumes by grade were calculated for each tree. Finally, the author sorted the data again by treatments and replicate for final valuation. Valuation of Lumber Yields Lumber yields for each treatment were evaluated using prices from the Hardwood Market Report (Lemsky, 1971, 1981, 1991) and the Wisconsin Forest Products Price Review, lumber edition (Peterson 1981, 1991). Each sawtimber ( > 9.5 in. d.b.h.) tree's lumber yield value was calculated in 1971 and in 1991 for residual stands, and in 1971, 1981, and 1991 for harvested trees. Thus, the value gain or loss from grade changes was shown for each individual tree, and by summation, for each treatment over time. Overall Economic lvaluation Net treatment values were calculated from the value of all harvested stumpage, plus residual values of all commercial (> 4.6 in. d.b.h.) trees in 1992, less treatment costs. A total of nine treatment case files (not including sensitivity analyses) were created in Quick- Silver version 2.0 (Vasievich et a1. 1984). Quick-Silver has a simplified method for conducting marginal analyses. This method is called a combined case, where two separate Quick-Silver cases are entered, one for the treatment in question, and one for the control treatment. Quick-Silver then automatically changes all control transaction quantities to negative values for comparative 42 analysis. Thus, the program treats control revenues as opportunity costs (negative values) and control costs as opportunity revenues. Analysis of the combined cases showed the differences in financial performance between each treatment and the control. In this study, the difference in net present value (NPV) was the desired decision criterion. Analysts consider net present value to be a superior investment criterion in most benefit/cost applications (Randall 1981, Davis 1966). Net present values were calculated using real discount rates of 2, 4, and 6 percent and adjusted to 1990 U.S. dollars for reporting. Values of the residual merchantable stands in 1992 were included in all analyses to reflect total value for each treatment, after 40 years of management. Although a few other cutting methods studies do not include residual value in their economic analyses (Reed et a1. 1986, Erickson st a1. 1990), this study explicitly included residual value. Comparison of stand treatments over time must include changes in total treatment value, not just periodic revenues. Summary A benefit/cost analytical framework was used to compare total economic returns for nine northern hardwood cutting methods. Cutting yields and returns were analyzed for 40-year (1952-92) and 70-year (1952-2022) investment horizons. Tree quality changes were explicitly included in an analysis of treatment lumber values that considered potential grade and yield of every sawtimber-sized tree for 1971-92. Species diversity trade-offs were also analyzed using Shannon's diversity index to measure hardwood regeneration (Chapter 7). Regression analysis was used to find price trends from time series data and to analyse changes in hardwood tree quality and lumber yield. Data from surveys were used to estimate hardwood treatment costs. Sensitivity analyses tested economic results under changing price, cost, and discount rate assumptions. CHAPTER ‘8 HARDWOOD PRICI EXPECTAIIOIB Introduction This section presents real stumpage price trends in Wisconsin and Upper Hichigan for selected hardwood species and products for the period 1950 to 1991. It also reviews past research on hardwood prices and expectations of future hardwood prices. Based on these sources likely hardwood price scenarios are reported for the commercially important northern hardwood species occurring on the study site. In a market economy, prices provide an intricate system of signals and incentives to coordinate the activities of firms and individuals. Market prices serve as signals to producers and consumers alike, showing scarcity or abundance, regulating production and consumption. 8ut market prices may not always be efficient prices, because suppliers of resources often have insufficient information to assess the future. Timber markets often contain imperfect information, making them inefficient. However, the advent of stumpage price reporting by various agencies (Rosen 1984, Hichie and kamets 1987) and widespread dissemination of price information has improved this situation in recent years. Stumpage price is the price that a buyer, usually a logger, pays to the owner of forest land for the privilege of cutting and taking timber. This payment is the timber's scarcity rent (Tietenberg 1988). Stumpage can also be defined as the value of standing timber. It is determined by the complex interaction of timber stand, location, and market characteristics. Since the characteristics determining stumpage price are site specific and vary over time, price forecasting is difficult. However, current and past prices inevitably shape landowners' expectations of future timber prices. These future expectations about stumpage prices can have a dramatic effect on forest cutting decisions both now and in 43 44 the future (Dennis and Remington 1985). For northern hardwoods, stumpage price expectations can influence the cutting system chosen (even- or uneven-aged), the particular methods used, and economic returns achieved. Therefore, it is important to consider hardwood stumpage price expectations in this study. Historical Price Trends Historically, forest products have been atypical of most natural resource commodities because they have had real long-term price increases (Hanthy 1978, Potter and Christie 1962). Host natural resource industries have had stable commodity prices and decreasing employment-per-unit-of-output ratios, suggesting widespread adoption of capital-intensive technology (Hanthy 1978). Forest products, and sawlogs and lumber in particular, are resource commodities that have not always followed these patterns. Hanthy's (1978) composite index of forest product prices showed rising real prices between 1870 and 1950, relatively stable prices between 1950 and 1970, and rising real prices since then. He attributed most increases in real forest product prices to sawlogs. Hanthy described sawlog prices for the 100-year period 1873-1973 as a rising ratcheting trend, with 1950-70 being a fairly stable period. Barnett and Horse (1963) examined the real cost trends for sawlogs, and the trend of cost of net sawlog output relative to a deflator of Gross National Product for non-extractive industries. They found that the relative cost of sawlogs rose substantially, a fact they cited as confirmation of the historical scarcity of sawlogs. Potter and Christy's (1962) work on lumber prices found similar price trends to those reported earlier for sawlogs. They suggested that the rising real price trend for lumber may have been understated. This occurred because for many years the shift of the industry from the East to the West Coast had put the output (priced at the mill) farther from 45 the market. Lagging productivity had been cited as the chief cause of lumber price increases until the 1950's (Duerr 1960). Irland's (1974) work showed that increases in lumber prices had not come steadily, but mainly in building booms and inflationary periods following several wars in this century. His interpretation of the rising real prices for lumber parallels Manthy's ratcheting rising price trend for sawlog prices. Irland also cautioned that lumber prices may not be a good proxy for timber prices. For example, from about 1860 to 1900, lumber prices were stable while timber prices rose. Scarcity shown in the rising price of timber had been camouflaged by declining transportation and milling costs. Recent work by Potter-Witter and Lacksen (1993) has found rising real sawtimber stumpage prices for red maple, sugar maple, and basswood in Hichigan between 1954 and 1991. They documented somewhat different market trends between Hichigan's Upper and Lower Peninsulas, with most of the larger price increases occurring in the Upper Peninsula. In contrast to sawtimber (and lumber), pulpwood prices have been remarkably stable since 1900. Hanthy (1978) noted this real price stability as unusual, given rapidly rising consumption of pulp and paper. The pulp and paper industry, unlike the lumber industry, increased its productivity (output per unit of labor) by an average 133 percent in the 40 years prior to World War II, blunting the effect of increasing wages and costs (Duerr 1960). However, a few researchers have recently documented significant real price increases for some pulpwood species and real price declines for others. In New Hampshire real stumpage prices for hardwood pulpwood increased an average of 3.9 percent annually while spruce/fir pulpwood declined 3 percent annually between 1964 and 1983 (Dennis and Remington 1985). Further work (Dennis 1989) found that real prices for both hardwood and spruce/fir pulpwood declined 0.5 percent for the 1979-88 period. 46 These trends reflect the substitution of hardwood for softwood pulp as the Eastern U.S. region has shifted from sulfite mills to semi- chemical mills. Other regions of the U.S. have been making similar changes in pulp and paper making industries, with similar changes in relative pulpwood stumpage prices. The result has been short-term periods of rising real prices for hardwood pulpwood and declining real prices for softwood pulpwood, followed by a stabilizing trend as Dennis (1989) observed in New Hampshire. Despite varying interpretations of the past century of stumpage and lumber prices, most analysts conclude that, in the long run, pulpwood prices have generally remained stable, while sawlog and lumber prices have seen real price increases averaging between 1.5 and 2.0 percent per year. Analysis of historical stumpage prices for this study is made in the context of these more general trends. These historical price trends are pertinent to recent work done on the demand for hardwood stumpage and hardwood lumber. The reason is that hardwood price expectations are always shaped by the current and expected supply and demand for hardwood stumpage and lumber. The following sections examine the status of hardwood lumber markets in the United States and the various factors that influence the supply and demand for hardwood products and, therefore, hardwood stumpage prices. Hardwood Markets and Prices-~Wational Demand and Price Trends In the past 15 years hardwood lumber markets have had generally strong growth (Dempsey and Luppold 1992). Demand for hardwood lumber increased especially strongly in the period 1982-89, from 8 billion board feet to 11.2 billion board feet. In the decade of 1977-87, overall hardwood lumber use increased by 38 percent. However, these overall upward trends include dramatic changes in end uses of hardwood lumber. For example, in the same lO-year period hardwood lumber exports were up by 287 percent, while hardwood lumber usage in railroad ties and 47 timbers (-46 percent) and wood household furniture (-15 percent) both declined. Some current research (Dempsey and Luppold 1992) suggests these may be long-term trends. The major domestic uses of hardwood lumber in the past two decades have been pellets and containers, hardwood dimension, wood household furniture, and millwork. These four industries combined accounted for 7.6 billion board feet or nearly three-fourths of domestic consumption in 1987 (Dempsey and Luppold 1992). Thus, the future of hardwood stumpage prices is, to a large extent, determined by the outlook for these four industries. The following sections explain the hardwood exports and wood household furniture sectors in greater detail. W The export market nearly tripled its consumption of hardwood lumber between 1977 and 1987, and continued a positive real growth rate through 1990 and 1991, despite a U.S. recession. Host hardwood exports are in high quality logs and lumber, particularly of premium species such as red and white oak, white ash, and black cherry. Sugar maple, though a valuable hardwood, has not shared this status as a ”glamour" species. However, this situation may be changing. Sugar maple exports have shown a strong increase since 1989, as Japanese and Duropean buyers and brokers have begun to purchase large volumes of high-grade logs in the northern Lake States. lxports of hard maple logs to Korea have shown particularly strong growth in the past five years (LuppoldU. Interest in bird's-eye maple has been particularly strong. Prices for Michigan's Upper Peninsula bird's-eye have gone as high as $28,000 per thousand board feet, or more (Hros et a1. 1990). Demand for bird's-eye 'Luppold, W. Personal Communication, October, 6, 1992. 48 maple may be contributing to upward price pressure on woods-run sugar maple (lrdmann’) . need_lenaehel§_znrnitnrs The wood household furniture industry is the third largest domestic user of hardwood lumber, and the largest end user for total value. Hew housing construction feeds demand for hardwood lumber, though there is also a significant market for wood furniture in remodeled homes. Two opposing trends have affected hardwood demand for household furniture. First, the amount of hardwood lumber used in each unit of furniture has been reduced by half as less expensive reconstituted wood products such as hardwood plywood, particle board, and medium density fiberboard have replaced solid wood components. This trend reflects the ingenious nature of furniture manufacturers to find less expensive substitutes for hardwood lumber (Cardellichio and Binkley 1984). Second, the actual number of furniture units produced has more than doubled between 1940 and 1980, and has shown continued strong growth since then. If the decreasing use factor is combined with the growth in furniture production, there is a slight upward trend in the consumption of hardwood lumber for furniture for 1940-80, a decline of 15 percent for 1977-87, and a stable trend in household furniture demand since 1987 (Cardellichio and Binkley 1984, Dempsey and Luppold 1992). Recovery from the 1990-91 recession has resulted in a modest upward trend in consumption of hardwood lumber for furniture. For many years, perceptions of the hardwood sector have been shaped by the belief that sawtimber of the select species is declining in availability. These perceptions are not always confirmed by official reports of hardwood sawtimber availability but recent studies have aErdmann, 0.0., Consulting Forester. Personal Communication, July 26, 1992. 49 confirmed forest industry's beliefs that the availability of quality hardwood sawlogs is in fact declining (Dennis and Remington 1985, Kingsley and De8ald 1987). Real prices in the Lake States have increased for several furniture-grade hardwood species, including red oak, white ash, and black cherry. Declining real prices seen in the 1960's and 1970's for red and sugar maple and basswood have stabilized and increased since about 1975. The quality issue has dominated most discussions of the future of hardwood markets. However, increases in the production of low-grade hardwood products (pallets, 088, fuelwood, etc.) and substitution of hardwoods for softwoods (e.g., pulp and paper) also may be a driving force in increasing hardwood demand (Haynes 1990). Therefore, the species most able to meet a variety of product specifications may be those that sustain future real price increases. During the 1980's hardwood harvest from growing stock grew by over 40 percent (Haynes 1990), and the share of harvest used as pulpwood is expected to increase to nearly 50 percent by 2040. Analysts expect continued growth in hardwood timber inventories will prevent stumpage price increases for hardwoods, except for sawtimber, especially in the north (Haynes 1990). These forecasts also show that tree quality is important in determining future hardwood price increases. The world furniture industry generates most demand for high-value U.S. hardwood sawlogs. Several global economic trends are also affecting supply and demand for U.S. hardwood forests and the U.S. furniture industry. Host of these trends have developed rapidly since the late 1970's. They include: increased U.S. wood furniture imports: increasing substitution of American hardwoods for Southeast Asian hardwoods by Pacific Rim manufacturers: development of a strong "baby- boomer“ market for high-end furniture: strong interest by Asian furniture producers in buying U.S. hardwood timber supplies; and changes in global exchange rates that have increased economic incentives for SO U.S. firms to export the better hardwood logs (Smith and West 1990). These trends point to an emerging global demand, and possibly higher prices for, the better-quality U.S. hardwood timber resources. Hardwood Price Analysis an The basic price data analysed for this study came from stumpage data from Hichigan's Upper Peninsula. Statewide stumpage sales from Wisconsin and local sales from the Nicolet National Forest were also considered for comparative purposes. The Hichigan data were used because the study area is within the Upper Peninsula market area. The Wisconsin data originated from Hinnesota and Wisconsin Forest Product Prices: A Historical Review 1950—1980 (Lothner et a1. 1982), and from the University of Wisconsin Extension (UWEX) Forest Products Price Review, Timber Edition (Peterson, various dates). The Lothner document summarises past trends in forest products prices for stumpage, harvested wood, and lumber sold in Hinnesota and Wisconsin. The UWEX price review newsletters report similar information by species in a market news format. Both sources report ranges and average prices for stumpage. The prices reported in these sources (and used in this research) are median stumpage prices, unweighted by sales volume. All reported prices represent timber sales throughout the state of Wisconsin and reflect stumpage prices at the time of harvest. All stumpage prices are on a Scribner Decimal C log rule. The major data source used in the analysis was the Region 1 (Upper Peninsula) stumpage data provided by the Hichigan Department of Natural Resources. Sugar maple stumpage price data from the Nicolet National Forest (1977-91) were checked for comparative purposes only. The Hichigan Upper Peninsula data are for the region just to the north and east of the study site. These data generally are for the 1954-91 period, from competitively-bid government sales. The Nicolet 51 data are from the National Forest area that surrounds the Argonne Experimental Forest. They are also competitive stumpage prices for locally sold, public timber sales. The Forest Economics Research Unit (of the USDA Forest Service North Central Forest Experiment Station, in East Lansing, Hichigan) compiled price data into a uniform reporting format and coded into ASCII data files. This research project made liberal use of this consolidated data base. Price series for nine species and product combinations were analyzed, representing nearly all of the commercial volume on the cutting methods study site. These species and product combinations are: W 1. Sugar maple 6. American elm 2. Basswood 7. Hixed hardwoods 3. Yellow birch 8. Eastern hemlock? 4. White ash Eglpgggg 5. Red maple 9. Hixed hardwood The mixed hardwood category was used to estimate price trends for other hardwood sawtimber on the Argonne. These other hardwood species included small (< 1 percent of basal area) amounts of white birch, black ash, black cherry, and northern red oak. Except for American beech (not present on the study site), the above species, in various combinations, are typical of northern hardwood forest types in the northern Lake States. ’Though not a hardwood, hemlock prices were analyzed because Eastern hemlock is a significant component of this and other northern hardwood forests. 52 W Three different methods are commonly used to evaluate stumpage: l) transaction or market evidence: 2) value in use or cost/income; or, 3) residual value approach. This study used transaction evidence for several reasons, discussed below. The transactions evidence approach to stumpage valuation is based on statistical analysis of past timber sales in the same area. It involves obtaining the actual sales records from sales of similar and nearby timber. This is generally the preferred method because it directly measures what actual stumpage has sold for, in the market. Transactions evidence, also called 'appraisal-by-comparison,' has been used effectively to value stumpage for public and private timber sales (Anderson 1976). Legally, transactions evidence has been the method most often upheld by judges in court cases, rather than expert opinion (Sizemore et a1. 1965). One Kentucky judge summed up the practical advantages of transactions evidence this way: ”Of course market value is the price at which an article sells in the open market. This price is fixed by sales actually consummated. Such sales, when made under normal and fair conditions, are necessarily a better test of the market value than the speculative opinions of witnesses; for truly here is when 'money talks." (Sizemore et a1. 1965) To be valid for valuation purposes, a transaction evidence price must represent the high bid in a competitive market. The market defines competition as a case in which there are at least two parties bidding for the same stumpage. Advantages of the transaction evidence approach include: its reliance on existing data from past timber sales, simplicity and ease of operation, low cost, and less reliance on end product valuation (Vasievich et a1. 1988). The cost or income approach has limited application, mostly to the valuation of immature timber, and is not appropriate for short-term timber sale contracts. The residual value approach is sometimes used by 53 the USDA Forest Service. It works backward from an estimate of the value of the manufactured end product such as lumber. The value of various costs of logging, hauling, manufacturing, and profit and risk margins are subtracted from end product value, leaving a ”residual” value for the stumpage. A major problem with the residual value approach is that it is costly and time consuming to use (USDA Forest Service 1988). It also has the disadvantage of overlooking competitive factors that can lead to higher or lower prices than the average. Being based on what the firm of ”average“ efficiency can afford to pay for stumpage, the derived residual value approach does not necessarily evaluate market price where marginal costs equal marginal revenue but instead, price at a lower level (Buongiorno and Young 1984). Since the concern here is with fair market value rather than appraised value, the derived residual value approach was not appropriate to this study. Given the availability of appropriate data and its acceptance by the research community, transactions evidence was deemed the best method to use. W Long, medium, and short-term price trends (Tables 4-1, 4-2) for the nine northern hardwood species product combinations were analyzed using FORSight (Vasievich et a1. 1988), a timber price analysis software program. FORSight generates a variety of useful statistics. These include mean values and standard deviations for model variables, t-statistics, F-statistics, slope and intercept coefficients, ANOVA, and R-squared and adjusted R-squared values. Unless otherwise specified, the long trend was for 1950-90 or 1954-90, the medium trend for 1970-90, and the short one for 1980-90. Each price series was analyzed in FORSight using a natural log transformation model of the form: 54 ln(P,.) - a + b (Year) Where ln(P) - natural log of the reported stumpage price, and the corresponding year is the dependent variable. The b parameter is interpreted as the annual rate of compound price change. In the above model, a regression coefficient (b parameter) above zero indicates that an increase (decrease) in the independent variable (time) results in an increase (decrease) in the dependent variable (natural log of real price). The Student's t-statistic is used to test the null hypothesis that the b parameter is statistically equal to zero. The model's significance is tested similarly with F-values. W In determining price changes, the regression method is preferable to merely taking prices at the beginning and end of the period(s) and calculating the annual average increase. By incorporating timber price variation in all years of the series, the regression method includes more information in the price estimates (Holmes et a1. 1990). The natural log transformation is used because it transforms the multiplicative coefficients into a more convenient sum that can easily be estimated by using linear regression. Once transformed, the regression is kept in its log form because of the ease of interpreting the b coefficient, because the change in the log of x (yearly price) is equal to the relative change in x itself (Wonnacott and Wennacott 1984). RIBLB 4-1. Peninsula sawtimber stumpage price trends. Regression equations used to estimate Hichigan Upper “ Dependent Independent variables Variable: natural log of reported stumpage Trend price for: Period Constant (a) Year (b) Sugar maple 1954-91 3.0201 0.0180** 1970-91 1.1788 0.0404** 1980-91 -S.1971 0.1141** Basswood 1954-91 3.2966 0.0150* 1970-91 -0.7068 0.0576** 1980-91 -5.9920 0.1191** White ash 1980-91 -0.8889 0.0560** Yellow birch 1954-91 4.6797 -0.0008 1970-91 3.1703 0.0174* 1980-91 2.1083 0.0295* Red maple 1954-91 1.5022 0.0290** 1970-91 0.5613 0.0406** 1980-91 4.2974 -0.0026 Hixed hardwoods 1972-91 -2.3720 0.0722** 1980-91 -6.2461 0.1170** Hixed hardwood 1957-91 -0.5393 0.0221** pulpwood 1970-91 -2.9327 0.0510** 1980-91 -11.5075 0.1511ff m“ * - Significant at the a - .05 level ** - Significant at the a - .01 level 56 TABLE 4-2. Regression equations used to estimate Wisconsin sawtimber stumpage price trends. I Dependent Independent Variables Variable: natural log of reported stumpage Trend price for: Period Constant (a) Year (b) Sugar Haple 1950-90 5.3328 -0.0105* 1970-90 4.4318 0.0004 1980-90 1.1377 0.0388* Basswood 1950-90 5.5337 -0.0143** 1970-90 4.5518 -0.0023 1980-90 -0.4563 0.0562** White ash 1967-90 1.8179 0.0307** 1980-90 -2.3108 0.0791** Yellow birch 1950-90 6.2273 -0.0209** 1970-90 6.6328 -0.0263** 1980-90 2.2098 0.0254* Red maple 1950-90 4.2845 -0.0001 1970-90 3.6568 0.0074 1980-90 -0.0330 0.0504** American elm 1950-90 3.8064 0.0055* 1970-90 1.7425 0.0310** 1980-90 -3.8647 0.0965** Hixed hardwood 1972-90 5.2692 -0.0129 1980-90 3.7212 0.0049 Eastern hemlock 1950-90 6.0567 -0.0310** 1970-90 5.9851 -0.0302** 1980-90 3.3975 -0.0003 Hixed hardwood 1972-90 -l.1111 0.0323* pulpwood 1980-90 1.0342 0.0074 * 8 Significant at the u e .05 level ** - Significant at the a - .01 level 57 TABLE 4-3. Stumpage price scenarios used in analysis, expressed in average real price increases (decreases) per year. Low Price Hedium Price High Price Scenario Scenario Scenario Sugar maple 0% 1.8% 4% sawtimber Basswood 0% 1.5% 4‘ sawtimber White ash 2% 3.8% 6% sawtimber Yellow birch -1% 0% 2% sawtimber Red maple 0% 2% 4% sawtimber Hixed hardwood 0% 2% 4% sawtimber Hixed hardwood 0% 1% 3t pulpwood Softwood pulpwood inc. hemlock -2% 0% 1% sawtimber Results and Discussion W Price trends for sugar maple were especially important to this study, because 60-85 percent of the study site's individual trees and volumes were sugar maple. Over the 1954-1991 period, real sugar maple sawtimber prices in Hichigan's Upper Peninsula showed a real price increase of 1.80 percent, while 1970-91 was up 4.04 percent annually and the last decade (1980-91) increased by 11.41 percent (Figure 4-1 and Table 4-1). The Nicolet National Forest data showed that sugar maple sawtimber had an average annual real increase of 3.33 percent for the period 1979-90. In Wisconsin, sugar maple prices for 1950-90 declined 1.05 percent (Figure 4-2 and Table 4-2). However, despite the long-term decline in Wisconsin sugar maple sawtimber prices, when the 40-year 58 WBF Scribner 200 150 100 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-1. Reel (1990 S) Michigan Upper Peninsula sugar maple sawtimber stumpage prices. 19544991 (Michigan Department at Natural Resources data). SIMBF Scribner 200 150 100 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4.2. Reel (1990 S) Wisconsin sugar maple sawtimber stumpage prices. 1950-1990 (Wisconsin Forest Products Price Review). S9 analysis period was divided into two 20-year periods, a different price trend picture emerged. Then, real sugar maple stumpage prices were nearly stable (+0.04 percent) for the 1970-90 period, and showed an average 3.88 percent annual increase for the 1980-90 period. Based on current market activity, sugar maple will experience the greatest demand for pallet wood and for high-quality furniture and export logs. The development of specialty markets such as bird's-eye maple is also indicative of the growing interest in the sugar maple resource. Demand for the mid-grades of sugar maple logs and lumber will not be as great, because the end users of these grades (such as flooring mills) have not experienced rapid growth. However, there is a large surplus of #2 sugar maple sawlogs in the region. New technologies that can produce high-quality furniture parts from lower-quality materials may create added demands on this currently under-used resource. Price scenarios for sugar maple were difficult to predict because of the resource's current abundant supply, yet its rapidly growing demand. However, price trends since 1970 were probably more indicative of future prices than the long-range price series. The author estimated that real sugar maple sawtimber prices would have a low scenario of a 0 percent increase, a medium scenario of a 1.8 percent increase, and a high scenario of a 4 percent increase (Table 4-3). am Real basswood sawtimber prices showed a 1.50 percent increase in the U.P. for 1954-91 (Figure 4-3), but an average 1.43 percent decline in Wisconsin for 1950-90 (Figure 4-4). In the past two decades, basswood showed a 5.76 percent increase in the U.P., but a slight decrease (-0.23 percent) in Wisconsin. Since 1980 prices have shown significant annual increases both in Wisconsin (+5.62 percent) and in the U.P. (+1l.9l percent). 6O $IMBF Scribner 200 150 100 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-3. Reel (1990 3) Michigan Upper Peninsula basswood sawtimber stumpage prices, 1954-1991 (Michigan Department of Natural Resources data). SIMBF Scribner 200 0 50 55 60 65 70 75 80 85 91 Year Figure 4-4. Real (1990 3) Wisconsin basswood sawtimber stumpage prices. 1950-90 (Wisconsin Forest Products Price Review). 61 Basswood has had increased demand for carving stock and, to a lesser extent, keystock in the past few years. Current technological trends may favor greater demand for basswood in the future. Hany mills now have improved debarking machinery that is making it easier to remove basswood's stringy bark. Also, basswood's versatility makes it usable for pulp, pallets, and solid wood products. Based on the more recent (1970-91) price history for basswood, the author expects at least modest real price increases in the future. Therefore, real basswood sawtimber prices were projected to have a stable low price scenario, a 1.5 percent increase medium scenario, and a 4 percent increase high scenario. My White ash was the only species that showed consistently high real price increases in both Hichigan's Upper Peninsula and Wisconsin. In the Upper Peninsula on state lands, white ash showed an average real price increase of 5.60 percent for 1980-91 (Figure 4-5). In Wisconsin, median white ash prices had average annual real increases of 3.07 percent between 1967 and 1990, and 7.91 percent annually for 1980-90 (Figure 4-6). As red oak sawtimber removals continue to approach or exceed annual growth (Cooney 1992), and as oak types convert to maple and other species, white ash will provide an increasingly attractive alternative to scarce red oak. White ash has several desirable qualities: fast growth, fairly easy regeneration in both even- and uneven-aged management (Godman 1985), and excellent machining and finishing properties. Because it can substitute for red or white oak, its price often follows the price of red oak, but at a lower level. 62 $IMBF Scribner 200 150 100 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-5. Real (1990 S) Michigan Upper Peninsula white ash sawtimber stumpage prices, 1978-91 (Michigan Department of Natural Resources data). $IMBF Scribner 200 150 100‘ A?‘ l r ' A ‘ ' so 1' 1 I , 0 50 55 60 65 7O 75 80 85 91 Year Figure 4-6. Real (1990 S) Wisconsin white ash sawtimber stumpage prices. 1950-1990. (Wisconsin Forest Products Price Review). 63 Significant real white ash price increases should be the norm for the northern Lake States. The following scenarios were based upon price performance since 1967: a low scenario of +2 percent, a medium scenario of +3.8 percent, and a high scenario of +6 percent. 111.191.111.911 Yellow birch sawtimber prices were essentially stable (-0.08 percent) for 1954-91 in the U.P. (Figure 4-7), but declined significantly for the long trend period in Wisconsin (-2.09 percent, Figure 4-8). For 1970-91 U.P. birch prices climbed 1.74, while in Wisconsin they declined at an average annual rate of 2.63 percent. Prices have appeared to ”bottom out," evidenced by rising real prices both in the U.P. (+2.95 percent) and in Wisconsin (+2.54 percent) for the last decade. However, median real prices (1990 $) for yellow birch stumpage in 1990 were only $77.00/H8F (Scribner) in Wisconsin compared with $153.46 in 1957. In the Upper Peninsula median real yellow birch prices (1990 $) had surpassed their 1957 historic high of $140.00, reaching $148.17 by 1991 (Figure 4-8). Scarcity of yellow birch sawlogs may contribute to modest future price increases for this species. Only a generation ago yellow birch was the premium northern hardwood species of the northern Lake States. It was used mainly for furniture, veneers, airplane parts, and doors from the late 1940's to the 1960's. Today the yellow birch resource and its economy are barely viable compared to their past prominence, the victims of excessive harvesting by industry and a lack of regeneration amid competition by sugar maple. Though among northern hardwoods yellow birch has slipped from second to fifth place in commercial importance on the Argonne and in the region, it is still a high-value species. Despite a long-term (1950-80) real decline in birch prices for Wisconsin, yellow birch prices have stabilized in the region. 64 $IMBF Scribner 200 / WM 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-7. Real (1990 S) Michigan yellow birch sawtimber stumpage prices. 1954-91. (Michigan Department 01 Natural Resources data for Upper Peninsula). SIMBF Scribner 200 150 VAN 1 "I" i ‘ 100" .“~'\ 4 V ’1’ < M 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-8. Reel (1990 6) Wisconsin yellow birch sawtimber stumpage prices. 1950-1990 (Wisconsin Forest Products Price Review). 65 The success of current yellow birch regeneration efforts will decide the future viability of these markets and the species' price performance. Price scenarios for yellow birch sawtimber were estimated as a 1 percent real decline (low), a stable scenario (medium), and a 2.0 percent increase (high). mm In the Upper Peninsula, a 37-year trend (1954-91) showed an average annual 2.90 percent increase (Figure 4-9) in real red maple prices. By contrast, Wisconsin real red maple sawtimber prices were virtually stable (-0.01 percent) over the 40-year period 1950-90 (Figure 4-10). The U.P. had a 4.06 percent increase between 1970 and 1990, more than five times as great as Wisconsin's (0.74 percent). However, while Wisconsin showed a highly significant 5.04 percent increase for the most recent decade, the U.P. showed a slight (-0.26 percent) decline. Obviously, there are local variations in markets that strongly influence red maple price behavior. Red maple has increased faster in volume and acreage since 1968 than any other species (or type) in the Lake States (Spencer and Hahn 1984, Spencer et a1. 1988). It has also seen greatly increased usage in products ranging from 088 to upholstered furniture to flooring. Along with its increased usage and volumes red maple has experienced marked changes in its price behaviors. Potter-Witter and Lacksen's (1993) work showed average annual real price increases of 3.2 percent for Upper Peninsula red maple since the late 1950's. This analysis showed similar (+2.9 percent) real price increases for a slightly longer period. Wisconsin's long-term red maple series was stable, but annual compound price increases of 0.74 percent and 5.04 percent were observed for 1970-90 and 1980-90, respectively. 66 SIMBF Scribner 200 150 100 50 50 55 60 65 70 75 80 85 91 Year Figure 4-9. Reel (1990 S) Michigan Upper Peninsula red maple sawtimber stumpage prices. 1954-1991 (Michigan Department 01 Natural Resources data). SIMBF Scribner 200 1 50 100A». RNA/me. n .A WV U . /\«. w u V341" 0 50 55 60 65 70 75 80 85 91 Year Figure 4-10. Real (1990 8) Wisconsin red maple sawtimber stumpage prices. 1950-1990. (Wisconsin Forest Products Price Review). 67 Based on the increase in its marketability, and on the proximity of the study site to Upper Peninsula markets, the author expects red maple to have price scenarios of stable (low), +2 percent (medium), and +4 percent (high). W American elm prices in Wisconsin were on an upward trend for all three projection periods. Between 1950 and 1990, prices rose an average of 0.55 percent annually (Figure 4-11). The increase was 3.10 percent for 1970-90 and 9.65 percent for 1980-90. Prices were essentially stable for 1950-78 (-0.24 percent), but large real price increases have occurred since then. Dutch elm disease has created a scarcity of elm sawtimber, driving up prices since it swept the northern Lake States in the 1970's. Although American elm is a minor component in today's northern hardwood forests, many stands still have an elm component, particularly those stands on more mesic sites. Due to the scarcity of mature elm and continued demand, American elm scenarios were estimated as: +1 percent (low), +3 percent (medium), and +5 percent (high). W Eastern hemlock in Wisconsin had average annual real price declines of 3.10 percent for 1950-90, and 3.02 percent for 1970-90 (Figure 4-12). A steep 30-year decline in real price may have ”bottomed out" by 1980, because prices have been nearly stable (-0.03 percent) since then. This long-term decline in real hemlock prices was even more pronounced than the real decline in yellow birch prices. No new commercial markets are expected for hemlock sawtimber. Hemlock price scenarios were estimated as: -2 percent (low), stable (medium), and a 1 percent increase (high). 68 SIMBF Scribner 200 1 50 100 501 VVYYY VW 0 50 55 60 65 70 75 80 85 . 91 Year Figure 4-11. Real (1990 3) Wisconsin elm sawtimber stumpage prices, 1950-1990. (Wisconsin Forest Products Price Review). SIMBF Scribner 200 150 100 50 0 . 50 55 60 65 7O 75 80 85 91 Year Figure 4-12. Reel (1990 8) Wisconsin hemlock sawtimber stumpage prices. 1950-1990. (Wisconsin Forest Products Price Review). 69 W Hixed hardwood sawtimber displayed contrasting price behavior in Wisconsin and the Upper Peninsula. This disparity may be due to changing definitions and composition of sales of ”mixed” hardwood. Upper Peninsula mixed hardwood sawtimber prices increased 7.22 percent annually for the period 1972-91 (Figure 4-13). But in Wisconsin, mixed hardwood sawtimber declined in value an average of 1.29 percent (Figure 4-14). For 1980-91, Hichigan's Upper Peninsula saw a very substantial increase (11.70 percent, Table 4-1), while Wisconsin's mixed hardwood sawtimber showed a slight annual increase in real price (0.49). Because of the difficulty of defining mixed hardwood, and because the observed trends in the U.P. and Wisconsin were distinctly different, price scenarios were estimated as: stable (low), a real increase of 2 percent (medium), and a real increase of 4 percent (high). W In Hichigan's Upper Peninsula (Figure 4-15), mixed hardwood pulpwood prices showed a 2.21 percent average annual increase for 1957- 91, a 5.10 percent increase for 1970-91, and a 15.11 percent real price increase for 1980-91. Available data for Wisconsin showed a 3.23 percent real increase in mixed hardwood pulp prices from 1972-90 (Figure 4-16). Real price increases (+0.74 percent) appear to have leveled off since 1980. Hichigan Upper Peninsula real prices (1990 S) were in the $1.75-2.75 per cord range from 1970-80, whereas Wisconsin's average real prices were in the $3.00-5.00 range for the same period. Since 1980, the steep climb in Upper Peninsula real prices has made them comparable to Wisconsin hardwood pulp prices (Figure 4-15 and Figure 4-16). Hardwood pulpwood stumpage markets have played an important part in the recent development of forest products industry in the northern Lake States. Several new mills and mill expansions led to significant real price increases (3-15 percent), especially in the 1970's in 7O SIMBF Scribner 200 1 50 1 00 50 0 50 55 60 65 70 75 80 85 91 Year Figure 4-13. Reel (1990 S) Michigan Upper Peninsula mixed hardwood sawtimber stumpage prices, 1972-1991 (Michigan Department of Natural Resources data). SIMBF Scribner 150 100 50 55 80 85 70 75 80 85 91 Year Figure 4-14. Reel (1990 5) Wisconsin mixed hardwood sawtimber stumpage prices. 1972-1990 (Wisconsin Forest Products'Price Review). 71 $lCord 10 4 ’1‘ ‘ 4 2 ‘V' 50 55 60 65 70 75 80 85 91 Year Figure 4-15. Real (1990 S) Michigan Upper Peninsula Hardwood pulpwood stumpage prices, 1954-1991 (Michigan Department of Natural Resources data). $lCord 10 0 50 55 60 65 70 75 80 85 91 Year figure 4-16. Reel (1990 6) Wisconsin mixed hardwood pulpwood stumpage prices. 1972-1990 (Wisconsin Forest Products Price Review). 72 Wisconsin, and in the 1980's in the Upper Peninsula. It is very doubtful that these price increases will be sustained. At this time it is unlikely that another large hardwood mill will locate in the region. Also, cheap eucalyptus fiber from Brazil and elsewhere can easily be substituted for northern hardwood fiber, especially in the production of high-quality printing papers, which are the economic mainstay of most northern mills (Stier 1990). Huch of the Lake States paper industry has an older plant infrastructure and heavy (38 percent) reliance on purchased woodpulp. Therefore, it is quite possible that cheaper imported fiber will have some future impact on Lake States paper producers (Stier 1990), exerting downward pressure on real prices. Despite recent real increases in hardwood pulp prices, the author expects them to stabilize within a few years. Price scenarios for mixed hardwood pulpwood were forecast as: stable (low), +1 percent (medium), and +2 percent (high). summary Northern hardwood sawtimber stumpage prices for the northern Lake States have generally shown real price increases of 0.5-7.5 percent since the early 1970's. Sugar maple, the species of primary interest, has shown price scenarios ranging from stable (Wisconsin 1970-90) to a real increase of over 11 percent (Upper Peninsula, 1980-91). Because of high value and excellent growth rates, price projections for basswood and white ash are also of primary interest. All three species are likely to have real price increases in the next 20 to 30 years, ranging from a low of 1-3 percent for sugar maple sawtimber stumpage to a high of 5-8 percent for white ash. Hardwood pulpwood markets are more uncertain, and may have modest (1 percent) increases or decreases in the future, with a stable real price trend quite possible. CRAPTIR 58 "081383. IARDWOOD “IMAGINE“! 00828 This section presents available cost estimates for various northern hardwood management practices. These estimates are applied to the eight treatments in the Argonne cutting methods study. A review of the scanty northern hardwood cost literature follows. Results of recent studies and the derived cost estimates are then made in the light of the literature. Introduction Land managers need information on the costs of northern hardwood timber management to assure full knowledge of forest management options and to make efficient resource allocation decisions. Landowners and managers need this cost information when choosing between even- and uneven-aged management and the various cutting methods available to achieve their silvicultural goals. Unfortunately, there have been few studies on hardwood management costs in the northern United States. The lack of published information has contributed to uncertainty and imprecise estimates of the costs and returns in northern hardwoods management. Literature Review Hardwood timber management is different from conventional enterprises since it has a very long time horizon. Production costs are affected by land productivity, resource or stand characteristics, available capital, available technology, local, regional, and global markets, and by landowner preferences (Duerr 1960, Davis 1966, Davis and Johnson 1987). Costs of timber production are also affected by social and institutional factors including zoning and land use statutes, taxation policies, forest practice acts, traditions, ownership patterns, and environmental concerns. 73 74 Host empirical work on timber management costs has involved softwoods, particularly in the South (Straka et a1. 1989, DuBois at al. 1991). Recent southern studies suggest several important overall cost trends: 1) Forest management costs in the South (1982-88) increased by less than inflation, as measured by the implicit Gross National Product price deflator: 2) labor costs are the most significant factor associated with recent increases in forest management costs: and, 3) costs associated with regeneration of forest sites account for a very high proportion of a forestry firm's total operation expenditures (DuBois at al. 1991). Characteristics of Hardwood Hanagement Costs The above results are general trends in costs of softwood management in the South. Although some softwood management cost categories are common to hardwood management, there are several differences in costs of management between the two. Usually, hardwood management involves moderate or no site preparation for regeneration, while softwoods often require intensive site preparation. Natural regeneration is common in hardwood management: artificial regeneration in softwoods. Hardwoods often require additional costs for crop tending or quality development. Establishment costs for natural hardwoods are less than for softwoods, while tending costs are higher (Smith 1976, Erdmann and Godman 1981). Host variable costs faced by owners of northern hardwood forests are from labor-intensive practices (Winebar and Gunter 1984). Included in these high labor cost practices are non-commercial tree removal, timber sale layout, timber stand improvement, timber marking, and pre- commercial thinning. These are the costs of primary concern in this study. Limited research on Lake States forest management costs has been accomplished. Researchers in Wisconsin first tried to establish cost 75 relationships for common forest management practices by regression analysis, using time and materials data from forest management agency records. They constructed dollar-cost estimates from the physical cost relationships (Hilliker at al. 1969). Another study (Olson et a1. 1978) developed cost equations for the most common forest management practices, using agency data from across the Lake States. Recent efforts to document costs have concentrated on surveys. In these studies, researchers have asked both public and private forest managers to provide detailed responses on the actual or estimated costs of their operations (Winebar and Gunter 1984, Vasievich and Potter-Witterfl. Hardwood management costs may vary depending on whether the silvicultural system chosen is even-aged or uneven-aged. Hanaging forests under even- or uneven-aged management systems usually leads to different cost functions. Variation in stand structure and ‘ silvicultural goals leads to differential timing of stand entries and, therefore, treatment costs. The uneven-aged methods have the greatest number of stand operations, and thus higher periodic costs. The even- aged methods have fewer operations and thus lower periodic costs, but regeneration treatments are riskier and are often more costly. This is especially true where clearcutting is the harvest method. Annual costs (e.g., taxes, administration) are generally the same for both even- and uneven-aged methods. One recent study (Lewiifi attempted to assess if there were differences in stumpage price received for timber sales using even-aged versus uneven-aged cutting methods. This study would show if there were 'Vasievich, J.H.; Potter-Witter R. Costs of timber management practices in the Lake States, 1987-88. Unpublished draft manuscript (dated 8/16/90) of a cooperative study. East Lansing, HI: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station and Hichigan State University, Department of Forestry: 16-19. (In author's personal file.) ?Lewis, Paul J. Unpublished. The impact of cutting method on northern hardwood stumpage prices. East Lansing, HI: Hichigan State University, Department of Forestry. H.S. thesis. 1990. 155 p. 76 cost differences between even- and uneven-aged methods. The study found that, usually, the stumpage price received varied because of the species being out, not the cost savings of a particular method. There were no significant effects on stumpage value from shelterwood cuts, selection cuts, overstory removals, and improvement cuts. Buongiorno and Young (1984) also found that individual species volume was the most important factor in predicting the high bid for northern hardwood timber sales. Hethods Only direct or variable costs of management were analyzed to compare the economic performance of the eight cutting methods in this study. The advantage of only including direct costs is that the fixed costs of the operation as a whole do not affect relative choice and thus can be excluded in a comparative analysis (Davis 1966, Davis and Johnson 1987). The financial performance of each treatment is then compared with the control or no management treatment. This analysis relied heavily on a few studies of Lake States forest management costs (Olsen et a1. 1978, Winebar and Gunter 1984, and Vasievich and Potter-Witter”. Survey data specific to northern hardwoods management were used to construct variable cost estimates of treatments. Costs not identified as common to northern hardwoods management were not included. Since nearly 90 percent of the commercial northern hardwood forest types in the Lake States are found in Hichigan and Wisconsin (Eyre 1980, Spencer 1983, Spencer et a1. 1988), cost survey responses from Hinnesota were not included. Cost Assumptions Several cost assumptions were made in this analysis. First, costs were discounted at 4 percent, a rate that represents the real, long-term opportunity cost of capital (Row et a1. 1981). All costs are expressed 3Vasievich and Potter-Witter, pp. 16-19. 77 in real 1990 dollars, unless otherwise noted. Land purchase costs are not included in the analysis, because the author assumed that owners purchase land for non-timber reasons (Young st a1. 1985), or evaluate timber management marginally where land is a sunk cost. Taxes are also not included in the analysis. Harvesting costs were assumed to be reflected in the stumpage prices of competitively-bid sales. Results Results of gleaning cost estimates from several surveys and studies are summarized below. Because there were a small number of forest managers who responded to the various surveys, results by management treatment varied greatly. Accordingly, numbers of responses and standard deviation are given for each treatment cost category (TBblO 5-1). We: Though timber stand improvement (TSI) can involve both chemical and non-chemical means, in Lake States hardwoods landowners most often use non-chemical methods. Labor is the most expensive cost component, accounting for 76 percent of all TSI costs in one study (Winebar and Gunter 1984). Wisconsin managers reported average TSI costs of $15 per acre in 1987, while Hichigan forest managers had higher average costs of $22 per acre (Vasievich and Potter-Witter‘). Earlier work in Hichigan (Winebar and Gunter 1984) reported average per acre TSI costs of $36.76. W Felling and killing of unmerchantable trees is an important albeit labor-intensive silvicultural practice in many forest types, both following commercial clearcuts and partial cuts. Its purpose is to improve regeneration and promote stand growth. This practice had ‘Vasievich and Potter-Witter, pp. 16-19. 78 reported average costs in 1987 of $24 per acre in Wisconsin and $16 per acre in Hichigan (Vasievich and Potter-Witter“. Another study (Winebar and Gunter 1984) reported that removal of non-commercial trees after a harvest had labor costs averaging 69 percent of total costs. Total costs averaged $24.51 per acre. TABLE 5-1. Summary of cost survey results: variable costs, number of observations, and standard deviation by category. — L ‘l — 1 Type of variable Costs variable Costs Variable Costs Treatment Wisconsin* Hichigan* Lake States** (per acre) (per acre) Timber Stand $15 (3,510) 522 (7,517) $36.76 Improvement (13,$17.60) Fell/kill 524 (12,518) $16 (9,514) ------- unmerch. stems Cruise/mark $16 (3,54) $23 (5,518) ------- (partial cut Cruise/timber 516 (3,54) 517 (26,514) 515.63 (37,512, mark, imp. cut mark only) Clearcut Layout 584 (6,5156) 516 (25,516) 58.97 (26,511.6) Periodic road $256 (9,5189) 5871 (7,51031) -------- maintenance Reduced current/ $0.50 (8,50.560 $0.47 (2,50.25) -------- special use tax Yield tax rate 7.55 percent 10 percent (7,0 10 percent (8,3.67 percent) percent) L W— . * Vasievich and Potter-Witter (unpublished) ** Winebar and Gunter (1984) 79 Writing These important practices are often lumped together as a single cost category because they occur simultaneously in stand management (Vasievich and Potter-Witterfi. Some studies, however, report them as two separate cost categories (Winebar and Gunter 1984). Because of the similarity in these cost categories for improvement and partial cuts, the following data is reported for both types of cuts. Individual tree selection marking/cruising is perhaps the single most important cost component in uneven-aged northern hardwood management. Harking stands for commercial thinning and partial sawlog cuts occurs every 10-20 years in uneven-aged management, and occurs at least once, about one-half to two-thirds of the way to rotation age, in even-aged management. Because of the differences in timing and frequency, marking is one key cost that differentiates uneven-aged management from even-aged management. Therefore, it is important to pay particular attention to cost estimates for this practice. Winebar and Gunter (1984) reported an average 1982 total cost for selection marking of 512.62 per acre in Wisconsin, slightly less than the average cost of $13.85 per acre reported for Hichigan (37 responses). These figures compare favorably with selection marking costs reported in another study (Vasievich and Potter-Witterfi. That work showed an average cost for cruising and selection marking of 516 per acre in Wisconsin, with a range of $12 to $20. The cruising portion of this estimate is about $2.75 per acre, leaving $13.25 as an average for marking. In real terms (net of inflation during 1982-87) the two cost estimates are virtually identical. There is a significant degree of variability between selection marking costs as reported by various agencies and types of ownerships. Costs of selection marking in one of the studies ranged from a low of ’Vasievich and Potter-Witter, pp. 16-19. 80 $9.31 reported by industry sources to a high of $24.26 per acre reported by Federal agencies. In all ownerships labor costs were about three- quarters of total selection marking costs (Winebar and Gunter 1984). A few researchers have cited the potential for diameter-limit cutting in northern hardwood stands because of the reduced marking costs (Lyon 1986). Based on work with forest industry, these workers believe that marking costs could be avoided altogether when corporate foresters and logging contractors work closely together. At least two large corporate landowners in Hichigan have tried this practice, but its long- term feasibility has yet to be proven, particularly on non-industrial lands. Thinnina_£2111 Olson at al. (1978) derived equations for manual thinning based upon thirty-eight operations in Hinnesota and Hichigan state forests. With an average basal area of 38 square feet removed (BAR), their estimated total project cost equation was: r - 120.29 + (0.418(8AR)] so For a 40-acre tract of northern hardwoods this would translate into a cost of 513-14 per acre (1978 5), or about 519-21 (1990 5). The cost compares well with the $18-19 per acre cruising/marking cost reported by Vasievich and Potter-Witter‘. Cost of lighter or heavier thinning would vary accordingly. Since the Argonne cutting methods averaged 25-30 square feet of basal area removed, estimated thinning costs would be somewhat less in this study than in Olson et a1. (1978). Thiede (1986) reported that northern hardwood thinning costs on .Hichigan state forests had been about $30 per acre in 1974. The development of commercial firewood markets in the mid- to late-1970's ‘Vasievich and Potter-Witter, pp. 16-19. 81 caused thinning costs to drop to about $16 per acre in 1978. By 1984, nearly all hardwood thinning on state lands with access to good firewood markets was being done profitably, through commercial timber sales. Here, a net cost had become a net return. £1!!!£H&_LIIQES Reported costs for laying out clearcut timber sales averaged $5.15 per acre for Wisconsin and $8.14 for Hichigan in the early 1980's (Winebar and Gunter 1984). Laying out clearcuts ranged from a low of $3.15 per acre for state and industry to a high of 519.77 per acre for the federal agencies, where the use of landscape architects and wildlife biologists, and higher overhead, add appreciably to sale layout costs (Blumfl. By 1988, the latter survey reported cruise and sale layout costs averaging 516 per acre for Hichigan and $84 per acre for Wisconsin (Vasievich and Potter-Witter‘). These figures include clearcut sale layout costs and 52-3 per acre for cruising costs. These costs are significantly higher than costs for the same practice reported in the earlier survey. Because clearcutting is becoming more controversial, and a variety of expertise is now commonly used before clearcuts take place, this study will use an estimate intermediate to the extremes listed above. A cost of 526 per acre (the mean for all respondents in the most recent Lake States survey) is a reasonable compromise, reflecting the increased labor and environmental compliance costs for clearcut management. ’Blum, John. Assistant Ranger, Timber Hanagement. Baldwin District, Hanistee National Forest. Personal communication. 'Vasievich and Potter-Witter, pp. 16-19. 82 Cost variation Between Even- and Uneven-aged Hethods In this analysis, three methods (BA 60, 75, 90) are clearly uneven-aged. Two of the methods (clearcutting and shelterwood) are even-aged. The crop tree cutting method, though classified by Stoeckler’ as uneven-aged, is closer to an even-aged or a two-aged method because of the stand structure it creates (Erdmann 1986). Similarly, the diameter-limit method was called an uneven-aged method (Stoeckler’), but results in stands that are structurally quite similar to the commercial clearcut (Erdmann”). Hardwood management costs occur each time there is a stand entry: thus, costs are more frequent and usually higher for uneven-aged management. Application of Cost Estimates to Argonne Cutting Hethods Cost estimates from above were applied to the Argonne cutting treatments. All costs are variable costs, and were expressed in 1990 dollars, then adjusted for inflation by the Producer Price Index for all commodities (1990 8 100). Finally, all costs were discounted with a 4 percent discount rate. Wham This cost is consistent with Olson st a1.'s (1978) analysis. Each basal area treatment was selectively marked every ten years. Average cruising and marking costs from the three studies was $17.60 (1987 5) per acre for both improvement and partial cut marking; this is about $20 per acre in 1990 dollars. Therefore, a cruising and marking cost of $20 per acre was assigned to each basal area treatment. ’Stoeckler, J.H. Establishment report for cutting methods study (A-3) in pulpwood sized northern hardwoods. Unpublished manuscript (dated January 1955). St. Paul, HN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station, Argonne Experimental Forest. 60 p. I‘Erdmann, G.G. July 17, 1992. Personal communication. 83 W This study included two types of clearcut treatments: a silvicultural clearcut with a l-in. residual limit, and a commercial clearcut with a s-in. limit. Eoth treatments were out only in 1952. Though reported costs for clearcut layout in Visconsin averaged $84 per acre in 1987, this cost category had only five responses. An average of 57 responses from the three surveys indicates that a cost of $26 per acre (1990 $) is a figure reflecting current operational costs. Post-harvest of non-commercial trees is a cost that needs to be assigned to the silvicultural clearcut treatment, where l-S in. non- merchantable stems were removed. In 1990 dollars, this would be about $23 per acre (Vasievich and Potter-Witter"). MW The Argonne cutting methods study is adjacent to three unpaved USDA Forest Service roads. The study replications are no more than one- half mile from these dirt access roads, and approximately one to two miles from Wisconsin Route 32, a paved, all-weather secondary state highway (Figure 3-2, page 32). Typical logging access conditions in the Lake States northern hardwoods region vary widely; there is really no such thing as a “typical“ logging chance in the Lake States. Some areas, such as the uplands of the northern Lower Peninsula of Michigan, are crisscrossed with logging roads that make logging access low in cost (if trespassing is not a problem and rights-of—way can be obtained). Other areas, such as the central and western parts of the Upper Peninsula, have more difficult physical access that occasionally require access roads of two to five miles in length. Spencer (1983), reported that 39 percent of commercial forestland in Hichigan is within one-quarter mile of a "Vasievich and Potter-Witter, pp. 16-19. 84 maintained road, and SS percent is within one mile. Northern Wisconsin has average logging access conditions that tend to be intermediate between these two extremes. Documentation of differences in road costs for even-aged and uneven-aged hardwood management is poor. Initially, road costs for even-aged systems tend to be less per unit of wood harvested than for uneven-aged systems, yet limited research indicates that in the long run the type of harvest system has little influence on road costs for similar site conditions (Smith and DeBald 1978). Given this poor information, and the marginal cost framework of this analysis, differences in road costs between the two systems are assumed to be minimal. Therefore, road costs are not included in this analysis, with one exception discussed below. Read maintenance, a variable cost, is an important cost category, particularly for uneven-aged management, where stand entries occur every 10-15 years. Vasievich and Potter-Witter's'2 Lake States study found average annual Wisconsin road maintenance costs ranging from $110 to $330 per mile, with 5190-200 the average for most conditions. Variation in these costs may be a function of different state, agency, and corporate standards (Winebar and Gunter 1984). Based upon common practice and ”average“ conditions, a maintenance cost of $200 per mile was included every ten years (stand entry interval), for the uneven-aged methods. This cost averages about $5.00 per acre, with a $3.00-7.00 range, assuming an SO-acre parcel size (Table 5—2). Because even-aged management requires only two or three entries per rotation, road maintenance costs in these treatments were applied only for final harvest and thinnings. “Vasievich and Potter-Witter, pp. 16-19. 85 TABLE 5-2. Average nominal road maintenance costs used in the analysis. (These costs are on a per-acre basis, assuming an SO-acre parcel sise.) Year Low ledium Eigh Costs Costs Costs 2022 $7.25 $12.10 $17.00 2012 5.40 9.00 12.65 2002 4.05 6.70 9.40 1992 3.00 5.00 7.00 1982 2.55 4.30 6.00 1972 1.00 1.70 2.40 1962 0.80 1.35 1.90 1952 0.75 1.20 1.65 Sensitivity Analysis: Selection Harking, Sale Layout, and TSI Costs Results of the 1952-2022 analysis were tested for sensitivity to changes in individual tree selection marking, TSI, and sale layout costs. Table 5-3 lists these variations in cost. Opportunity Costs Opportunity costs measure the cost of making an investment in terms of lost opportunities or alternative courses of action. In forestry, opportunity cost is an important concept because cutting decisions, once made, affect future forest conditions for decades. The ”wrong“ decision can lead to high opportunity costs for forest managers and forest users. Once a stand is high-graded, marked too heavily, out below a specified diameter-limit, or fails to regenerate, the landowner pays a substantial opportunity cost for many years to come. Those landowners with large acreage (industry, state, and national) are less adversely affected by these decisions than smaller, non-industrial private landowners who have fewer acres to manage and fewer alternatives to choose. One problem in dealing with opportunity costs is that they 86 are not easily anticipated and cannot be quantified readily. Uncertainty about future timber markets and forest conditions complicates landowners' ability to anticipate Opportunity costs. TABLE 5-3. Data used to test sensitivity of treatment results to changes in marking, TSI, and sale layout costs; per acre (1990 S). _—_ _ Cost Category Low Wedium Bigh Interval Cost Cost Cost Cruise/Hark Every for Partial 10 years $16 $20 $24 Cut 1952-2022 I Shelterwood 1957, 1965, 16 20 24 2022 Clearcut Sale 1952, 2002, 12 26 40 Layout 2022 Post-Harvest Removal Non- 1952, 2022 20 23 26 Comm. Stems (Silv. CC) Periodic Woods 10 years or 3 5 7 Road before cuts ($120/mi) ($200/mi) ($280/mi) Maintenance - Opportunity cost is also important when comparing even- and uneven-aged management because timber growing stock functions differently in the two systems. In even-aged management, the growing stock is merely the crop to be harvested, while in uneven-aged management, the reserve growing stock is both the crop and the "factory” that produces the crop. In either case, the reserve growing stock represents accumulated invested capital. A landowner's marginal cost of capital decides the level of growing stock allowed to accumulate. In economic theory, if a landowner cannot achieve a rate of return on growing stock that is above his or her alternative rate, then s/he will harvest the excess growing stock (Davis 1966). 87 Social Costs of Timber Wanagement Because the great majority of commercial hardwood resources are in the Eastern United States where population pressures are greatest, people demand that hardwood forests provide a greater variety of resource benefits than other forest types. Thus, social concerns can have a greater impact on costs in hardwood management than in softwood management, particularly in the Worth, where hardwood resources are often close to large population centers. Costs associated with environmental regulation of timber management, though important, are poorly documented. A few forest management practices, such as clearcutting, have been restricted or eliminated entirely in certain areas. The federal government, for example, recently enacted a policy that restricts clearcuts larger than 40 acres, and eliminates clearcutting as the standard silvicultural technique on most national forests (Robertsonu). Legislative restrictions on cutting practices impose higher costs, yet few regions have consistent evidence of the extent of these costs. Many states have enacted forest practices acts that add to timber management costs, particularly for non-industrial private forest owners whose transaction and administrative costs for complying with forest practice acts are much higher than state government (Ellefson 1992). At this writing a forest practice act has been proposed for Michigan, but has not been passed. Neither Wisconsin nor Hinnesota has a forest practices act yet, nor anticipates enacting one in the near future. Evaluating the impact of higher social costs in hardwood timber management is next to impossible. However, sensitivity analysis can be used to allow for much of this uncertainty. I’Robertson, F.D. “Ecosystem Hanagement of The National Forests and Grasslands.“ Letter from Dale Robertson, Chief, USDA Forest Service, to Regional Foresters and Station Directors. Dated June 4, 1992. 3 p. plus enclosures. Suuary Hardwood management costs from four Lake States studies were reviewed and compared. Costs from the two more recent studies (Winebar and Gunter 1984, Vasievich and Potter-Witter“) were used in this analysis. Variable costs included in this cutting methods study were: sale layout, timber stand improvement, individual tree selection marking, and periodic road maintenance. These variable costs were used as inputs to the marginal economic analyses reported in Chapters 7-9. These highly variable and site specific hardwood management costs provide a basis for the sensitivity analysis in this study. However, since there are very limited empirical data from northern studies, future northern hardwood cost research is crucial, and may change the results of this study. “Vasievich and Potter-Witter, pp. 16-19. CHAPTER ‘8 GROWTH AND YIILD P38010110"! Introduction Forest managers have a variety of programs available to project northern hardwoods growth and yield in the Lake States. This study used a 30-year growth and yield projection (1992-2022) to supplement the original 40 years (1952-1992) of data on cutting methods. The growth and yield projection was needed for a more complete description of stand condition and response to treatment. Instead of only 40 years, a management horizon close to northern hardwood rotation age was desired. Comparing even- and uneven-aged management in northern hardwood forests is complex because cutting cycles, stand entries, and stand volumes vary between the two systems. In this study, the stand that originated in 1905 was even-aged with most of the oldest trees about 90 years old. Even though there were 40 years of data (1952-92) on nine cutting methods, the data base was inadequate to compare even-aged with uneven-aged management. Rotation ages for even-aged management of Lake States northern hardwoods are usually 110-120 years for sawtimber and 50-70 years for pulpwood (Erdmann 1986). Cutting cycles for uneven-aged management in northern hardwoods range from 10 to 20 years (Erdmann 1986, Tubbs 1977b, Arbogast 1957). The available data were only sufficient to observe the silvicultural effects of four or five periodic cuts in the uneven-aged treatments and about one-half of a sawtimber rotation in the even-aged treatments. Growth and yield models, developed since the early 1970's, can help close this knowledge gap. A variety of models exist for projecting growth and yield of mixed hardwood stands (Belcher 1982, Belcher et a1. 1983, Brand 1981, Adams and Ek 1974, Hoser and Hall 1969, Hoser et a1. 1979). Available growth and yield models can be divided into two major categories: stand level and individual tree models. Stand level models are used where individual tree data are not generally available. 89 90 Individual tree models are further classified into three types: individual tree, distance-independent: individual tree, distance- dependent: and deterministic models. Because this study's data are specific to individual trees on plots, the individual tree, distance- independent models are of greatest interest. Wethods Lake States TWIGS (The Woodsman's Ideal Growth Projection System) (Belcher 1982) was used in this study. Experts in hardwood growth and yield (Hiner et a1. 1988, Belcher 1982) suggest that for short-term projections of up to 30 years where individual tree data is available, the STEHS or TWIGS growth and yield model is very appropriate. TWIGS is an individual tree, distance-independent growth and yield simulation program developed by the USDA Forest Service's Worth Central Forest Experiment Station. TWIGS projects the diameter-at-breast-height growth and the mortality of individual trees in the context of a stand, whether of single or mixed species and sizes. Thus the computer program grows the forest stand through time based on the characteristics of the stand (Hiner et a1. 1988, Kowalski and Gertner 1989). TWIGS is actually the microcomputer version of a main-frame growth and yield program developed earlier, called STEHS (the Stand Tree Evaluation and Modeling System, Belcher 1981, Belcher et a1. 1983, Holdaway and Brand 1986). One of the specific purposes of TWIGS is to evaluate the productivity and economic effects of different silvicultural prescriptions, a purpose consistent with this research. Characteristics of TWIGS TWIGS and its predecessor, STEHS, a mainframe computer model, are both regional growth projection systems. Data used to develop the Lake States version of TWIGS came from Hichigan, Minnesota, and Wisconsin (Hiner et a1. 1988). Nearly all of the data for northern hardwoods came 91 from Wisconsin and Hichigan's Upper Peninsula (BrandH. Because the TWIGS data came from across the region, it results in average species growth and mortality for the Lake States region; it will not necessarily give accurate localised estimates. When localised estimates are required, the developers of TWIGS recommend the use of methods for improving estimates with additional diameter-growth information (Smith 1983, Holdaway 1985). The TWIGS growth function equation has three subcomponents: potential diameter growth, a competition modifier, and a diameter adjustment factor. These three components of growth are related as follows: annual diameter growth equals potential diameter growth times the competition modifier plus the diameter adjustment factor. Projections from each diameter growth model are used in each succeeding projection. Each component is modelled in TWIGS based upon a set of equations developed for that purpose. In Lake States TWIGS, estimated board-foot, cord, and cubic feet volumes are based on Upper Peninsula of Hichigan equations developed for the program (Raile and Smith 1982). Volume equations were developed from measurements of d.b.h., merchantable height, top diameter outside bark, site index, and tree class. There are also three condition classes that trees can be assigned in TWIGS. They are: l) acceptable, 2) undesirable, and 3) cull. In this study, all trees were given acceptable condition classes except for ironwood, which was classed as undesirable, and obvious cull trees. By 1992, after 40 years of management, virtually all cull trees were gone from the treatment study plots. Therefore, the 1992-2022 projections were done with all trees classed as acceptable. Assuming reasonable accuracy of the projections, this combination of 40 years of actual data with 30 years of projection creates an expanded view of stand development. It gives the study about three- fourths of a typical sawtimber rotation, and 20 years longer than a 'Brand, Gary. November 5, 1992. Personal Communication. 92 fairly typical pulpwood rotation. With this longer analysis period, based on actual and projected growth, even- and uneven-aged methods can be evaluated on a more comparable basis. Validation of Diameter-growth Predictions Diameter-growth predictions of TWIGS were validated by projecting sample plots from all cutting treatments for the 1951-81 growth period. Actual plot growth was compared with the performance of TWIGS over 30 years. A total of 17 sample plots were tested from the three replications; most were from replications one and two, which had more white ash and basswood. Average annual prediction errors for sugar maple were -0.015, -0.0008, and -0.012 for the 10, 20, and 30-year growth periods tested, respectively. This level of prediction error was well within the plus or minus 0.020 acceptable prediction error calculated by Holdaway (1985). Therefore, no adjustment of growth coefficients was made for sugar maple. However, basswood and white ash exhibited much higher prediction errors than sugar maple, particularly for the 9, 11, and 13 in. diameter classes. Annual prediction errors for white ash ranged from a low of -.025 in. per year to a high of -0.093 in. per year, depending on diameter class (Table 6-1). Diameter growth for sawlog-sized white ash had an average underprediction of nearly two inches within a 20-year period. Growth and yield for basswood was also strongly underpredicted for the Argonne sites. Twenty-year prediction errors for basswood ranged from a low of -.027 in. per year to a high of -.074 in. per year (Table 6-1). 93 TABLE 6-1.- Comparison of growth predictions for three species before adjustment (in inches, measured at d.b.h.). 20- ear annual rediction error Potential underprediction Species Low High Ten years Twenty years Sugar maple -.006 -.015 -0.15 -0.30 — Possible Reasons for Underpredictions Holdaway (1985) suggested the causes of poor predictions of tree diameter growth in TWIGS (or STBHS) growth models could include several factors: limitations in the original data base; the preponderance of experimental plots, which are usually carefully managed; changes in climate and environment; and the fact that 90 percent of the hardwood trees used to calibrate the model were between 5-11 inches. Another factor Holdaway (1985) cited was the aging of the model development data. Researchers took most initial measurements between 1945 and 1960, and final remeasurements in 1975 and 1976. Changes in general forest conditions since 1976, such as droughts, pest outbreaks, and warming trends, could lead to over or underprediction of d.b.h. growth for several species. For example, in some forests high mortality in American elm could have led to other hardwood species having greater- than-expected growth during a 20-30 year period. Another factor is the limited amount of data for white ash and basswood used to calibrate and adjust the model. Finally, it is likely that site indices for basswood and white ash were higher at the Argonne (8.1. 70-72) than for the region-wide average originally used to calibrate TWIGS. Although basswood and white ash together made up only 15 percent of the study's trees in 1951, both species had high growth rates, good quality, and higher-than-average prices. For these reasons, they contributed disproportionately to stand value. Therefore, an adjustment 94 or correction to their growth rates was necessary before the growth projection could be completed. Diameter-growth Adjustment Hethodology When making adjustments to growth and yield programs, there are two methods: additive and multiplicative. Holdaway's (1985) additive method was chosen here for several reasons. First, it provides more realistic projections of overall stand characteristics than the multiplicative adjustment model. Second, it has the flexibility to increase or alter growth for large diameter trees, a trait lacking in the multiplicative adjustment models. Finally, it has a proven record of reducing d.b.h. prediction errors specifically for Wisconsin forests (Holdaway 1985). Prediction errors for basswood and white ash were calculated for a 20-year period by comparing actual field data with individual tree growth predicted by TWIGS in each treatment plot for these periods. The 20—year prediction errors were used because this was the best compromise between having a short growth interval for accuracy and a longer period to maximise sample points (Holdawayfi. When lO-year diameter correction coefficients were used initially, 10-year corrections obtained were good for both species, but 20 and 30-year growth, especially for white ash, were overpredicted. Prediction errors were stratified across treatments to determine if there were differences in errors due to treatment effects. An analysis of variance was run on all the 20-year growth prediction errors in HIWITABO. Though there were some differences due to treatment, they were not found to be significant at the a - .05 level. Although one might expect treatment differences on an individual tree basis, this was not the case, perhaps because basswood and white ash trees are both somewhat tolerant of shade, and if they survive heavy early competition, aHoldaway, Hargaret. October 5, 1992. Personal communication. 95 can achieve good growth rates regardless of cutting history. This effect was seen in several of the control plots, where white ash in particular exhibited excellent growth. Because no significant treatment effects were found, two prediction equations were fitted to the data, one for basswood and one for white ash. For each tree in each species, the midpoint d.b.h. was calculated by: (final observed d.b.h. - initial d.b.h.) / 2. Each observation was placed in the appropriate 2-in. diameter class; e.g., 5, 7, 9, 11, 13, and 15 in. classes. The annual growth error was found by: (predicted - observed growth) divided by the number of years (10, 20, 30) in the growth period. The average adjustment for each diameter class was thus the average growth error, or: §2m_i29§i_:_2£29i1 Eq.1: Avg. adj. 8 s no. obs. in d.b.h. class The average 20-year adjustment was calculated for each diameter class for basswood and white ash. These "new“ adjustments were added to the 'old' adjustments calculated by Holdaway (1985) to yield a “final” adjustment for each diameter class. Average calculated adjustments for basswood and white ash, by d.b.h. classes, are shown below (Table 6-2). Once the above average adjustment was calculated for each diameter class, a polynomial function of the form: «.2: Adj. - a + b(d.b.h.) + c(d.b.h.)’ was fitted to the above data using HINITABO. The fitted value was calculated at each 2-in. d.b.h. class, then added to the old adjustment value. Finally, another polynomial function of the same form as above was fitted to this data to yield the new adjustment coefficients. These coefficients were entered into the TWIGS coefficient file (LS-TWIGS.dat) 96 in place of the previous values for basswood and white ash, following instructions given by one of the designers of TWIGS (Brandfl. TABLE 6-2. Basswood and white ash coefficient adjustments by d.b.h. class. Basswood: fl Old adjustment New adjustment Final d.b.h class Holdaway 1985 Twenty-year data adjustment 5 -.035 .003 -.032 l 7 -.034 .042 .008 9 -.031 .066 .035 11 -.029 .056 .027 13 -.025 .074 .049 White ash: Old adjustment New adjustment Final d.b.h. class Holdaway 1985 Twenty-year data adjustment 5 -.042 .025 -.017 7 -.049 .055 .006 9 -.050 .070 .020 ‘_ - The final fitted equations for basswood and white ash were as follows: 3.1.3: W... Eq.4: - 0.0312 d.b.h.-0.00123(d.b.h.)’ -0.154 mm - 0.0007 d.b.h.-0.00036(d.b.h.)’ -0.24. ’Brand, Gary. November 5, 1992. Personal Communication. 97 The above prediction equations were highly significant at the a -.05 level. The F-value for the fitted basswood equation was 20.38 with an R, of 93.1 percent (adjusted R, . 88.6 percent). The F-value for the final fitted white ash equation was 17.92 with an R, of 92.3 percent (adjusted R, I 87.1 percent). For comparative purposes, the growth adjustment coefficients from these equations are displayed with Holdaway's (1985) basswood and white ash adjustment coefficients to the original STEHS (Belcher 1981) model (Table 6-3). TABLE 6-3. Comparison of TWIGS diameter adjustment coefficients before (Holdaway 1985) and after final adjustments. - Coefficientle lSpecies Term Before After I Basswood a, 0.00000 0.03120 I a, 0.00007 -0.00123 a, -0.03700 -0.15400 White ash a, -0.01130 0.00070 a2 1.00066 -0.00036 -0.00200 -0.02400 Once TWIGS was modified with the new coefficients, the modified program was run to determine the level of prediction improvement for basswood and white ash. The 20-year white ash prediction error, that before adjustment had averaged -0.083 annually, now averaged only -0.01137 annually, based upon 22 observations from nine different sample plots. The 20-year basswood prediction error, that had averaged -0.059 before correction, was now a slight overprediction of 0.0176 annually (23 observations). Another result of the change in coefficients was a slightly improved predictability for sugar maple, even though its prediction errors were already within acceptable limits for the program 98 before the changes. The overall result of the changed coefficients was a substantially improved version of TWIGS for the prediction of white ash and basswood growth on the Argonne study site. Stand Growth Parameters for Eunning TWIGS A number of stand growth parameters must be specified before TWIGS can be run. Due to the 5-year (growing season) remeasurement period throughout this study, a 5-year projection interval was chosen. This is the frequency at which TWIGS applies the growth and mortality equations. The TWIGS technical guide (Niner et a1. 1988) also suggests a projection interval of not over five years to maintain accuracy. A related parameter, years in growth cycle, was also set for a 5-year interval. This parameter determines the frequency of stand and stocking table generation, and allows for the management Option. In the uneven-aged treatments, management only occurs on a 10-year interval, but the 5-year growth interval allows for checking remeasurement data on a basis consistent with the study's original design. For each 0.1 acre sample plot included in the study, trees present in 1992 were entered into a companion computer program called TRBBGEN (Brand 1981). A total of 27 plots were projected; nine treatments were projected for each of three replications in the study. Host of the treatments in each replication had five 0.1 acre sample plots, and these were aggregated for each tree list. Consequently, an expansion factor of two was entered to put all treatment projections on a per acre basis for each of the following treatments: the light, medium, and heavy selection, crop tree, control, diameter-limit, and shelterwood. For the commercial and silvicultural clearcut treatments, only two 0.1 acre sample plots were available in each 5-acre clearcut block. Therefore, an expansion factor of five was used to put these treatments on a per acre basis. To give even-aged treatments about three-fourths of a sawtimber rotation, six 5-year growth cycles were chosen over 30 years 99 (1992-2022), with cuts occurring every 10 years in the uneven-aged treatments. In the prior TWIGS validation, six S-year cycles were also chosen over 30 years (1951-81). Other relevant parameters chosen before running TWIGS were as follows: management option on; economic evaluation option Off; volume option chosen was Scribner Decimal C rule; and five diameter classes were chosen. The economic evaluation Option was kept Off in this study because all growth and yield results were later evaluated with the use of Quick-Silver v. 2.0 (Vasievich and Frebis 1984), a forest investment program. The Scribner Decimal C volume rule was used for consistency with study measurements and northern Lake States' protocol. Also, five diameter classes were chosen for all TWIGS simulations, to be consistent with the study's diameter classes and to give greater detail Of projections. In the validation all management activities were done on an actual, individual tree basis. If a 7.5 in. d.b.h. red maple was cut, it was also 'cut' in TWIGS. For the 1992-2022 stand projections, uneven-aged treatments were managed through a thinning from above to a specified residual basal area (BA 60, 75, 90). This TWIGS treatment option simulates the harvest of all volume growth in the form Of large sawlogs, an acceptable treatment once the selection stand is fully regulated. The crop tree treatment was cut by individual tree selection, favoring existing crop trees and leaving a residual basal area Of 85 square feet per acre after each cut. This treatment differed slightly from the original crop tree treatment imposed in 1951. The original crop tree treatment had thinned trees tO BA 60-70, following New England guidelines, which later proved to be too heavy. Because it was cut in 1992, no out was modelled in the TWIGS projections for the 8-in. diameter-limit treatment. The growth of the control treatment by definition was modelled with no changes. Due to 100 the small sise Of the residual stands, the commercial and silvicultural clearcuts only received treatments in 1952. Similarly, the shelterwood was cut only in 1957, 1965, and 1975. Of these three treatments, only the commercial clearcut had a stand nearly ready for commercial treatment in 1992. A regeneration cut to a 5-in. residual diameter was simulated for the commercial clearcut in 2002, while a similar cut was simulated for the silvicultural clearcut in 2022, the end Of the projection period. The shelterwood plots were not cut in 2022, because they had not reached commercial sise. Results after Forty Years of Nanagement Over the 40-year management period, the diameter-limit treatment had the greatest total volume Of sawtimber (8069 board feet per acre) removed (Table 6-4). The shelterwood treatment ranked second in total sawtimber removals (7664 feet per acre) while the heavy selection treatment ranked third at 7101 board feet removed per acre. The medium selection treatment ranked fourth at 6775 board feet, the light selection next at 5475 board feet, and the crop tree treatment came next at 5119 board feet removed. The silvicultural clearcut had 3510 board feet removed (all in 1952), while the commercial clearcut ranked last (in 1992) at only 3325 board feet harvested (Table 6-4). Total pulpwood removals were greatest over the 40-year period in the diameter-limit treatment at 25.4 cords per acre removed. The medium selection treatment ranked a close second at 25.1 cords. The heavy and light selection treatments ranked third and fourth at 21.5 and 20.9 cords, respectively. Crop tree pulpwood removals were 12.5 cords, and shelterwood removals were 10.6 cords. The silvicultural clearcut to a 1 in. limit had 8.2 cords removed through 1992, while the commercial clearcut ranked last with 6.7 cords removed (Table 6-5). 101 .maco ome>asm + .emOstm cw muaaeuuoe onus won an 0mm cosmos“ 00: noon cc .>Ho>«uummu0u .mhma use meme .mmmu on one: nuuo>wsc ooozuouaocm c Oouosnonm o homma mauve «mmofl emcee veep Fonda nomad meow a Heaven AmuOH o amaa ona amaa mena 0 meme 0 o aaoa o vwha mesa mama amma o o o o aHoa o eflmm emma mnha owed o o mwam o aooa OceannOhm o mbvm memo aoab mHHm veer oflmn mamn moon deuce 0 meme cwflmoa HMMH o o o o amev amma o mmaa mmoa omoa come 0 o o o amma 0 mac haoH new omaa canoe o o o ahoa o mmwfl coma Hmoa +va some 0 o o awma o mum voaa mama mmvH cmmma onm mann been amen Howunoo .Hom .Hom .Aem ooue COOS 2H gm :0 uno>hmn unwed suave: m>som mono lusuaonm van 0:0 valid as use» usoao usouu .san .swosxasem mason wenndwum .wssh use uselesewu ha .nonedh wenl408sm .vle “Adda 102 uouuofioum o o.oo h.a¢ o.mm H.an o.oH m.mn H.mm m.oa a Amount annoy o o.m o.m m.v m.m o n.0n o o aaoa o o.o o.o m.m m.o o o o o aHoa o m.h o.o H.o m.n o o ¢.an o aooa unevenoum o m.oa H.ma m.Ha m.aH 0.0H a.m h.w o.ma Hmsuot Annoy o H.m o.m a.m o o o o a.Ha amma o a.n a.m h.m v.a o o o 0 Need 0 H.m m.m m.m w.o m.n o o o ahma o H.a a.a o.n H.H a.a o o 0 Noon 0 ¢.H a.a H.H a.a w.¢ a.m h.o a.o ammfl HOhudOU .Hum .Hom .Aom oouu 6003 3n 8:0 gm #50 so unabumn nomad snaps: m>som mono luouaoom useau ussau eased no use» .sao .suos new svwoo :4 .ussh one uneluesuu he ended» nocthusm .nlo manta 103 Results of Nanagement including TWIGS Projections Total treatment harvests and rankings changed when the results Of 30 years (1992-2022) of projections were added to the 40 years of actual management (1952-1992). The medium selection treatment now had the highest sawtimber production at 14,708 board feet per acre, followed closely by the heavy selection (14,594), the light selection (13,847), and the crop tree (13,100) treatments (Table 6-4). The commercial clearcut was highest Of the even-aged treatments and fifth overall at 11,593 board feet per acre, followed by the silvicultural clearcut with a production of 11,397 board feet per acre. The diameter-limit cut ranked next at 8069 board feet per acre, and the shelterwood treatment was last with 7664 board feet per acre. Rankings for total production Of pulpwood also changed. Including TWIGS projections caused the commercial clearcut tO move up to third for total pulpwood production (39.1 cords per acre) and the silvicultural clearcut to fourth (38.9 cords per acre). The medium selection treatment now had the highest pulpwood production at 42.7 cords, with the light selection now second (40.0 cords). The heavy selection was fifth (38.0 cords), the crop tree sixth (32.1 cords), and the diameter-limit had fallen to seventh (26.3 cords). The shelterwood treatment remained last for total pulpwood production (Table 6-5). Several interesting management trends emerge from the projections. First, the commercial clearcut was ready for harvesting again in the year 2002, and yielded a great deal Of low-grade (mostly sugar maple) sawtimber in that cut--an average Of 8268 board feet per acre across the three replications. The projected cut ranged from 5936 (replication 1) tO 9512 (replication 2) board feet per acre. Pulpwood removals in 2002 totalled 32.4 cords per acre. Average tree size in the cut was 8.8 in., with a range Of 5.5 to 21.7 in. d.b.h. Species composition Of the sawtimber cut was overwhelmingly sugar maple (82 percent Of cut BA), with much smaller amounts Of basswood (7 percent Of BA), white ash (6 104 percent Of BA), red maple (3 percent of BA), and bigtooth aspen (2 percent Of BA) harvested. By contrast, the simulated silvicultural clearcut in 2022 had most (> 65%) Of its sawtimber volume in valuable white ash. Another intriguing development shown by the TWIGS projections is the sustainability Of all three selection cuts. Each cut harvested only the larger diameter trees after 1992, yet the residual stand diameter distributions were sustained. Only the heavy selection cuts in Replications 2 and 3 had diameters Of harvested trees that decreased tO below 20 in. when thinned. In actual management practice, however, although large trees will continue tO be harvested on the basis Of financial maturity, some smaller trees will also be selected for reasons of high risk, etc. Therefore, actual volumes harvested in the next 30 years may be less than shown here for the projected selection cuts. Summary Lake States TWIGS was tested for accuracy in predicting the growth and yield Of the Argonne study treatments for 1951-81. Host species were predicted reasonably well, but there were substantial and consistent underpredictions for white ash and basswood. Using Holdaway's (1985) additive adjustment methodology, diameter coefficient adjustments were made for these two species, and an adjusted TWIGS model created. The adjusted model performed well for the 30-year (1992-2022) period needed in this study. Growth and yield results for treatments were consistent with past growth Of similar-sized stands Obtained in actual silvicultural trials. Therefore, the modified TWIGS results were used in the extended economic analysis in this paper. CHAPTER 78 “1818 0? TRADE-0P1! BETWEEN TREE IPECIES DIVERSITY IND ECONOMIC RETURN. Forest ecologists have long believed that greater tree species diversity in northern hardwood forests reduces biological risk, but researchers have not yet linked diversification with economic returns for managed hardwood forests. This chapter shows how management Of northern hardwood forests affects tree species diversity and economic returns. Shannon's index (Drawer and far 1984) was used to measure overstory and regeneration diversity for the nine even- and uneven-aged cutting methods from this 40-year study on the Argonne Experimental Forest. These indices Of tree species diversity were then compared with the potential economic returns for each cutting method. Introduction There has recently been a rapid development Of interest in ecosystem management among forest managers (Salwasser 1990). Haintaining biological diversity in forested ecosystems has become an important resource management issue. Bcologists consider biological diversity at four different levels: genetic, species, ecosystem, and landscape (HcNeely et a1. 1990, Cleland and Scott 1990). Yet economists have not studied the implications Of preserving diversity until recently. The paucity Of economic studies results from difficulty in defining and measuring biological diversity on an economic scale. Lacking clear definition, resource allocations for biodiversity will be less than Optimal (Leefers 1990, Randall 1988). This analysis defines and measures species diversity as a parameter Of average rarity within a forest stand (Patil and Taillie 1982). In a diverse community, average rarity is the probability that a particular species will be comparatively rare, whereas in a single 105 106 species community such as a red pine plantation the average rarity or diversity measure is expected tO be near serO. As a portion Of the overall economic analysis, this chapter addresses tree species diversity in a northern Wisconsin hardwood stand. The Objective is to learn which cutting methods are profitable, yet also maintain or enhance tree species diversity. Both even- and uneven-aged management can be used to regenerate northern hardwood stands (Tubbs 1987, Godman 1985). Conventional wisdom holds that even-aged management produces more diverse stands than uneven-aged management. However, either system can increase or reduce species diversity under varying site conditions (Stearns 1986). Past research on the various cutting methods used in northern hardwoods management has shown that tree diversity is strongly affected by the choice and timing of a specific cutting method (Chapter 2, pp. 17-19). Initial tree species composition also affects potential stand diversity. Species composition of the stands in this study in 1951 was 63 percent sugar maple and 4-9 percent each Of basswood, yellow birch, white ash, red maple and hemlock (Table 3-1, p. 32). Diversity Neasurement The Shannon diversity index is derived from information theory. Its premise is that natural system diversity can be measured similarly to the information contained in a code or message (Hagurran 1988). The Shannon measure of rarity of species, H', is calculated by: 3.1.1: H'I-Zpflogmpi where p" -log.a (pg, reflect the degree Of uncertainty Of the identity Of an individual chosen at random from a community. Shannon index values range from 0 to l; the higher the index value, the more diverse 107 the forest community“. The Shannon index satisfies all essential criteria needed for a diversity index to be useful (Hunter 1990): 1) the diversity index should be maximized when the proportion Of each species is equal (an even community): 2) for two completely even communities, the one with the greater number Of species should have the larger index value; 3) for a community that can be classified in more than one way, the sum Of the index values in each classification should equal the index value using all levels Of classification: 4) diversity indices should increase with transfers of abundance from one species to other less abundant species within the same community. Analysis Of variance was used to test differences between treatment means. Differences between individual mean values were calculated with Duncan's multiple range test when the p-value was less than 0.05. Results of Management Effects on Stand Diversity Initial (1951) stand basal area Of trees 4.6 in. d.b.h. and larger did not differ significantly between treatments and averaged 92 ftZ/acre (Table 3-1). Sugar maple was the largest component in all three replications but was more abundant in replications 2 and 3 (Table 7-1). Replication 1 was much more diverse than replications 2 and 3 (Shannon indices Of 0.864, 0.358, and 0.408, respectively, Table 7-1). Forty years later, basal area differed by treatment and ranged from 147 ft? in the control plots to 73 ft? in the heavy selection cut plots. Initial cutting decreased the diversity Of trees 4.6 in. and larger only in the diameter limit cut (Table 7-2). Shannon's index dropped from 0.310 before cutting to 0.219 after cutting. The change in ‘Some researchers use log base e (natural logarithms) for scaling the Shannon index; however, either method is acceptable (Drawer and Zar 1984). 108 diversity from the initial cutting was not significant for the other treatments. After 40 years Of management, diversity had decreased in all uneven-aged treatments (Table 7-2). The change in diversity was not significantly different between treatments and no pattern Of change was apparent. Shannon's diversity index decreased the greatest in the medium selection cut (0.220) and the least in the light selection cut (0.107). In 1992 basal area Of species other than sugar maple (as a percentage Of total residual basal area after cutting) was by far the highest in the silvicultural clearcut treatment (85 percent). This was followed by the shelterwood (56 percent), the light selection (45 percent), the commercial clearcut (38 percent), the control and crop tree treatments (37 percent), and the medium selection (34 percent). The heavy selection (19 percent) and the diameter limit (18 percent) cuts had the lowest percentages Of basal area in trees other than sugar maple. Results of Management Effects on Regeneration Diversity Saplings at the time the study was established were almost nonexistent. In 1957, some seedlings had grown into the sapling size class, but diversity was low and not significantly different among treatments (Table 7-3). Shannon's diversity indices in the uneven-aged treatments ranged from 0.100 in the heavy selection tO 0.207 in the light selection. By 1990, significant differences in diversity occurred among treatments. Shannon's diversity indices ranged from 0.080 in the control to 0.350 in the heavy selection plots for all trees (Table 7-3). Generally, the heavier the cutting or less the basal area in 1990, the greater the diversity. A strong negative correlation (p < 0.01) exists between current basal area and Shannon's diversity index (Figure 7-1). 109 0.4 ., Heavy selection 6 ' I I Medium selection .2 0 3 In _ I Light selection >.. .1: - Crop tree I a) In- 0 e e .2 0-2 ‘ Diameter limit '0 I .0) q C 8 c: 0 1 " Control (0 I S . U) 0 I I I I I I E U 70 90 1 10 130 150 1990 stand basal area (ftz) Figure 7-1 Comparison of 1990 overstory basal area and Shannon's index of saplings 2.0 to 4.5' d.b.h. for uneven-aged treatments. 110 TABLE 7-1. Species composition by percentage of total in each of three replicates prior to initial harvests in 1951. I Species Rep 1 (t) Rep 2 (8) Rep 3 (S) Nean (8) I Sugar maple 32.2 79.9 77.7 63.3 White ash 9.3 3.9 1.6 4.9 Basswood 15.3 6.8 3.0 8.4 Yellow Birch 11.1 5.3 2.7 6.4 Red maple 12.0 0.0 0.0 4.0 Hemlock 10.3 0.1 4.1 4.8 Paper Birch 3.7 0.2 0.0 1.3 American Elm 1.9 1.8 1.6 1.8 Black cherry, 0.0 0.0 1.0 0.3 Black ash 1.9 0.0 0.0 0.6 I N. Red oak 0.1 0.0 0.0 0.0 I Balsam fir 0.2 0.2 0.8 0.4 I Ironwood 1.2 1.8 7.5 3.5 Aspen 0.5 0.0 0.0 0.2 White cedar 0.2 0.0 0.0 0.1 Shannon's index 0.864 0.358 0.408 0.622 111 TABLE 7—2. Shannon's diversity values and changes for trees 4.6“ d.b.h. and larger in 1951 (before and after cut) and in 1990. Changes from initial condition Partial 1951 1951 After cut Before After cut treatments cut cut 1990 1951 1990 Control 0.438 0.438 ab| 0.305 0.000 0.133 Diameter 0.310 0.219 c 0.184 0.091 0.126 Limit Crop tree 0.509 0.501 a 0.305 0.008 0.204 Light selection 0.433 0.424 ab 0.326 0.009 0.107 Hedium Heavy selection 0.326 0.311 abc 0.198 0.016 0.128 p value 0.205 0.014 0.364 0.001 0.169 — _ 'Heans followed by the same letter are not significantly different at the 0.05 probability level based on Duncan's multiple range test. TABLE 1-3. 112 by treatment in 1957, and in 1990 for all trees and commercial trees. Shannon's diversity values of saplings 2.0” to 4.5“ d.b.h. Partial out All trees Commercial trees treatments 1957 1990 1990 Control 0.128 0.080 c' 0.038 a Diameter limit 0.122 0.183 be 0.144 ab Crop tree 0.138 0.277 ab 0.175 b Light selection 0.207 0.250 ab 0.141 ab h Medium selection 0.118 0.345 a 0.238 b Heavy selection 0.100 0.350 a 0.229 b p value 0.3640 0.0001 0.0013 Even-aged All trees Commercial trees treatments 2.0“ 8 larger 2.0” 8 larger Shelterwood --- .441 .351 Silvicultural clearcut --- .465 .322 Commercial clearcut --- .477 .368 _ i 'Means followed by the same letter are not significantly different at the 0.05 probability level based on Duncan's multiple range test. Sugar maple comprised greater than 50 percent Of the saplings in all treatments except the medium selection, shelterwood, and complete clearcut. The proportion Of sugar maple was related to harvest intensity: the heavier the cut, the less sugar maple, except in the diameter limit cut. Stocking of saplings was greatest in the medium selection cut and least in the control (171 and 41 saplings/acre, respectively, Table 7-4). Considering the stocking Of commercial trees only, the heavy selection had the highest number Of saplings at 134 per acre. Comparing the regenerated stands, stocking is about the same in the complete clearcut and shelterwood (693 and 734 trees/acre, respectively) and much less in the commercial clearcut (347 trees/acre). The shelterwood had 113 more uniform stocking (plot stocking levels ranging from 640 to 920 trees/acre compared to the complete clearcut (ranging from 150 to 1210 trees/acre, Table 7-4). Diversity values were similar for all regenerated treatments with diversity indices ranging from 0.441 to 0.477 for the shelterwood and commercial clearcut, respectively (Table 7-3), as compared with the entire study area's initial average index Of 0.622 (Table 7-1). Species composition in these three even-aged treatments varied. The high percentage Of white ash (50.7 percent) in the complete clearcut and black cherry (20.1 percent) in the shelterwood were Of particular interest (Table 7-4). Economic Analysis Methods Price estimates for each species were Obtained from reported prices (Chapter 4, pp. 62-67) using the linear regression model: Eq. 2: ln(P”.) s a + b(Year) Where ln(P)- natural log Of the reported stumpage price, and the corresponding year is the dependent variable. The b parameter is interpreted as the annual rate Of compound price change. Each cutting treatment was evaluated using these estimated prices for the four major hardwood species and hemlock present on the study sites. Other hardwood sawtimber2 comprised less than 6 percent Of stand volume. Management costs for cutting treatments were derived as discussed in Chapter 5. 2Other sawtimber included small (<1.5t Of basal area) amounts Of American elm (Ulmus Americana L.), white birch, black ash (Fraxinus Nigra Marsh.), black cherry (Prunus serotina Bhrh.), and northern red oak (Quercus rubra L.). 114 ucoaummuu ooozuouaucn ecu :0 moeuomn Hocuo Has no an.oa oomaumeou muuocu madam. usuumoHo hem m.n m.w~ m.a o.H o.v o.hH o.om HmwoumEEOU usousmao new o.oH o.oa 0.0 0.0 5.0 b.0m a.HH .uH=O«>me vmm _O.Ha m.m~ o.o o.o 5.0 h.mH a.hv voozumuaocm eudolasouu com-Ice>fl moH 0.0 h.bH 5.0 o.o o.m m.va H.mo .Hmm >>m0= aha 0.0 m.Ho m.o m.H m.m a.HH h.ov .Hom ESHOOE ooa o.o H.ma m.o m.o m.w o.m o.mo .Hmn unqu mm m.v o.aH o.a w.o o.va m.m o.mo moms mono UHEAH mm m.o H.o H.m o.o m.H a.HH a.mh Honesman Ho H.a m.h h.H 0.0 0.0 0.0 h.mm Howucoo un\oouu mowuomm COOS sauna nomad COOS nun Sandi nomad: nonuo down pom Sodas» neon nude: ummsm nudoflusouu can unwound EOOHU HHI N0 “HOOHOQ .omeu nu nude-esouu one newuoma ma ansuusm dd mosesveuu one swosxessuu n4 unsold: A.£.n.o .du m.e|o.av mswdmsm .olh manna 115 Volumes for the 1952 and 1962 cuts were derived using the volume line equation method (Erdmann and Oberg 1973), due to a lack of merchantable height data early in the study. Individual tree volumes were calculated in the 1972, 1982, and 1992 cuts from merchantable height measurements, diameters, and appropriate deductions for cull. A marginal analysis was conducted using Quick-Silver v. 2.0 (Vasievich et a1. 1984). Financial performance for each treatment was compared to the control (no management) treatment. Values of the residual merchantable stands in 1992 were included in all analyses to reflect total value for each treatment. Economic returns did not include tree quality changes due to treatments except as reflected in residual stand values. Net present values were calculated using a 4 percent real discount rate and adjusted to 1990 U.S. dollars for reporting. Economic Results Net present values (1990 U.S. $ per acre) for the differences among the seven treatments and the control ranged widely (Table 9-1, p. 143). The medium selection had the highest net present value ($54/ac) followed closely by the heavy selection ($51/ac) and the diameter-limit treatment ($48/ac). The light selection ($40/ac) and the commercial clearcut ($32/ac) ranked fourth and fifth, respectively. Although the diameter-limit cut ranked third among uneven-aged treatments, its actual value was probably lower due to two reasons: the heavy early harvest took small sawlog material before it had reached financial maturity, and the residual was degraded primarily to a low- value pulpwood stand. A subsequent analysis considered the economic effects Of these changes in stand tree quality (Chapter 8). The high residual sawtimber volume Of the medium selection and heavy selection treatments combined with the low percentage Of below-grade material in their residual stands helped give them high 116 financial rankings. The lack Of a commercial residual in the shelterwood ($14/ac) and commercial clearcuts ($32/ac) caused their rankings to be much lower. Financial performance for these even-aged treatments may improve when current volumes attain commercial size in the next few decades. Discussion Traditional views in the forestry community have held that, when selecting a silvicultural system to manage a stand, the first criteria should be the Objectives Of the landowner (Smith 1986). Economic concerns Often dictate these objectives. Aesthetics and tree species diversity, however, are Of increasing concern in forest land management. The discussion Of these results focuses on the trade-offs between these Objectives (Figure 7-2). W The 8-in. diameter-limit resulted in the least amount Of total regeneration diversity Of any treatment except the control. Northern hardwood stands that begin as diverse mixtures can easily become sugar maple monocultures through diameter-limit cutting. Sugar maple monocultures also may develop under light selection treatments, but diameter-limit cutting brings on the near-monoculture condition faster than any other treatment except the control. Heavy cutting Of mature intolerant hardwoods appears to have led tO ideal conditions for sugar maple reproduction. Many Of today's nearly-pure stands Of sugar maple may have been created by diameter-limit cutting in the past. 117 0.5 l I X Shelterwood Clearcut .8 0.4 __ .................................................................... 5 Heavy > I s ’3 Medium 2 0'3 _ ............................... . ................... . ................ 0) .2 a ‘0 ...................................................... C 0.2 .............. O C Diameter . 3 Limit z 0.1 L— .................................................................... CD 0 2 Control , 1 L , 1 0 10 20 30 40 50 60 NPV in $Iac. Figure 7-2. Comparison of net present value and Shannon's diversity index of saplings 2.0 to 4.5“ d.b.h. NPVs are based on the differences between the control and cutting treatments. 118 Landowners with short time horizons may favor diameter-limit cutting because it gives quick financial returns, and avoid the more labor-intensive and expensive basal area cuts. Other studies (Reed et a1. 1986, trickson et a1. 1990) have also found diameter-limit cuts to be economically attractive for short-term management regimes. The light selection treatment in this study was not as profitable as the other uneven-aged cutting methods. This result comes as no surprise, since others (e.g., Lyon 1986 and Reed et a1. 1986) have demonstrated the economic infeasibility of these cuts. Also, the light selection was the only uneven-aged treatment other than the control that lost diversity in the commercial sapling class (Table 7-3). After four cuts, stand structure in the medium selection treatment came very close to the structure suggested by Arbogast (1957). The Arbogast guide is widely applied in selection cutting throughout the Lake States. This guide produces sustained yields of around 250 board feet per year for 50 years once the desired structure is reached (Crow et a1. 1981). Rays and Buongiorno (1989) suggested alternative economic cutting strategies, based upon the uncertain growth of forest stands and fluctuations of product prices. These results, however, suggest the medium selection treatment not only sustains yield but has the highest economic value of the treatments in this study. Furthermore, tree sapling diversity is greater in the medium selection than all other uneven-aged treatments. Initial heavier cutting and creating of canopy gaps in unmanaged hardwood stands (Erdmann 1986) may increase sapling diversity even further. I!!!:l§l§.fl£§l§!l£ni Shelterwood and clearcutting both maintain or enhance tree species diversity, while diameter-limit cutting will not increase diversity in hardwood forests (Table 7-3). Also, shelterwood's potential for enhancing diversity can be increased by modifying the shelterwood to 119 include scarification (Godman 1985) and by controlling deer densities when needed through heavy hunting (Relty and Nyland 1981). The higher the value placed on tree species diversity, the more appealing shelterwood will be as a silvicultural system. However, overall financial performance of the shelterwood in this study falls considerably short of the uneven-aged methods (Figure 7-2). Past research has shown clearcutting to produce highly variable regeneration results in Lake States northern hardwoods (Tubbs 1977a). Our results also suggest that clearcutting may not be as financially attractive as other methods; however, the Argonne clearcuts occurred in pulpwood and small sawlog stands that may not represent commercial stands today. Summary This study shows that there are clear differences in economic returns and regeneration diversity under various cutting methods. The medium selection treatment produced the highest diversity index of all uneven-aged treatments and the highest economic returns. Production of quality sawlogs and compatibility with aesthetic objectives are additional benefits. Results also show that regeneration diversity in the diameter-limit cut was half that in medium and heavy basal area cuts. Both the medium and heavy selection treatments also provided greater economic returns and species diversity than the diameter limit cut (Figure 7-2). Results indicate that the crop tree method may be profitable if landowners are willing to wait. For even-aged treatments, the commercial clearcut produced moderate financial returns and high species diversity. Shelterwood cutting did provide high diversity and initial revenues but little commercial value after 40 years. If landowners have tree species diversity as their primary objective and financial return is secondary, then clearcutting is the 120 preferred treatment in Lake States northern hardwood management. However, if higher economic returns are the key objective, with tree species diversity as a secondary management goal, then a medium selection treatment produces the best combination of both economic and diversity results. CHAPTER .8 HARDWOOD IAWLOO CHILI!!! DCOIONIC EVILUAIIOI OP CUTTIDO METHOD IMPACTS Lake States northern hardwood stands today are largely second- growth, resulting from heavy commercial cutting at or near the turn of the past century. Tree quality has been a concern in these forests because many stands have suffered degradation due to high-grading. These stands could produce far better quality than in the past, yet higher tree quality depends upon strong markets to remove low quality hardwoods from heavily overstocked stands. Recent development of vigorous pulp and fuelwood markets in many areas of the Lake States has created a new economic opportunity for silvicultural treatments to improve tree quality (Webster 1986). Forest managers already have most of the silvicultural techniques needed to produce higher quality in northern hardwoods (Erdmann 1986). What they lack is adequate information on the returns they can expect from silvicultural treatments that may improve tree quality. Managers need to know if investments in tree quality are worth the time and expense involved, and what methods are most profitable under specific stand and site conditions. This chapter answers some of the above questions, based upon a 20-year tree quality study. In 1971 this study was superimposed on the Argonne cutting methods study. Changes in the silvicultural and economic yields due to improvements in individual tree quality were analysed for six of the original study's nine cutting methods. Literature Review Several past studies (lrdmann 1986, Nyland 1986, Tubbs 1977b, Eyre and Zillgitt 1953) suggested that northern hardwoods on good sites are best managed for the sustained production of high grade veneer and sawlogs. However, there is scant economic evidence to support these 121 122 assertions. Most studies on tree quality have been theoretical or silvicultural in nature (lcCay and Denald 1973, Brisbin and Dale 1987). tconomic evaluation of long—term tree quality response to treatment has not been possible previously due to a lack of adequate data. Accurate evaluation of hardwood tree quality is crucial to the economic management of high-value hardwoods, because the differential in price between high-quality and low-quality end products is large. For example, a recent issue of the weekly fig;§ugg§_figgigw (March 1991) reported northern white ash 4/4 green lumber prices (delivered) of $775 PAS grade and only $215 IZA grade. Price differentials for northern basswood were similar--$625 FAS grade and $175 #23 grade. Available data for delivered logs by quality, though scarce, also show that prime and #1 hardwood logs often command more than twice the price of #2 or #3 logs (Hoover'). Obviously, improvement in log quality of one or more grades could produce handsome financial returns. However, tree quality can only be estimated accurately after end products to be produced are defined (Rast et. a1. 1987). Hardwood tree quality, therefore, is a function of tree and wood characteristics (Brisbin 1985), the most important of which are: 1) growth rate and bole form of the tree; 2) knots and limb-related defects; 3) decay and insect damage; and, 4) size of tree. Prior research has shown that much of the improvement in stand tree quality is the result of increasing sise of hardwood sawlogs. Larger siaes make increased yield and utilisation possible (Brisbin and Dale 1987, Sonderman 1987, Banks 1976, Boyce and Carpenter 1968). Improved growth rate in response to crown release and thinnings has three important functions relative to tree quality. First, it makes individual grade improvements due to size increases possible much sooner (trdmann 1986, Johnson 1986). Second, it improves the overall stand 'Hoover, W.L. Unpublished data on delivered log prices according to grade. West Lafayette, IN: Purdue University, Department of Forestry. 123 quality through increased vigor. Finally, it can stimulate growth that can cover up or reduce bole defects (Godman and Books 1971). Yet there is a definite trade-off between stand growth response and stem quality-- thinnings that are too heavy will stimulate epicormic sprouting and lowering of crowns, increasing defect and decreasing quality and value (Godman and Books 1971, Crow 1985). Although there is a close relationship between diameter growth and tree grade improvement, grade improvement does not depend on diameter increases alone. Diameter growth is a necessary but not a sufficient condition for grade improvements in hardwoods (Herrick 1952, Godman and Mendel 1978). Loss of quality may occur because of crook, rot, branch defects, sweep, seams, mineral streak, bird peck, insect holes, limbs, knots, and bud clusters (Banks 1976). The species of tree or inadequate growth between stand entries also can contribute to trees that do not improve in tree grade. In medium- to high-value hardwoods, research has confirmed that quality change over time is at least as important, if not more important, than diameter and height growth in contributing to timber stand value (Rast et a1. 1987, Sonderman 1986). In Lake States northern hardwoods, increases in tree grade over a 10-year period consistently added greater value than either gains in merchantable height or diameter growth (Godman and Mendel 1978). Previous work in Ohio oak stands by McCay and Denald (1973) also confirmed the financial superiority of grade increases over height or diameter gains alone. Their study showed a doubling in the rate of value increase for lB-in. red oak trees that changed from a butt-log grade 2 to grade 1, versus those with no change. On a stand basis, particularly in vigorous pole-sized hardwoods, diameter growth is the primary contributor to grade improvements. Properly timed thinnings greatly reduce the time necessary for northern hardwoods to reach larger sizes and, therefore, improvements in tree grade. Johnson (1986), for example, found that the time required to 124 reach minimum financial maturity (16 in.) was 45 percent less on thinned versus unthinned trees in over-stocked pole-sized northern hardwoods of Michigan's Upper Peninsula. Such reductions in maturation time can yield significant increases in rates of return to hardwood management. Merchantable height differences also occur between even- and uneven-aged hardwood management treatments. Prior research has not considered the variation in hardwood quality and yield because of these differences. Erdmann (1986) and Crow et a1. (1981) mention this difference, but little empirical data has been available to consider economic impacts of this growth difference between even- and uneven-aged management. One ”rule of thumb” sometimes used by managers is that uneven-aged management can produce an extra half-log of merchantable height in Lake States northern hardwoods. However, this judgment has not been verified by research. Mautiyal (1983) has proposed a theoretical model for managing uneven-aged selection hardwood forests based on the theory of financial maturity. Using tree vigor as a criterion, he derived optimal rotations from crops of trees that have increasing stumpage price functions, thus effectively “safeguarding" the penultimate (next-to-last) crop before financial maturity. Tree grade and size changes caused most of these increasing stumpage prices. Nautiyal's work ties in with studies on hardwood tree quality because tree grade improvements lengthen financial maturity (Godman and Mendel 1978). The financial value attributable to tree quality changes has two components: value from trees already harvested, where treatment may have affected quality; and the value of residual trees, which may or may not change in quality in response to treatments. Higher quality logs yield a greater proportion of higher value lumber, the key factor in valuing changes in tree quality. Herrick (1956) used log and lumber price relatives rather than dollar values to analyse differences in tree and log quality. Expressing the prices of the various grades as ratios, 125 using log or lumber grades as reference grade, Herrick found that differences in tree and log value could be calculated more quickly and easily than with other methods. Hanks (1976) developed equations that gave graded lumber yields for tree-graded hardwood logs. He developed these equations for 17 hardwood species. Hanks (1976) used hundreds of logs from the Lake States, Midwest, Appalachians, and Hortheast. These lumber yield equations have become the standard for developing hardwood yield estimates from graded trees throughout the last. This analysis used these equations as a principal tool in this portion of the study. Data Diameter at 4.5 feet (d.b.h.) was measured on 5-year intervals beginning in 1951 from five, 0.5 acre sample plots within each treatment plot. Sawlog trees (trees larger than 9.6 in. d.b.h.) were graded according to Hanks' tree grading system in 1971 and 1991 (Hanks 1976). Cull percent was calculated by techniques reported by Gevorkiantz (1947). Merchantable sawlog heights (number of 16-foot logs) to a diameter inside bark of 7.6 in. on the small end, or to a merchantable log stopper, were recorded in 1971 and 1991. The stand in 1951 primarily consisted of pole-sized trees (4.6-9.5 in.). Scattered among the treatments were some larger diameter residual trees from the early 1900's logging. Because grade changes are unlikely in these residual trees and most often these trees are poor quality, only those trees that were pole-sized in 1951 and had grown into sawlog-sized trees by 1991 were analyzed for tree grade improvement. The silvicultural analysis of tree quality did not include trees cut during the period. Final economic analysis, however, included these large holdover trees because of their contribution to total value. For the silvicultural data (Tables 8-2 and 8-3), analysis of variance tested differences between treatment means. Differences 126 between individual means were calculated with Least Significant Difference when the p-value was less than 0.05. The silvicultural data were originally in a variety of hand- written formats. They were put into an individual tree data base computer format using PARADOXO v. 3.5 (Borland International 1985,1990), a relational database management software package. The query and table functions in PARADOX. made it possible to link tree grade, diameter growth, and species data for each treatment. Tree grades for the 1952 and 1962 stand entries were not available. However, log grades were available for 332 trees from 1951 that were in the original stand. Of these 332 trees, about 85 were trees that were harvested in 1951 or 1961. Analysis of log grades for these 85 trees showed that most (60) of these graded trees were either of small diameter or grade three. Therefore, they contributed very little #1 common and better lumber to the analysis. However, 25 trees in the selection and diameter-limit cut were of log grade 1 or 2 or were grade 3 trees above 15 inches in diameter. Tree grades were derived for these 25 trees so their contribution of upper—quality lumber could be included in the regression analysis of lumber quality changes. Bconomic Hetbods Reported Wisconsin hardwood lumber prices were used for economic analysis of cutting method treatment values. These are the prices paid for five grades of green, 4/4 hardwood lumber at the mill--FAS, selects, #1 common, #2 common, and #3 common. Three sources of Wisconsin data were consulted--the Wisconsin Forest Products Price Review, Lumber edition (Peterson 1981, 1991), the Hardwood Market Report, Northern Market (Lemsky 1971-91) and the Weekly Hardwood Review (1971-92). Gevorkiants' and Olsen's (1955) volume equations calculated individual tree gross board foot yields, according to the formula: 127 volume (Scribner) I AHTILOO 10 {-.121 - 4.21(1/d.b.h.) + l.16(loglo d.b.h.) - .306(l/merch. length) + .723(loglo merch. length)). Het board foot yields were determined by multiplying gross board foot yield derived from the above equation by the percent cull for each tree, a variable measured in the study. Tree grades (Hanks 1976) of both harvested and residual sawlog trees, by treatment, were measured for 1971, 1981, and 1991 following Hanks' (1976) methodology. Merchantable lengths of each log were also measured, and used to derive factory grade lumber yields for each major treatment. These yields were developed by applying Hanks' (1976) yield equations to treatment data from the Argonne cutting methods study in 1972, 1982, and 1992. Treatment lumber yields were calculated for 1972 and 1982, while both treatment lumber yields and residual stand quality lumber grade distributions were calculated for 1992. A few changes were made in Hanks' (1976) lumber grade yields to make treatment comparison more feasible. Instead of using seven or eight lumber grades, this study used only six. A single FAS grade included the grades FAS, FASlF, and Selects. This change was made because differentials are very small between these grades, generally only 35-20 per thousand board feet. Similarly, the grades 2A and 28 were combined into a single 20 grade for red maple, because it was the only study species not having a single #2 grade, and because red maple yields were less than 2 percent of this study's volumes. However, because of large reported price differentials between grades 3A and 38 (Peterson 1981, 1991), particularly for sugar maple, these grades were included in the analysis for all species. The five grades used in the analysis were: FAS, #1 Common, #2 Common, #3A, and #38 (or #3C). These 128 grades and their prices per thousand board feet (Scribner) by species and pertinent year are shown below (Table 8-1). Economic values of treatment yields were calculated before the 1972 cut and after the 1992 cut, to determine the effect of quality treatment on value yield. Total economic values also included residual treatment values that reflected gains in the quality of the residual stands. Prices for the graded hardwood lumber yields came from the Wisconsin Forest Products Review, Lumber edition (Peterson 1981, 1991) and the Hardwood Market Report (Lemsky 1971, 1981, 1991). Only variable costs for the 20-year period were used in the analysis (Chapter 5, pp. 82-85). A marginal cash flow analysis was conducted using Quick-Silver v. 2.0 (Vasievich et a1. 1984), where each treatment's performance was compared with the control. Silvicultural Results: Changes in Tree Quality and Grade Because grade improvement is the primary contributor to increased value (Godman and Mendel 1978), discussion of silvicultural improvements in quality focused on grade changes (Tables 8-2 and 8—3). Average tree grade improved the most and was highest in 1991 for the medium selection treatment (Table 8-2). The medium selection was the only treatment that averaged more than a full grade in improvement. It also had the most (32 percent) grade 1 trees in 1992 after cutting and the most (66 percent) trees in grade 1 and 2 combined (Table 8-3). 129 TABLB 8-1. Lumber prices by grade used to evaluate lumber yields (per 1000 MBF, 4/4 lumber F.0.B.). * Prices for ash in 1971 were quoted for brown ash, the trade name for black ash. In most markets white and black ash lumber were mixed, with the same price in 1971. Notes: Above are median prices at the mill. These are nominal prices. FAS, PAS IF, and Selects combined into a single FAS grade. #2A and #28 red maple were also combined into a single #2C grade. Grade premiums of 380-100 per MBF were applied for FAS/Selects alone grades for sugar maple, basswood, and yellow birch in 1981 and 1991. Species ,7 FAS #IC #2C I #3AI #38 Hbvember 1971 Sugar maple 270 180 90 70 65 White ash* 250 175 78 --- 65 Basswood 260 157 80 --- 59 Yellow birch 345 175 100 83 65 Red maple 240 180 85 --- 65 November 1981 Sugar maple 450 335 210 190 140 White ash 540 350 182 152 127 Basswood 462 295 136 145 100 Yellow birch 585 325 210 192 137 Red maple 370 207 160 140 127 Hovember 1991 Sugar maple 620 415 305 285 165 White ash 755 515 215 170 140 Basswood 650 295 170 110 100 Yellow birch 625 325 220 190 150 Red maple _____ 485 335 __ 200 160 150 130 .mo.o v m .ucououuep maucsoeuecmea so: one cesaoo mas sq weaved ease or» he veSOHHOw assesn ..mHucv scausw>ep owsossumu .Hmmd on area sown women new» :4 masseuse an oewoowo masseusoua. — n Ace.oe 0.0 an .ov.o. m.H a .Hv.ov o.~ coauooao- aroma a Aac.o. e.o o mam.oe n.~ n .5H.o. o.n asses Iweueeedn n 3.13 To on 32% Ta an 33.8 o5 monsoon r n :28 m6 8. ~26. «A a 323 o.~ 33032. m>sem nu ion.ov m.o n ime.o. m.H no AmN.ov m.N easeaow eewu mono e Amn.ov H.H e Avn.ov m.H nan ".mn.o. to.“ noduoeaon seeps: essence” Ham“ amen wselwsewfl _ .«men 00 upon loam esseuudw evswm use .ueeu use area aw aselwsew» he sensum sew» ensue»: .mlm mamdfl 131 .mo.o v m .useuOHuuo haucsuwuacmen uo: mus cesaou has :4 weaves same on» he oOBOHHOu asses. .Amnuce :oeuue>ou ouaocuuma .H spawn seen» uo codewomoum ca eaeowoou an pewsuuo suceausoua. a Am.mfls an n .m.m~. cm u Av.oae «a u .o.m . m assamnuouoano a «w.FMe em on .o.H~e mm on .o.nae mm on An.¢av «a Aouucoo n .e.~ae as an .m.m~. mm as .m.o~. on n .o.oav he manages new» mono n ~m.me N van Am.n~. on a ~m.aav me n .m.o~v mm .Hon menu: a .m.ofi. an on Am.ma. an on Am.hae mu a .p.mue an .Hou uzoea n .e.hse om u Ae.o~. em nu Ah.o~. en .a ~.m.p~. .un .Hoa snows: oushm laden n opsho a snake a evshO analwsewa .nmmu .svsum he sees» me useuuem .nlm manta 132 Grade improvement was slightly less than one grade for the crop tree treatment, and 0.8 grade for both the light and heavy selection. The control (2.1) and diameter-limit (2.3) had by far the lowest average tree grades in 1991, but their increases (0.7) were not significantly different from the heavy or light selection treatments. After 20 years of quality management, the light selection had the same number (32 percent) of grade 1 trees as the medium selection, but fewer grade 2 trees. The heavy selection had more grade 2 trees (43 percent), but fewer (19 percent) grade 1 trees. The diameter-limit had the fewest grade 1 (5 percent) and grade 2 (14 percent) but the most grade 3 (50 percent) and total number of trees that were grade 3 and below (81 percent). The control treatment had the highest number (37 percent) of below grade trees, while the heavy selection had only two percent of its trees below grade. These results strongly confirm the effects of silvicultural treatments that remove cull and low-grade trees in selection management?. Results: Regression Analysis of Lumber Quality Changes In order to proceed with the economic analysis, it was first determined whether there were, in fact, significant changes in treatment quality. A linear regression model was used where the change in #1 common and better lumber was analysed by treatment and plot. Independent variables were the cutting treatments, stocking of white ash and basswood, and average initial tree size. The dependent variable was the total change in #1 common and better lumber (both harvested and residual). ’Strong, T.F.; Erdmann, G.G.; Niese, J.H. (in preparation). Forty years of alternative management practices in second growth, pole-sized northern hardwoods. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station (unpublished manuscript). 133 In the best regression model, treatment effects explained 32 percent of the variation in #1 common and better lumber. Adding a term for percent initial stocking of white ash and basswood improved the explanatory power of the model to 51 percent. A third term, average initial (1951) tree sise, boosted the model's explanatory power to 59 percent. In this model, five of the six treatments had significant t-ratios (Table 8-4). The analysis of variance is presented in EARLS I-C. IQUAIION: Regression statistics for plot analysis of variation in #1 common and better lumber. Total #1c* :- «19 + 184 LIT! + 123 use + 176 m + 31.: case - 108 nus + 11.a avatars + 20.3 ans 1. ass —_ — _1 Independent Coefficient Standard t-ratio p-value Variable Deviation Constant -418.90 124.40 -3.37** 0.001 Light 183.65 55.89 3.29** 0.002 Medium 123.15 56.53 2.18* 0.032 Heavy 175.54 56.22 3.12** 0.003 Crop 87.21 55.03 1.58 H.S. 0.117 Diameter -108.07 54.93 -1.97* 0.053 Limit Average 77.80 15.07 5.16** 0.000 tree sise Basswood] 20.29 3.64 5.57** 0.000 White ash s - 140.4 a’ - 58.7t Adjusted a” - 55.0: * - significant at the a - .05 level ** - significant at the a - .01 level 134 Analysis of variance for #1 common and better lumber. m Mean Source DF Sum of Squared F-Value p-Value Squares Brror Regression 7 2212976 316140 16.04 0.000 Srror 80 1557239 19712 Total 87 3770216 T _ Source DF SEQ Sum of Squares I Light 1 476757 Medium 1 224590 IHeavy 1 221386 Crop 1 164855 I Dia.-limit 1 107693 IAvg. tree 1 405081 Basswood 1 612614 /ash The regression model above indicated that changes in #1 common and better lumber were significant. Then, a one-way analysis of variance was conducted to determine how significant the changes in #1 common and better lumber were by treatment (Table 8-6). The light selection treatment had the greatest change (562 Bdft/plot) in #1 common and better lumber over the study period. Changes for the medium (477 Bdft/plot) and heavy (449 Bdft/plot) selection treatments were also significantly (a - .05) higher than the other three treatments, but did not differ from each other. The crop tree (374 Bdft/plot) and control (351 Bdft/plot) did not differ from each other, but were also significantly greater than the diameter-limit treatment (214 Bdft/plot). For comparison, these data are also shown on a per acre basis (Table s-s). 135 TAILS S-6. Increase in #1 common and better lumber, by plot and treatment (1951-92). — l i Mean change in Increase in #1C 8 better #IC 6 better lumber, Standard lumber, avg. Treatment 8d.ft[plot Deviation 8d.ft[ac Light 561.6 195.5 5616 Selection Medium 477.1 182.9 4771 Selection Heavy 449.4 175.2 4494 Selection F Crop 374.1 218.1 3741 Tree Diameter 213.8 115.0 2138 Limit rControl 350.6 250.4 3506 The above results showed that there were significant changes in lumber quality variation by cutting method. This result gave credence to the economic analysis that follows. Sconomic Results new The diameter-limit treatment had the greatest net harvested volume for the study period, 4452 net board feet (HBF) per acre (Table 8-7). Its lumber value cash flows averaged $1763 per acre. The heavy selection treatment was second with 3216 net board feet harvested per acre with cash flows of $1405/ac. Ranking third was the medium selection with 2180 NBF/ac and $1079/ac. Fourth ranking went to the light selection (2175 HBF/ac.) with $933/ac. Finally the crop tree treatment had the lowest net harvested volumes, 1916 HBF/ac., and total cash flows of $9l8/ac. 136 TAILS 8-7. Met harvested lumber volumes and total cash flows, 1971-92 (per acre basis, 1990 S). — F Treatment Met volume Total Cash Harvested Flows Lumber Value Discounted at(4t) Diameter-limit 4452 $1763 $805 I Heavy 3216 1405 953 I Medium 2180 1079 738 I Light 2175 933 583 I I Crop 1916 918 827 — — TABLS 8-8. Residual values of lumber yields, 1992 (1990 S). nu- [Treatment Residual Value I Light selection $3173 Medium selection 2671 Heavy selection 2182 Crop tree 2329 Diameter-limit 0 Control 2899 I We” Residual lumber value (real 1990 S) was greatest in 1992 in the light selection treatment ($3173/ac) (Table 8-8). The control treatment ranked next ($2899/ac), followed closely by the medium selection ($267l/ac), then the crop tree ($2329/ac). The heavy selection ranked last of the four uneven-aged treatments in residual lumber value ($2182/ac). The diameter-limit treatment (SO/ac) had no residual lumber value in 1992 because all sawlog-sized material was harvested in that year. 137 The above cash flows from harvests and residual lumber values were combined and then discounted to give a proper measure of economic values. Marginal costs (see pp. 82-85) were also applied to obtain the net present value rankings below. WW Marginal analysis (1971—92) of the remaining treatments as investments gave top ranking to the heavy selection ($542/ac) treatment (Table 8-9). The medium selection ($488/ac) came next, while the light selection ranked third ($443/ac). Met present value was lowest for the diameter-limit cut (-$247/ac). Lumber value in the diameter-limit was far below the value of lumber in the control treatment. TABLE 8-9. Comparison of net present value of the difference between the control and cutting treatments for total lumber value yield (1990 8, 4‘ discount rate). I— Met present value, Treatment Marginal revenues Heavy selection $542 I Medium selection 488 I Light selection 443 I Diameter limit -247 I Discussion Considering the need to balance both harvested and residual lumber values, both the heavy and medium selection treatments clearly rank ahead of the others. The heavy selection treatment had the highest total value of marginal revenues (Table 8-9, Figure 8-1), the highest discounted lumber value of harvests (Table 8-7), and the lowest percentage of poor quality trees in 1992 (2 percent). The medium 138 NPV $/acre 600/ 400 " ’ / / _ . .. /// /// 1:: _ W / / , / 0 . / :2:::::::..:: ::: ’ -200 " f / 7 7 / / / / / / / / Heavy Medium Light DL Figure 8-1. Comparison of harvested. residual. and total lumber values (1971-92) for four tree quality study treatments. I Total Value Residual Harvest Values are for differences between treatments and control. 139 selection treatment ranked second in MPV of marginal revenues and had the greatest improvements in tree grade, merchantable height, and number of residual grade 1 and 2 trees (Table 8-3). It also had the most favorable distribution of size classes that approached recommended (Arbogast 1957, Crow et a1. 1981) stand structure guidelines (Figure 2-1). Though it also had a high total lumber value, the light selection's value was heavily weighted by its residual value, and harvests tended to be light and marginally economic from an operability viewpoint. The crop tree treatment had a high total lumber value, but over 70 percent of its harvested value came in the first cut, which gave it excessive weight under discounting assumptions. Also, the crop tree treatment was overcut in 1972 at a level that was not sustainable and may have resulted in subsequent wind damage and mortality to the residual stand. The diameter-limit treatment had the highest volumes and cash flow for its harvested lumber, but was only third in present value of lumber (Table 8-7). Met present value of marginal revenues (-$247/ac) was lowest for the diameter-limit cut, and it had by far the highest (81 percent) number of poor-quality (grade 3 and below) trees before harvest in 1992. The light selection had the highest percentage (71 percent) of its total lumber value in the residual stand (other than the control). The medium selection had 62 percent of total lumber value in its residual, the crop tree 57 percent, and the heavy selection only 51 percent. The diameter-limit out had none of its total lumber value in its residual stand, because a complete harvest of all sawlog-sized material had taken place in 1992. These economic results can be used to elaborate on results of the overall economic analysis (Chapter 9). When the economic returns to tree quality change are included, then the economic performance of the 140 medium and heavy selection treatments improves, relative to the diameter-limit and clear-cut treatments. The gap in economic performance between the medium and heavy selection treatments and the diameter-limit is much greater because this quality analysis explicitly includes information on low-value, below- grade sawtimber, while the overall analysis does not. Also, this quality analysis only includes relative lumber value returns of sawlogs, and does not include pulpwood, which is the primary product of the most heavily-cut treatments. The quality results do not include potential returns to veneer production, although veneer trees may number 50-75t of residual large sawlog trees in the carefully-managed selection treatments. These quality analysis results emphasise the importance of one's management perspective and underlying assumptions. Taking a long-term view of the forest enterprise makes the realisation of these much higher returns to quality improvement possible. A shorter-term view will favor a greater proportion of pulpwood production, as shown in the diameter- lbmit results of Chapter 9. This analysis of tree quality returns does show that with current high prices in hardwood lumber markets, long-term management for tree quality can yield excellent returns for some owners. Summary W Analysis of six cutting treatments regressed on total change in upper-quality lumber showed that there were significant differences in hardwood tree quality of the six cutting methods over a 20-year period. Initial stocking of white ash and basswood, and average initial tree size, also helped explain differences between treatments. 141 new The heavy selection treatment had the highest economic returns in this study, yet its silvicultural quality--in terms of tree grade, merchantable heights, and lumber quality--improved nearly as much as the more lightly cut selection treatments (Strong et al. in preparation). The medium selection treatment also had high total economic returns, but a greater proportion of its value was in the residual stand, and it had greater improvements in silvicultural quality. Diameter-limit cutting in this study resulted in surprisingly low net revenues, and by far the lowest lumber and stand quality. For managers primarily desiring the highest economic return while enhancing tree quality, the heavy selection treatment was best in this analysis. More conservative managers desiring the greatest improvements in silvicultural quality, but also high total value, will find the medium selection treatment desirable. Under no situations where quality management is a goal is the diameter-limit treatment recommended. CHAPTII ’8 ICOMOMIC IMILIOII, IISULII, IMD DISCUBSIOM The following sections contain the results of the overall economic analysis for the nine cutting methods. All analytical results are stated in terms of marginal revenues that accrue to treatments as compared to the financial performance of the control treatment. The base analysis covered the 40-year (1952-92) period at a 4 percent real discount rate. Later sections discuss sensitivity to discount rate and marking cost changes. Then, the analysis was extended to include the 30-year (1992-2022) growth and yield projection. The extended analysis tested results for their sensitivity to the low, medium, and high price scenarios defined previously. Also, a sensitivity analysis was done to test the extended treatment response to changes in marking, sale layout, and road maintenance costs. Finally, the treatments were ranked and discussed according to their financial performance under the various economic conditions. Results of Analysis for 1952-1992 Mn The medium selection treatment ($54 per acre) had the highest net present value of difference with the control, using a 4 percent real discount rate (Table 9—1). Ranking second was the heavy selection ($51/ac) followed by the diameter-limit ($48/ac), the light selection ($40/ac), the commercial clearcut ($32/ac), and the crop tree selection ($30/ac). The shelterwood treatment ($14/ac) ranked last of all the treatments considered. The silvicultural clearcut was not included in this analysis due to lack of species-specific volume data in 1952. W Both the medium ($70/ac) and the heavy ($58/ac) selection treatments still ranked highest with a low discount rate of 2 percent. 142 143 However, the light selection treatment ($57/ac) now ranked far ahead of the diameter-limit treatment ($30/ac). The crop tree treatment ($30/ac) was about the same in financial return as the diameter-limit. The commercial clearcut ($20/ac) and shelterwood (-$19/ac) were at the bottom of the rankings under the low discount rate assumption. Raising the discount rate to 6 percent (real) changed the treatment rankings for the 1952-1992 investment period. The diameter- limit cut now had the highest financial return ($56/ac), followed by the medium ($45/ac) and heavy ($45/ac) selection treatments, in a tie for second ranking. Mext came the commercial clearcut ($38/ac), the light selection ($31/ac), the crop tree ($28/ac), and the shelterwood ($25/ac). TABLI 9-1. Met present values per acre (1990 S) of differences between treatments and control for 1952-1992 analysis. Discount Rate Treatment 2‘ 4t 65 Light $57 $40 $31 I Medium 70 54 45 Heavy 58 51 45 Crop 30 30 28 I Diameter-limit 30 48 56 Commercial CC 20 32 38 Shelterwood -19 14 25 -- The medium selection treatment was best at two of the chosen discount rates (2 percent and 4 percent) and tied for second ranking at the high discount rate. The diameter-limit cut ranked first at the high rate, third at the medium (4 percent) rate, and fifth at the low rate. The heavy selection treatment was second at all three discount rates (Table 9-1). The light selection, commercial clearcut, and crop tree generally fell in the middle of the rankings for all discount rates, 144 although the commercial clearcut improved its ranking at the high rate, while the light selection improved at the low rate. At all three discount rates, the shelterwood consistently ranked last of all treatments analyzed. W The 1952-92 HPV results were tested for sensitivity to changes in marking costs. For the basal area cuts, individual tree marking is the single highest cost category, while sales costs for all other treatments are lower, and less frequently applied. The medium cost for selection marking was $20 per acre in 1992 (1990 S). In the sensitivity analysis, this was varied by 20 percent to reflect a high cost of $24 per acre, or a low cost of $16 per acre. Higher marking costs for the basal area cuts (4 percent discount rate) caused the gap between the first-ranked medium selection cut and the diameter-limit cut to decrease by a marginal NPV of $2/ac. Rankings between the medium selection (first), heavy selection (second), and diameter-limit remained the same. Only by allowing sales costs for the diameter-limit to fall to low levels, while keeping selection marking costs high, did rankings change. Then, the diameter-limit was second, the heavy selection fell to third, and the medium selection still had the highest economic return. Financial ranking for the other treatments did not change when marking costs were varied, with one exception. Applying high marking costs to the selection treatments with a 6 percent discount rate caused the shelterwood treatment to move from last ranking to fifth, while the crop tree fell to last and the light selection fell to sixth. 145 Results of Analysis for 1952-2022 with TWIGS Growth Projection 3:11.9111 The medium selection treatment had the highest net present value ($75/ac) in the 70-year (1952-2022) base case, using medium price and cost assumptions and a 4 percent real discount rats (Table 9-2). Following it were the heavy selection ($68/ac), the diameter-limit cut ($65/ac), and the light selection ($61/ac). These rankings parallel those of the base case for the 40-year analysis (1952-1992), discussed previously (Table 9-1). The commercial clearcut ($53/ac) ranked fifth overall and highest of the even-aged treatments. Rounding out the rankings were the crop tree ($51/ac), the silvicultural clearcut ($46/ac), and finally the shelterwood ($27/ac). TABLS 9-2. Met present values per acre (1990 S) of differences between treatments and control for 1952-2022 analysis, using a 4 percent real discount rate. ---L, l --r Treatment Low Prices Medium Prices High Prices Light $45 $61 $153 Medium 61 75 131 Heavy 57 68 104 I Crop 40 51 119 Diameter-limit 64 65 11 I Commercial CC 51 53 32 Silv. CC 45 46 73 I Shelterwood — 49 27 -9 J WW Treatment rankings changed when discount rates were changed. At a 6 percent real discount rate, the diameter-limit treatment ($7l/ac) was first, followed by the heavy (SSS/ac) and medium ($54/ac) selection treatments. Both the commercial ($51/ac) and silvicultural ($46/ac) 146 clearcuts now had higher rankings than the light selection ($39/ac) treatment. As before, the shelterwood treatment ($36/ac) was last (Table 9-3). With a lower (2 percent) discount rate, rankings changed again in favor of the selection treatments. Again, the medium selection ($12l/ac) had the highest returns, followed this time by the light selection ($110/ac), heavy selection ($88/ac), and crop tree ($87/ac) treatments. The silvicultural ($40/ac) and commercial ($38/ac) clearcuts both were ahead of the diameter-limit cut ($24/ac). The shelterwood (-$36/ac) was last of all treatments (Table 9-4). Wanna Two other price scenarios were used in the analysis--a low price case and a high price case. In the low price case, the dominant species (sugar maple, basswood, and hardwood pulp) had stable real prices for the 70-year period. White ash had a real 2 percent price increase, while yellow birch had a 1 percent decrease. With the low prices and a 4 percent discount rate, the diameter- limit ($64/ac) had highest ranking followed closely by the medium selection ($61/ac). Then came the heavy selection ($57/ac), commercial clearcut ($51/ac), shelterwood ($49/ac), silvicultural clearcut ($45/ac), and light selection ($45/ac). The crop tree treatment had the lowest returns ($40/ac) (Table 9-2). The high price scenario assumed real 4 percent annual price increases for the major species (basswood and sugar maple), a 6 percent increase for white ash, and a 1 percent increase for yellow birch. Because of the variation in species composition and timing of harvests, the high price scenario had a major effect on treatment rankings. 147 TAIL! 9-3. Wet present values per acre (1990 8) of differences between treatments and control for 1952-2022 analysis, using a 6 percent real discount rate. Treatment Low Prices Wedium Prices gighiPrices Light $32 $39 $79 Medium 47 S4 84 Heavy 46 55 77 Crop 32 38 63 Diameter-limit 62 71 59 Commercial CC 47 51 S4 Silv. CC 45 46 52 Shelterwood 43 36 46 TABLI 9-4. Wet present values per acre (1990 8) of differences between treatments and control for 1952-2022, using a 2 percent real discount rate. F"""""""""r""""" """"' Treatment Low Prices Medium Prices nigh Prices Light $75 $110 $341 Medium 90 121 219 Heavy 73 88 111 Crop $8 87 281 Diameter-limit 58 24 -233 Commercial cc 53 38 -133 Silv. CC 48 40 152 Shelterwood 42 -36 -276 With a 4 percent discount rate and high prices, the light selection treatment ($153/ac) had highest ranking. The medium selection ($131/ac) and crop tree (5119/ac) followed: then came the heavy selection ($104/ac) and the silvicultural clearcut ($73/ac). Much farther behind were the commercial clearcut ($32/ac) and the diameter- 148 limit cut ($11/ac). The shelterwood treatment ranked last and had a negative value (-$9/ac) when compared to the control (Table 9-2). W Low prices and a low (2 percent) discount rate gave the medium selection ($90/ac) a clearly superior ranking over the next treatments-- the light ($75/ac) and heavy ($73/ac) selection cuts. Next came the crop tree ($58/ac), and the diameter-limit (SSS/ac). The commercial ($53/ac) and silvicultural ($48/ac) clearcuts followed. The shelterwood ranked last of all treatments again ($42/ac) (Table 9-3). High prices and the 2 percent discount rate favored the more lightly cut selection treatments. The light selection ($341/ac) and the crop tree ($281/ac) had, by far, the highest rankings under this very optimistic scenario (Table 9-3). They were followed by the medium selection ($219/ac), the silvicultural clearcut ($152/ac), and the heavy selection (Sllllac). Because the control had a residual stand with many valuable white ash, it had a higher value than the most heavily cut treatments. Accordingly, the commercial clearcut (-$133/ac), the diameter-limit (-$233/ac), and the shelterwood (-$275/ac) all had negative marginal values in this scenario. Higher real discount rates (6 percent) and a low price scenario again gave the diameter-limit cut (562/ac) the most favorable ranking (Table 9-4). The medium selection ($47/ac) and commercial clearcut ($47/ac) tied for second ranking, followed closely by the heavy selection ($46/ac), and silvicultural clearcut ($45/ac) treatments. The shelterwood treatment ($43/ac) moved from last to sixth ranking, while the light selection ($32/ac) and crop tree ($32/ac) tied for last ranking under the low price, high discount rate assumptions. With a high price scenario and a high discount rate, the medium selection treatment again had the greatest financial return ($84/ac) (Table 9-4). The light selection ($79/ac) and heavy selection ($77) had 149 a virtual tie for second ranking, followed by the crop tree ($63/ac) treatment. The diameter-limit ($59/ac) fell to fifth ranking, followed closely by the commercial ($54/ac) and silvicultural ($52/ac) clearcut treatments. The shelterwood ($46/ac) treatment was last under the high price, high discount rate scenario. In this extended analysis (1952-2022), the discount rate and price assumptions caused frequent shifts in rankings, especially between the medium selection and diameter-limit cut treatments. The diameter-limit treatment tended to be first under low price, high discount rate scenarios (three 1st rankings), while the medium selection was first under most of the medium and high price scenarios and at medium or low discount rates (four 1st rankings). The light selection only had a first ranking under the high price assumption cases with low or medium discount rates. Overall, the medium selection treatment ranked either first or second in seven of the nine price/discount rate scenarios. The diameter-limit ranked first or second in only three of the nine scenarios, while no other treatment was first in more than one case. W The treatments were tested for sensitivity to changes in road maintenance costs. The medium (base) cost was SS/ac in 1992 (1990 S) and was varied to reflect low ($3/ac) and high (37/ac) road maintenance costs. Because the selection treatments had costs applied at more frequent (lo-year) intervals, the change in their net present values was greater (53-4 range) than for the even-aged treatments where these costs were applied infrequently (Si-2 change in HPV's). The overall treatment rankings, however, were not sensitive to these changes. 150 W The extended (1952-2022) results were tested for their sensitivity to higher or lower marking and sale layout costs. Costs have a range of 20 percent for this analysis. Varying marking costs for the base (4 percent discount rate) analysis caused WPV for the selection treatments to fluctuate over a $5 range (e.g., between $75/ac and SBO/ac for the medium selection). In the medium price scenario, variation in marking costs did not change rankings unless the diameter-limit sale layout cost fell to low levels while the selection marking cost was at high levels. Here, similarly to the 1952-1992 sensitivity analysis, the diameter-limit moved to second ranking while the heavy selection fell to third. The medium selection (first) remained unchanged. Changes in marking and sale layout costs with a high price scenario did not change investment rankings. With a low price scenario, the diameter-limit (first) increased its relative performance over the medium (second) and heavy (third) selection treatments, but rankings of the first five treatments remained unchanged. The silvicultural clearcut, however, did move past the light selection treatment into sixth ranking when marking costs were at high levels. Discussion Several considerations must be kept in mind as a framework for the discussion of these results. first, all of the analyses used total economic value: that is, the value of both harvested and residual trees. Some analyses of hardwood management regimes have only accounted for the values of periodic revenues. Under such an analytical framework different results and treatment rankings may be obtained than those presented here. This author strongly believes that total economic valuation is the proper basis for forest management, particularly as scientists and managers struggle to make ecosystem management decisions. 151 secondly, these results do not reflect value differences obtained from improved or degraded tree quality. They assume that a buyer would pay the same amount for sawtimber from a poorly-managed stand (e.g., diameter-limit) as from a well-managed stand (e.g., heavy or medium selection) over the long term. However, both empirical research (Nautiyal 1982) and experience have shown that buyers do pay premiums for quality hardwood logs. A 20-year economic analysis of hardwood quality considered these value premiums explicitly (Chapter 8). Finally, these results must be considered from a long-term perspective. Porest land managed on a short-term or “cut-out and get- out' basis is not included in this analytical framework. These results apply to owners and managers who have a long-term (i.e., a generation or more) perspective for their forest management goals. W The medium selection had the highest net present value of all treatments at both 2 percent and 4 percent real discount rates. With a 6 percent discount rate, the diameter-limit out had the highest NPV, while the medium and heavy selection treatments were tied for second. These results should not be surprising to most foresters. Owners and managers with short-time horizons (e.g., high real discount rates) will tend to favor management regimes where as much timber as feasible is harvested early. This often means heavy diameter-limit cuts. With more moderate or low time preferences, landowners will be more willing to pursue management regimes balancing current revenues against future gains. Both the medium and the heavy selection treatments provided similar high total financial returns that exceeded the diameter-limit cut on all but the highest discount rate (6 percent) analyzed. The light selection treatment did not perform well at medium and high discount rates, but nearly matched the performance of the heavy 152 selection at the 2 percent rate (Table 9-1). Light selection treatments, therefore, in northern hardwood timber types are only economically feasible for landowners with very low discount rates and very long time horizons. Results for the light selection treatment are tempered by the fact that its replicates had a high initial endowment of valuable white ash and yellow birth. This comparative advantage was still not enough to give the light selection higher than a third ranking at the lower discount rate. Because of differences in the timing of cash flows, the range of returns varied by treatment across the three discount rates in the analysis. The narrowest range ($28-30) of return for 1952-92 was for the crop tree, while the greatest range was for the shelterwood treatment (-$19-25). Variation in range of returns was also high for the light selection (331-57) and diameter-limit (530-56) cuts, but moderate for the commercial clearcut (520-38) and the heavy selection (545-58). These ranges in economic returns are important. some treatments (light selection) will have a large shift in economic return given varying time preferences, while others (heavy selection) will not. The even-aged treatments (commercial clearcut and shelterwood) generally had low economic returns for the 1952-1992 analysis period. It is fair to question whether a different set of initial conditions, or a different analysis period, would have changed their investment rankings. for the 40-year period, all revenues from these cuts came at the beginning of the study, and since the stands were mostly pole-sized in 1951, returns were low. If the stands had been large sawtimber in 1951, then the commercial clearcut would have had a much higher ranking, perhaps the highest under certain conditions. It is obvious that the 40-year time horizon is not sufficient to evaluate the even-aged treatments comparably with the uneven-aged treatments. Therefore, the discussion of the 70-year analysis (1952- 2022) attempted to address this concern. 153 Winn In the extended analysis, a 30-year growth and yield projection was added to the original 40 years of data. Similar (2, 4, and 6 percent) discount rates were used, but low, medium, and high price scenarios were now included in the analysis. The 70-year analysis helped to account for the fact that at the 40—year (1992) mark, many of the even-aged stands were of pre-commercial size. Investment rankings of the 70-year base case (4 percent discount rate, medium prices) were nearly the same as for 1952-1992. The medium selection, heavy selection, and diameter-limit cuts still ranked first, second, and third, respectively (Figure 9-1, Table 9-5). The only difference was that the silvicultural clearcut was added as an eighth treatment, and it ranked number seven, ahead of the shelterwood (Tables 9-2, 9-5). Net present values for all treatments were somewhat higher than for 1952-1992. Host differences between treatments were also similar to those in 1952-1992, except that the commercial clearcut improved substantially in itsWPV, while the medium selection slightly increased the favorable gap in its performance over the heavy selection and diameter-limit treatments (Table 9-2). This result shows that investment rankings remained basically unchanged between the 40-year and 70-year horizons (Figure 9-1). A high price scenario at the 4 percent rate still favored the selection treatments, but the light selection was best, followed by the medium selection and crop tree treatments (Table 9-2, 9-5). High prices and high projected residual sawtimber volumes gave the light selection top ranking. However, the residual stands contained nearly all of this value. High prices caused the diameter-limit cut to fall greatly in NPV and in ranking. The silvicultural clearcut, however, rose in its rank (8th to 5th) and in NPV ($46 to $73). This is mostly because the silvicultural clearcut had a projected stand in year 2022 with a large, valuable white ash sawtimber component. NPV per acre (1990 S) 80 “Em-yr. returns I .40-yr. returns ' ______J 60 “C 3 , 40 "V‘W’VV‘ V" o 0’. "e a”. .{og’ O 20 O .0 V 9: Shelt CCC SivaC Light Medium Heavy Crop 8"DL Figure 9-1. Comparative economic returns, 40-year and 70-year analyses (4% discount rate) NPVs reflect differences between control and cutting treatments with medium prices. 155 real discount rate (1952-2022). I Low Prices Ranking of treatment net present values using a 4 percent Medium Prices High Prices Diameter-limit* Medium Light I Medium* Heavy Medium l Heavy Diameter-limit Crop I Commercial CC Light Heavy I Shelterwood Commercial CC* Silv CC I Light Crop* Commercial CC I Silv. CC Silv CC Diameter-limit I Cro Shelterwood Shelterwood * Shows treatments having nearly the same NPV. Low prices at a 4 percent discount rate gave the diameter-limit cut a slight edge over the medium selection treatment (Table 9-5). other treatment rankings were similar to the medium price case, but the shelterwood performed well, given the low price assumption. in performance is because low prices do not favor treatments that build up large sawtimber volumes and because the shelterwood, similar to the diameter-limit, had most of its volume removed early. We The diameter-limit out had the highest economic discount rate was raised to 6 percent, similar to the (Table 9-6). low and the medium price medium, light, and heavy selection treatments, and crop tree treatment, This top ranking for the diameter-limit scenarios. ranking when the 40-year analysis was for both the respectively, all ranked ahead of the diameter—limit (Table 9-6). Most This change Under the high prices, however, the 156 TABLI ’-6. Ranking of treatment net present values using a 6 percent real discount rate (1952-2022). Low prices Medium prices High prices I Diameter-limit Diameter-limit Medium I Commercial CC* Heavy* Light* Medium' Medium* Heavy* IHeavy Commercial CC Crop Silv. CC Silv. CC Diameter-limit Shelterwood Light Commercial CC* Light Crop Silv. CC* Cro Shelterwood Shelterwood * Shows treatments having nearly the same NPV. The commercial clearcut fluctuated from second ranking (low prices) to sixth ranking (high prices) depending on price assumptions. The even-aged treatments, in general, ranked better with the low price scenarios (Table 9-6). The shelterwood treatment was last in nearly every analysis except for the extended (1952-2022) analysis at low prices with the 6 percent discount rate. This result was surprising because the same treatment had ranked highest of all even-aged treatments in a previous analysis (Niese and Strong 1992) covering 30 years. However, the poor financial performance over the longer investment horizon is at least partly because the shelterwood was the only treatment with a non-commercial projected residual stand in 2022. Lower (2 percent) discount rates caused all of the uneven-aged selection treatments to move ahead of the even-aged treatments (Table 9-7), regardless of price scenario. The only exception was the silvicultural clearcut, which ranked fourth (ahead of the heavy selection) when the high prices were assumed at the low rate. 157 The preference for uneven-aged treatments, given low discount rates, is not surprising. Landowners with such time preferences are more willing to allow growing stock to accumulate, growing larger trees to harvest at some future date. The financial performances of three treatments, in particular, deserve mention. The medium selection was the overall best-ranked treatment at the low rate, while the shelterwood remained the worst. The silvicultural clearcut ranked highest of the even-aged treatments with either medium or high prices, despite higher costs. TABLI 9-7. Ranking of treatment net present values using a 2 percent real discount rate (1952-2022). [:Prices Medium Prices High Prices I Medium Medium Light Light* Light Crop Heavy* Heavy* Medium Crop** Crop* Silv. CC I Diameter-limit** Silv. CC** Heavy Commercial CC Commercial CC** Commercial CC Silv. CC Diameter-limit Diameter-limit Shelterwood Shelterwood Shelterwood * or ** indicates a virtual tie with nearest treatment having a similar sign. Summary Marginal benefit/cost analysis of the nine treatments in this study showed that, in general, medium or heavy selection treatments have higher total economic returns than either the more lightly-cut selection treatments or the heavily-cut even-aged treatments. The diameter-limit cut had the best financial ranking in most cases with higher discount rates or with low price assumptions. Clearcut treatments tended to rank in the lower third or middle, both in the 40-year and 70-year analyses. 158 Silvicultural clearcutting was preferable to commercial clearcutting under high price expectations. Commercial clearcutting, however, performed better with high discount rates or low prices, similar to the diameter-limit. In nearly every analysis, the shelterwood ranked last of all treatments, while in most analyses, except those with the most optimistic price/discount rate assumptions, the light selection or crop tree had the poorest financial performance of the uneven-aged treatments. These overall results confirm the findings of Chapter 8, that also showed medium and heavy selection treatments had the best economic returns when tree quality changes were included. CHIP!!! 10: CONCLUSIOI IND RICONNINDIIIONB Forest managers have known for many years that northern hardwoods can.be managed by both even-aged and uneven-aged methods. What has not always been clear is which silvicultural system is economically preferable under various conditions. Depending on their objectives, managers can use the results of this paper as a basis for making future northern hardwood management decisions. Results of this analysis indicate that under most conditions of long-term management, uneven-aged management will consistently produce higher total economic returns than even-aged management, on good (8.1. > 60) hardwood sites. In almost every case, medium and heavy selection treatments like those used in this study will generate higher economic returns than clearcutting treatments. Diameter-limit cutting can provide landowners with short-term profits, yet under many economic conditions its total economic return falls short of the best selection treatments. The medium and heavy selection treatments were first and second choices for total economic returns whenever discount rates were in the 2 to 5 percent range. At higher real discount rates (6 percent) these treatments were economically superior to the others only when owners expected high prices for their timber. Considering effects of tree quality, the heavy selection had the highest economic returns while the medium selection was second; however, the medium treatment had the greatest overall improvements in tree quality. The heavy (B-inch) diameter-limit treatment was the best of the ' heavily-cut treatments. When price expectations were low, it outperformed all other treatments at medium and high discount rates. It also was first-ranked in most cases where discount rates were high. Tree quality for diameter-limit cutting, though, was very low, with 81 percent of its trees becoming grade three quality or below-grade. 159 160 A light selection treatment had favorable economic returns only under very optimistic assumptions—-when discount rates were low and price expectations were high. In most management cases, clearcut treatments consistently had returns in the middle of the rankings. The silvicultural clearcut, however, outperformed the more common commercial clearcut whenever price expectations were high. Both the crop tree and the shelterwood treatments had consistently low total economic returns, and cannot be recommended. The study also shows that for even- and uneven-aged management, there are very tangible trade-offs between producing high economic returns and high species diversity. Managers who desire the highest tree species diversity must practice some form of even-aged management that produced only low to moderate economic returns in this study. Managers who choose medium or heavy basal area cutting to maximize economic returns will only maintain moderate species diversity. Similar trade-offs exist between even- and uneven-aged management for diversity, stand structure and aesthetics. Regeneration cuts in even-aged management enhance long-term diversity, but temporarily impair aesthetics. Selection management systems provide continuous forest cover, and favor tolerant sugar maple, producing high-quality sawlogs and sustaining the "big tree look” (Erdmann and Oberg 1973) desired by many nature lovers. Yet pursuing this aesthetic management goal is costly, because tree species diversity can be reduced. The overall results of this study show that medium to heavy basal area cuts that enhance both growth and tree quality are the economically preferable methods. Diameter-limit cutting that lowers stand quality and yields was preferred only when prices were low or discount rates high. Ho other treatments tested were comparable to these three for economic returns. These results do not reflect the reality of most northern hardwoods management today. Because most hardwood forests owners do not 161 employ professional foresters, they may use or allow inappropriate cutting methods on their property (Hiese and Vasievich 1988, Hyland 1992). Without professional forestry assistance, many landowners will make harvesting decisions with imperfect information. They may expect low prices for their hardwood resources when, in fact, prices are rising. As shown in this analysis, these low expectations may lead to excessive diameter-limit cutting or to the clearcutting of potentially high-value but small-diameter stands. Such an imperfect decision-making environment exacts high opportunity costs. However, when prices become high (as happened to oak in the 1980's, and to sugar maple in the Lake States in the early 1990's), landowners also may liquidate their hardwood resources through heavy cutting. This is particularly true for non-industrial private owners who often have high discount rates. lconomic returns to hardwood management are most sensitive to price expectations. This analysis has shown that real prices for the major northern hardwood species in the northern Lake States have increased at close to their historic (2 percent) rates over 40 years, and at considerably higher rates (3 to 11 percent) from 1970 to 1991. These higher real prices are indicators of economic scarcity for the sawtimber of most northern hardwood species. The author believes that the combination of a global economy and the lack of substitutes for fine hardwoods will cause these higher real prices to continue; such prices may become a permanent part of the forest products economy. Ho other country has a temperate hardwood resource comparable to that of the Iastern United States-~either in quantity or in quality. The current trend away from consumption of non-renewable tropical hardwoods is already focusing global attention on the United States' renewable hardwood resources. One large domestic furniture company, for example, recently changed one of its lines of fine furniture from using tropical hardwood to temperate northern hardwoods. A papular periodical 162 publicized this shift (Whodruff 1991) and cited it as an example of corporate environmental stewardship. last Asian manufacturers are also changing many of their product lines from tropical to temperate hardwood resources (Dempsey and Luppold 1992). What do these global trends and this study's results mean to the current owner of an overstocked and under-managed northern hardwood forest? Changing economic conditions generally mean that there will be good markets for northern hardwoods in the future. Low quality hardwood pulpwood and fuelwood can be thinned to provide current revenues that will release high-quality trees for faster diameter growth and quality improvement. Landowners who can afford to wait have an excellent opportunity to improve their rates of return through these carefully controlled but heavy basal area cuts. Because most cutting will continue to be of the opportunistic diameter-limit variety, those landowners who are patient may control an increasingly scarce, high- quality northern hardwood resource. Many northern hardwood landowners will continue to have short-term horizons and high real discount rates. Through ignorance and non- competitive timber sales they also may expect low prices. These owners probably will prefer some form of heavy diameter-limit cutting or commercial clearcutting. But there also will be a minority of owners who will wisely choose a longer investment horizon, who have low-to- moderate discount rates, who expect current moderate-to-high price trends to continue. For these owners, medium or heavy basal area treatments will consistently provide superior economic returns in northern hardwood management. 163 Future Research Weeds Research needs identified by this author fall into three categories. They are: 1) research on trade-offs between hardwood timber and non-timber outputs, 2) research on hardwood timber quality, and, 3) research related to the application and transfer of these results to northern hardwoods across the northern forest region. Future research is needed that specifies the non-timber opportunity costs of hardwood cutting methods. Managers need to know which wildlife habitats are lost or altered under certain cutting methods, and which cutting methods maintain or enhance habitats for unusual plant or animal species. In addition to research on the costs of foregoing non-timber outputs, managers need specific research tying economic returns of hardwood cutting methods to existing ecological land classifications. Preliminary research (Coffman et al. 1984) has shown that growth and yield of northern hardwoods vary by habitat class; however, this work needs to be refined and explicitly include economic returns. Another non-timber benefit that needs further research is species diversity. Some northern hardwood forests are highly diverse, others are nearly all sugar maple. A study could be initiated that would compare economic returns of the same cutting methods practiced on a diverse versus a homogenous northern hardwood forest. Managers might then obtain better information on the conditions where management for diversity could result in improved economic returns. Research on timber quality changes in northern hardwoods is still in its infancy because of the lack of good data. This study has shown that either of two basal area methods may be optimal, depending on whether objectives are defined primarily as economic or silvicultural in nature. This result needs to be refined through more careful observation and analysis. One question that needs to be answered is to what extent decreases in merchantable height may affect the (currently) economically optimal heavy selection treatment results. The economic 164 trade-offs and costs of maintaining specific levels of high-quality hardwood growing stock, under the varying cutting methods, also need further research. Technology transfer is needed to complete the northern hardwood economics research cycle. One way that these and other research results could be made more available to managers is through-specific northern hardwood cutting method profiles. Costs, benefits, and net yields of the most promising treatments could be summarized in short technical publications, such as Morthern Hardwood Motes, one that already reaches many managers. Finally, both the silvicultural and the economic results of this 40-year study need to be published jointly, in a general document accessible to the forestry community. APPENDIX A APPENDIX A Summaries of TWIGS Growth and Yield Projections 165 .c04uuofionm nuzoum use unsaomscni acousodo unusedsow>aqe « .msm manta us humlflsm t XHOzmmmt O O O O O O O. O mN O «mam: ban NNON av NNmn mN mama mth mama O. aha O mmn FOO ban NNON me NNmn mN mmmu mth mama m.m mom mN mmn machmm had NNON we Nnmn ON vaN hmmm mafia v.0 mma on non mzomwm Nag BHON no nab“ «N Noon omOm «mm m.h emu on Mac mmommm bod NuON «v MNwH MN vmhm vav who N.h ova me ave machmm NOH NOON an mvmn ON menu mfimm mmo m.m mma hm mmv mmOhmm hm NOON hm nee“ ma nnvn vmvN mhv m.m MN” eh «mm mmommm NO mama mm coma ma mafia hmON mom N.m Nag O mNo mucumm hm Nome mZOu. . hm OKOU . Pm . 8m . him zma mmoc OH H O M>HA maadhm EUC Edwin . so . so . om . :o m>¢ >3 5:. 53. are: ACDOHmmm OOOZOAOA mOOQZCm NNON ad acouucao .oswdmmm unslsmsns: o.mo .xsocn even came: unosm .usqusmm women seam em .smm Asqudcn Nmma answuusnoum no use» duwuwcn H “sowwsowamem modirmmuiwllilila u on cause .«Id Mundsmmt 166 O O O O O O O. O ee O mask: can NNON Nm OFHN 0N ONO— thm sea. O. OON O Ovv POO hnfl NNON Nm OhnN vN ONO— tho mean 0.0 OON cc Oev MOOhmO bud NNON mv ONON ON mHON "fine men“ m.h mm" mm vmv maOmmm Nnu FHON me vmmu hm mvoN mmhn Hum ~.h “be Oh mnm mmOamm hOn NnON we nnmu mN mFHN «man can v.o emu mm mom mmOmmm NOn bOON co more mN mama meN mm p.m Non OO~ new mmommm hm NOON vv amen ON mom“ O O ~.m mNn mm hmh szhmm No bmmfi no neon vn OOH" O O m.v Nag O mom szbmm hm Nmmn mZOF .._..m OzCU .9... . a...“ . km .55 mags OMHO m>Ha~ mOFCBm Maud Kama . so . no . a: . so ":2 2m 5:. cm... are: JCOOHmmm OOOZOJOO mOCO3~dn N .mom mauza HO ausllsm 4 x~ozmmm< N .co«ueo«~mom .Nlm xuocsmmm 167 O O O O O O O. O me O zmem< can NNON me NeON mu NNn— eanH eemN O. “as O hON BOD bun NNON me NvON ha NNnH vmon~ vemN v.O~ Hm“ OH th mmOmmm on” NNON mo Hem“ be men" mmvan hmnN s.m mo“ ma nmN mzommm Nae haON av OObu 5H mmnn meme mmha O.m va ON Hon mmommm he” NHON on menu mu comm mmnp coon n.m Ova «N uNn mmommm NOH hOON mm mvva mm v~¢~ new: mom m.h 0N: mN men mzommm hm NOON nn Nona ma Oeug hfibn Ohm m.o Nan an «en szhmm NO hmmu On HNNH Na Oom mNmN Hem H.o me O mOe mmommm hm Nmmn mz©fi firm OKOU .km .Pm .%h _=:. mzu< OMHO m>~A mOh< 2n 5:. 5:. .58: .HCDOHmm—z C8392...— wCOJZ’am NNON om nooummdo .csdammd ussflsmscsx O.mo .xoccu ovum cams: ummsm .nsdosmm mourn ovum so some usuuacn Neon accumusnoum we know usdawcu n .noaucuuunom mquOauommo ocean .=O«u00aonm :uSOum ocs unseemccme usuuoONU unusuazuq>~dn n .90: MOHSH «O husfilsm .nl‘ Mdvdommfl 4 XHOZHOQC 168 on “em h «mm moon wan n.o on n can access on” «now on one o no. amen mop o.p no u no" needed and snow nu Omm m nmn mONN Nee h.e me N no“ memmm nod N~ON on can n ~v~ ohm” emu o.m on n no" exodus «on poem on are a ecu «no and o.m an on an: «name he «com «a anon o" Hum omen mean 0. ma o no“ ecu so «oou on mono «H mam moon menu n.n pan on «ma muoaun be «com on one” «a sea «a». new m.o can an own accuse am been an spam an men «men can o.m cod 0 one arouse no «man mzoa .ah axon .ua .em .eu man are: ammo n>mn unseen uo< mama .oo .90 .on .oo m>< \rn rue «as are: neanumne nooznoaa mooazcm .nouuusnoun museum one anemones-I aroussuo unwouslloo u .msm menxa no musllsm NOON ma .:.n.o ca s>ons use xm.v noon» «an usuussao avedummm ans-suede: .xouau can» .ou: nausea: O.mm hm fl Knaznmmfl same: ususm Noun .usdusmm xenon seam .noduusnoum no ussw «seamen a .ooaueuwumsm mmqmmmumumaiumqum .eum xdvcsmmm 169 an ac. n «an omen ova 0.6: me a pm uzoauu ham "won on can a one «new . onm p.m an a on access and soon o «fin n ma" one” ncn 5.9 on a on «momma so" «non » mmu n mun soon «Nu s.» an a co exodus «on soon o «Na n ecu mac up o.o an an me «name pa «oo« No «man on «own poem one" 6. end 6 new you so «can no menu on puma mono” wean o.o nee pm «On axons: he «com me can“ an open coco ovum o.a «on an oan unease «a no»: no new: cu can" even «and n.o me” a con excuse so Noon mzoe .na expo .aa .eu .9: ran neon anon uses maaeem no: ¢< \au rue ea unseen nee «an» .au .ao .au .90 n>¢ \an ass any are: acsannux accesses anagram nomad» «oou an .:.n.v ca o>onu cc. .o.e noon» on. usuuuono .ooaaaua vacuum-an: .sOqusnoum museum one uselemsnsl asunssuo usuouellou a .mem manta no musllsm same: woman .uswusmm lemon swam «man .uoonoaa do an.» Andaman n .dONwIONAQsm auqmuduuuwaiuqqum o.mo .uovuu cues so .oe¢ uneven“ .oue «ecu-nae ¢ KHOZflamfl 170 in‘ no «on m gov nnnn com m.o on n on seeing sun -o~ «H as. m own «moN one s.o me a so axosmn «an anon no mac c Han soon omn a.» en a on nmoaun so: «Hon a men n no" ~oem now «.p an H on nuance «on room o men n on” «no or: o.o mm on an ensue so noon on we“: «a mom ~oho ommu 0. had 0 we" mmw so «oou co com“ mm mwwo vmmo when p.a men on can axonun ea «can a. when «H has: boom mMma H.a one a“ emn accuse rmw soon on he": on «an: v~oo «on: o.o on: 0 men neouun no wan: III‘ ['1 ll. mzoe .9: amoo .au .em .au mun ago: anon used unseen use m¢ \cn emu any axe: H q<=aommz noosagaa anagram “magma NOON cu .:.n.v :4 seems one um.9 sssuu Hus vsonssuu .vswummd unslensssi .dONOUsnoum museum use unslsmsnsl asuussuo unquusIIOU n .nem manta no musllsm sums: venom .nsuosfim Recon swam Nmmn .aoenonm no user Neuuunn n anodesuqnmsm gagging O.mw .MIUIN Dada so .ou¢ agenda: .ml¢ Hammond: ‘ NHOZHNN‘ 171 mm omna ma mama moao anaa m.m nma ca mom sconce paa «won an mnau ma mama mane com 0.» sea ma mam excuse aaa mama an mama «a mama mama can m.o ama mm «mm sconce boa aaoa we mama aa nooa avaa mun o.o «ma mm «mm smegma aoa moon ma umoa ma nova mam «ea m.m maa ma moo nzoaum «m «can an ooua ma unaa maa an o.m mm ma mum neoaum am mmma ma oaoa «a mam o o m.¢ am o ave umoamm mm amma mzoe .aa axon .aa .aa .aa can have mean n>an ashram . mom can» .oo .50 .on .60 n>¢ \en can aae are: seesaw»: noozaoaa neonate madam» scoz .msaammd unslsmsns: O.mm .wsmnu swam sans: noosm .usqosmm xenon seam mm .omc a-aaaaa amma .aoauoonoua no use» a-auana N .slsm meson mmdumdwummmlumqmm .nOawosnoum museum one unslemsnsl coosueaasmn a .msm manna no mnsllsm .mtd mandamus ¢ ”Nozummfi 172 ON ONNN mN vmmN hmo NNN m.m NNN NN Nov mmommm hNN NNON mN mom NN hOmN Nen mm v.0 mNN mN mam mmommm NNN NNON NN mNm mN QNmN mm NN m.m VON ON nnm mehmm NON NNON MN ONO eN NONN O O v.m mm wN men mmOhmm NON NOON me mnhN N nmN O O m.e Np MN new mmommm mm NOON Nn ONNN N mNN O O N.v em NN Ohm afloamm Nm NOON NN mam N Nm O O e.n en O mmm mmommm hm NmmN mZOH .ah OKOO .Hh .8h .9m sz mdU‘ ONNO m>Ha maatam mud ¢¢mu .60 .90 .mm .ao m>¢ \cn any any are: AdOONmmx OOOZNAON mooa3¢m madam» scoz .szNmmd unslsmsns: O.me .xsvnN seam sans: usoom .esaosmm xssz seam so .oma a-auaua amma .uoaaoonoam no use» a-auaaa N .noavsoaNmsm mmdwmdummmaludqdu .noauuenoum masonm mos unslsmsoel oooxusuNsms N .msm mauxh no husllsm .mtm xstemmm ¢ KNOZHNN£ 173 on mvna mN nmNN vaN OVN o.m ova me mom mmommm eaN NNON nn mNNN ON NNON nee mma O.m mnN mm NNm mmommm NNN NNON On ONNN «N mmma mom Nm m.m NNN Nm mum m¢Ommm mON NNON mm NmmN ON own no mN m.e mON Nn One memmm NON NOON ve ONON m mmm O O v.e Om ON Nos mmommm mm NOON nn NNNN e Noe O O m.n Np mN Nmp mmommm Nm mama eN nae v mmN O O n.n em O hem memmm mm NmmN @208 .Bh OKOU .ak .Hh .Hh TOO ”£04 OHNO m>NA mahthm ”0‘ Ktu» .OU ..OU .OO .90 n>¢ \dm ¢mfl (m8 920: ACOONmmd OOO3LAON mOOQZCm madam» scoz .szNmmd enslsmsns! O.mo .stnN euNm same: woman .ueNuemm xevna swam so .om¢ auauaaa amma .aoauoonoua no uses a-aaaua .nONaosnoua mvzoum one unslsmsnsd costustsma n .msm manta no musllsm 4 KNOZflmmd n .ooaasuaamsm mdammammmmaiuuqmm .mi¢ sacs-doc 174 ow mmOa m man mmmm mmma m.ea cm a mo «mesa maa «man u mmu a an mmmn one c. an o m eao maa «now an vnna m ame mmmma maau o.ma oaa a em mucosa maa «Non mu mmaa m men «maaa moma m.oa ooa a mm sesame «aa maoa ma aeoa e «an ammm omoa m.na cm a em «mane mOa Naou o mma a no ommu «no a. an o a you moa Naoa an ouna m men mmvua muaa m.oa aaa a co meonnm boa NaoN ow amaa v cmn momma omoa m.na ooa a mo excuse «ca moo“ «a neoa c mom «mam enoa o.na cm a we «use: mm «men s men a me mama can o. mm o m use mm noon Nn nmna m oav mmoua omam m.na maa a mm excuse mm Noam mm anaa m «me vmmoa nema a.na «ca a pm excuse am mmma cu aoOa o mac memo coma n.~a mm e um axonnn mo amma mzoe .an neon .nn .ea .en :mm axe: mean n>an «semen nor can» .oo .30 .om .oo u>< \cn can can are: Atacamaz moozmoam mooozmm NNON .NNON .NOON ca noun Nessa no menu snssvn on on s>oos Eouu .cazh .OONNmmt enslsmnns: O.me .xsmna seam sans: unosm .ssNusmm xscna suNm mm .om: a-auaua amma .aoauo-Noua no anon a-auaea N .noausuNNmsm mmammammmmmiumumu .noNausnoun muzoum fins unslsmsnsl noNaosNeu JAONN N .msm mONxH no musllsm .ONim xansmmm d NnOZflhh‘ 175 ha nooa v can mmmm mvoa m.va cm a a» «use: maa «New m man a mo mama men a. ea c w you baa «won an mama m nun maaaa «aoa «.ma moa a as excess maa anon ma maaa c own mmmoa aema m.ea mm a ma wzonum «aa mama mu coca c «an oaom omoa m.na cm n on «use: moa «aou m man a av coca non o. ma 6 m use moa aaom an naaa v nmn «mmaa anon m.ea moa a mo excuse moa «aoa mu moaa m «on amaoa Nona e.na mm a mo seamen aoa boon ma «mm m hon comm omma o.«a cm N am «can: em «mom a mma a cm oaan so. a. a" o m use am Noon an apaa o ave mamaa anon m.na aaa a mm sconce pm «com on amaa m mac vaaoa coma m.«a «ca a «ma mucosa um mmma ma mama o «me name vmca o.~a am 6 «ca neonmm mm amma mzoa .aa mmou .aa .en .ea 2mm emu: anam n>an unseen no: ¢¢ \rn can gas are: ncamauum noozmnam moonaam NNON .NNON .NOON ca noun Nsenn NO ussu susove on o» moons some .cazh .osNNmmd anslsmnns: .nOqusnonn nesonm one wnsnsmsnsl noNuusta ummNN N .mem mONTH no musllnm O.mm hm .uosua ouau .omc a-aaaaa ¢ NNOZNBG¢ N sans: usosm OON .usNoemm msonN stm .nONausnonm no use» NsNuNnN N .nONusUNNmsm qmammauuuuaimcqmm .aau¢ nave-one 176 mu mama p can mmmo mmca m.ea cm a we enema maa «wow u naN a mma vaoa mnu 0. ea c m ecu maa «Non mm onua m mum naooa mama m.ma 00a a am nuance baa «won an caaa a mom mmom mama m.va mm a um amonmm Naa mama ON NNON h mvm envm VNvN O.NN Om N am «meme mON NNON m ONN N Ov mOmN mmn O. ON O m BOO mON NNON mm enaa m vmm «noaa vama o.ea moa a an excuse pea «Non an uaaa 5 man oamm cmoa o.na mm a an ozone: «ca moon on n00a b can aoem oava a.na om N No anen< hm Noon oa can a mma comm mmv 0. mm a ma you mm «com on avna m «mo eohaa cuma a.va maa « «m excuse am Noon mm anua m «on aoeoa papa v.na mOa N cm accuse um mmma an eaaa m 5mm. emam ooma o.~a mm 6 mm excuse mo «mma mzoe .en axon .mn .en .eu can scum onao n>an unease no: «as» .90 .90 .on .90 m>¢ \rm (as mas are: Atacama: moosmnam mooozem NNON .NNON .NOON ca noun Nessa «0 none cusses OO 0» cache Eonu .caca eNOn: unosm .nONuusnoun avBOuO one unelsOsnnl nONuuste uoONN m .mem manta no husllsm O.mo mm d KNOZflmmt .xscna oaau .ome a-auaua NOON .osNNmmt enslsOsns! .usNusmm xsonN seam .noNuoenonm no nee» NsNuNnN n .noausuNNmsm O w an .aai¢ savanna: 177 ma mom a ama mmcm aava o.aa me a on «name paa aaoa m com a om amma me. o. ma o a «so baa aaoa ma maaa a pan mevaa aoma v.ca «m a an excuse . aaa aaoa ma omm a can mmmm avoa o.aa mm a an unease aaa aaoa aa one c man meme mama a.aa ma a we «sang aoa aaoa a mom a he mama we. a. an o oa you moa aaoa ma enaa m can mecca ovma m.na mm a no axoamm moa aaoa ma mom a man moom amma m.aa mm a co mzoamm aoa sooa an omm . can mace oena m.aa ma a om enema am noon 0 ama a maa mmaa man 6. na o m ecu am aooa ma aoaa o mac oamm mama m.~a mm a mm smegma am aooa ma nooa m mo. comm mpqa m.aa so a am uzoanm am amma «a mom m «a. mono aoaa o.aa ea c mm seeing so amma 2.2. .ea 9.8 .E a... .aa :3 axon mean ”53 3.23 sea in» . =0 . DO . an . 90 g‘ \Cn 5.9—. flan. 920! aaaoammm moozaaaa mooozcm aaoa .aaoa .aooa ca noun aauan «0 good cause. me on osona eoua .caze .eoaann¢ «announcer .noNausnonm autonO one anslsOsnsn noaaueNse lsNosl N .mem moNxa no musllsm O.mo hm .weonN saNm .sOm NsNaNnN ¢ KNOZflmht sts: usOsm NOON assNusnm xsonN swam .nONuosnonm no use» NsNuNnN N .noausuNNmsm O on O nu .naia uavaonaa 178 ON OOO N OON ONNO ONON N.ON ON N OO Kuhn: NNN NNON O OON N MO ONMN MOM O. NN O O _ BOO NNN NNON ON NOON n MMN OOONN OOON 0.0N NO N Om ”momma NNN NNON ON NOO M NON OONON OONN 0.0N OO N Om muchwm NNN NNON ON OOO M OON OMNO MOON 0.0N ON N Om Knahd NON NNON O ONN N OO OOON OOM O. ON O O BOO NON NNON OM ONON M MON MNNNN NOON 0.0N MO N MO udOhmm NON NNON ON NOO M MMN NNOO MNON 0.0N Om N MO Hicham NON NOON MN OOO M NMN NONO NOON O.MN ON N OO “Hahfl NO NOON O OMN N GO OOON MOM O. ON O O 990 NO NOON ON OOON O OOM OOOON OMON 0.0N OO N ON muchmm NO NOON ON ONO O OON OOOO ONON O.MN no N NN udOumm NO NOON MN OOO O NON . OOON NmMN O.MN ON O NN afloamm NO NOON OZOB .Hh OKOO .NL .Hh .Bu man 0‘ OONO m>NN OON‘BO m0¢ mdua .30 .50 .OO .50 m>¢ \tm HKON \muflmfi 8:0: NdOONOHK OOOCONOO mooattm NNON .NNON .NOON ca sous Nance uo assu.sussve ON 0» s>ons eouu .cara .osNNmmn unslsOsns: 0.00 .xeonN seam same: unOsm .usNusmm xeonN suNm mm .sOm NsNaNnN NOON anoNuosnona No use» NsNuNnN N .nONunuNNmsm a one o a .nonosnonm museum one anslsOsnsl nowaustu lsNosl N .msm meaxh no musllsm .ONIO xNonsmmm ¢ an2flmm¢ . 179 ea bmm a mom name «noa m.ma mm a am enema aaa aaoa o boa a me «man son a. ma o m aau aaa «mom on emoa a can oeoaa aoma «.ma mm a pm uzoamm eaa «non ha omm v mma mmom mama m.ma em a mm. nzoumm aaa naoa ea mam a oma mmam enea o.va mm a mm «use; moa «aoa m can a me mama mmn o. ma o a you aoa aaoa on mooa a nma onmoa mama «.ma mm a me access aoa aaoa ma amm e man mmvm mama m.va em a no nuance aoa soon «a man a mma ppm» mmna m.na mm a so «mane am «com o ama a om mama mac 0. on o a emu am noon on mmoa v awn omaoa mama m.va mm a on agenda am noon ma omm v won comm mooa o.na mm a ma nachos am amma ma emm e mom amma mmna e.aa mm c or season so amma mace .an once .9: .ea .ea an: acne mean u>aa «page» no: can» .60 .60 .am .90 u>< \rm «an ans are: aaaoamnx moosmaam mooazam aaoa .aaoa .aooa ca «on. aauan do noon nuance cm on usage eoua .cana .coaannc vacuum-ac: .noauosnonm masouO one ensssOsnnl nONuoste IsNosn M .msm neuxu no musllsm 0.00 NO .uooaa ouau .sme a-auaaa d KNOZHOO¢ sans: unocm NOON .ueNusOm xsonN swam .noavusnonm No use» NstNnN M .nONasuNNmsm ON» One s0 .ONI¢ MNQdOOh‘ 180 ma men n mnn mnmm amm n.aa on a am «mean naa anon m com o nn oonn nmn a. ea c n use naa «non nn men n non mnnn mana o.na on a no meouun naa anon on men v nmn onmo onaa n.aa no a mo nuance naa naon ma non e nan encm nnm n.aa on a an unea< nOa «aon o mma o «N «nan onn o. na o m use noa naon nu mom e mnn nvnn mvna n.aa nn a mn umonun noa «ago on mnn v man aamo ooaa n.aa mm a on «momma noa noon na mno e - mma ovum nnm m.oa on a an «meme nm noon n man a nv omen nnm c. on 0 na ecu nm «can mm «mm c onn «one evma a.na mm a nm muoann nm «can an «em v aon anon nmna n.aa an a mm neonum am nmma ma men v oan comm noaa n.aa on 0 mm acumen nm «mma mzou .ua nmoo .ea .an .eu can exec mma: n>ea manna» nor can» .8 .8 .8 .co m>¢ 2m ‘2. 5:. are: anamamuc moozmnam moonzam «non .naon .noon ca «on. among no uoou cease. .noNuusnoun museum one wnessOsnsl nONuosNee mason N .mem manta no musnlsm 0.00 NO .uucea suan .ome a-auana fl KNOZQOOC OO 0» s>onn Eouu .cana .osNNmmd unslsOsns: stnz unOsm .ssNusmm xeonN swam NOON .noNvusnouu no use» NsNuNnN N .noausoaNmsm mmammdummumimmnmm .eaie xNonsnnm 181 ma nee n can onne neaa e.ma me a an suede naa anon e nma a on nean men o. ea c n eao naa anon en ene n oen nmme emea e.ea en a on neonue naa «won an nnn n can aenn eana e.ea me a me museum naa naou ma nne n eon nave anaa e.ea me a me enema naa aaon e «an o en omen ooe o. ea c m use noa «aon en eme n men came anea n.ea en a me excuse naa «aon «a con n man «man anna n.ea me a me nuance Noa noon ma eee n «nu amme ana a.na me a me «near nm noon n acn o mu nana «me c. an o «a you nm noon en mom n aen vmee nvea o.ea an a on seamen nm noon an «on n new econ emna o.na an a an excuse nm nmma ma ane e mnn acne enoa o.na ae o «n axonme no amma e28. :5 9.8 are .a... .E zen neon mean 9a.. 3.23 use mean .ao .au .me .ao n>¢ \an ran eme are: amamaeux moozmaaa «oedema NNON .NNON .NOON ca nous Needs «0 ussu sussvn cm on s>ons noun .caoe .osNNmmm unsnsOsnsI O.mO .wsonN saNm stm: usosm .esNosmm xsonN suNm ne .om: aaaaana nmma .aoauuonoun do an.» a-auaaa N .noNusUNNmsm mmammqumuumimmmmm .noNausnoum avxouO one unsIsOsnsl noNuosts mason N .msm moNza we musllnm .NNIO wNonsmmd £ KNOZOON¢ 182 ma anaa a nna neOn eeaa e.ea me a we mean: naa «non e eaaa a nv enan ean 0. ea c m use naa «new on oona n nna neam aoea m.ea en a nv museum naa «non an eeaa n «an aemn «nna e.ea me a no vacuum naa naca ma nnoa n ann anee nnaa n.aa me a on «near noa naon e men a we enaa nan 6. ea c e ecu naa aaoa en aona n nna nene evea eo.ea en a ne vacuum noa naoa «a enaa n ova mven nena v.ea me a ne muonme «ca noon ma coca n can anee maaa e.na me a on «name nm noon m «an a ne neon emn a. ea o q ecu nm noon ea nena n ann «nee nava n.aa en a an excuse nm noon an anna n enn «man mana e.na ne a me vacuum nm nmma ma aeoa e enn maee enm n.aa me o oe seamen ne nmma ezoa .en anon .en .eu .en zen neon mean u>aa unease non «run .90 .ao .ae .ou u>< \rn can man are: mandamus moozmaam moonzce NNON .NNON .NOON :N nous Nessa no menu sumsve OO O» s>oan Eouu .caza .osNNmmm anslsOsns: .nONeusnoum nuSOuO one uncleOsnnl nONvuste mass: M .mem mONxa No NusIInm 0.00 NO .wsonN suNm .oma aaauaea t KNOZOOO¢ nNms: unOom NOON ansNosmm neonN suNm .nONuosnoum No use» NnNuNnN M .noausUNNmsm noau nous nu .eann navaonoc 183 en nooa n eon aenoa eona m.ea no a on xenon naa nnon e man o on emen nmn o. ma o m you naa nnon an nnna e oon oenna ooan n.ea ooa a me oooooo naa nnon on aoaa n non aonaa mmoa o.na em a ee oooooo naa naon en aom n men oaem eeea e.va no a ne «one: noa naon e enn a ee nnnn mon o. ma o na poo noa naon on nona e emn nooaa eeon m.ea ooa a mn ozonoo noa naon nn eooa e omn menoa maoa o.oa em a oo oooooo noa noon on nem o eon nneo oeea n.na mo n no «one: nm noon e nnn o on mann aoo o. on o na noo nm noon on enna e non emnaa oeon n.na moa n mm noonoo nm noon nn moaa e nnn nnem mena m.na om n ooa amouno nm nmma on enm m man. nnen eeea o.na oo o noa noonoo no nmma ozoe .eu oooo .na .eo .en zoo noon ooao o>aa eonoae no: one» .oo .oo .oo .oo n>< xco one one are: acooaeuo ooozmaom oooazoo nnon .naon .noon ca «one aouon no name cause. no on case .ooaanno vacuum-nu: O.mO .xsonN saNm stn: unOoO .eeNusmm xsonN saNm Nm .smm NsNUNnN NOON .nONaosnoum No uusm NsNUNnN N .nONasONNmsm gunman .nONuosnoum muxouO one unslsOsnsl ssua mouu N .msm moNxa no mus-Inn .Oaum xNonsmfim fl KNozummfi 184 NN OOO O OOO ONOO OOMN O.NN mO N OO mmhhd NNN NNON O ONN O MM MOON OOO O. ON O NN HOD NNN NNON MM OONN N OMO NNOON OONN O.NN OON N Om OOOEOO NNN NNON OM OOON O NOO MONO OOON 0.0N OO N OO OOOOOO NNN NNON NN NNO O OOO OOON NNON O.mN OO N OO Omahd NON NNON O OMN N NO MOON OOO O. ON O O BOO NON NNON MM NONN O NOO NOMON NNON 0.0N OON N OO NOOOOO NON NNON ON ONON O OOO ONOO NNON O.mN OO N ON OOOOOO NON NOON ON OOO O OMO NNON NOMN 0.0N OO N ON OOOOOO NO NOON NN NNO N ONO OOOO OOON O.MN ON N NN mmOhmO NO NOON ON OOO N NOO OMNO OOO O.NN OO O NN OOOOMO NO NOON OZON. . ab GOOD . .N...N . 8h . .NLN OOO mmoc Om N O W>N.N OOHOHO OOO ”NOON .30 .90 .OO .90 O>4 \dm OOH (OH 9202 Adan—H mam OOOSQADQ OOONZCVJ .nONwosNoun avSOuO one unsIsOsnsl ssuu mono N chem OONOH uo husllnm NNON .NNON ca nous Nonmn no comm muoovn mm on case came: unasm .mmus Nonmn 3ON ou moo NOON cN unannaou ozv 0.00 NO axeonN szm .ome aaauaoa d XNOZNOO¢ NOON N aoONNmOO unslsOsns! ansNoomm neonN suNm anONuusnoum no use» NsNuNnN .nOstONNmsm noNu Nuunsn onsum .oNin xaonsmmO 18S NN OOON O NOO NNMO MOON N.ON OO N NM amend NNN NNON O OON O NM OONN NOM O. NN O ON HOD NNN NNON NM OMNN O OMO ONONN OOON N.ON NON N OO OdOhmO NNN NNON ON ONNN O OOM NOMON OONN N.ON MO N NO OOOOOO NNN NNON ON NNON O OOM ONNO OOON N.NN OO N NO OONOO NON NNON O ONN N OO NOMN NOM O. ON O N HOD NON NNON NM OMNN O NNO NOONN OOON N.ON MON N OO OOOOOO NON NNON ON ONNN O ONM NMNON NNNN N.NN OO. N OO HOOKOO NON NOON ON OOO O NMM NMOO NMON M.ON OO N OO Omhhc NO NOON O ONN N MO NONN OOM O. ON O MN BOO NO NOON NM OONN O ONM ONONN NNON M.ON OON N OO OOOOOO NO NOON ON ONON O OOM ONMO OOON 0.0N OO N OO OOOOOO NO NOON ON OOO O OMO OOON MNMN 0.0N OO O ON HOOOOO NO NOON O28 . 9h GOOD ..N..N . uh . .N... NNOO ”OOO ONNO gNA OOHOHO n04 OOON .90 .30 .OO .90 fl>¢ \Om (OH COB N20: NOOONOOO OOOSONON mooaztm .nONwosnouO mesouO ons enslsOsnsI ssua mono M .nem moNxH No musllnm .uoua aauun 30a on one noon ca mcaccaeu oz. NNON .NNON nN sous Nsnso no ussu sussvn mm on cash .osNNmmc unsIsOsns: o.nO .xsonN szm sans: usOsm .usNusmm xsonN suNm nm .sOO NsNeNnN NOON .noNuusnoua no ussw NsNuNnN M .nONusONNOsm nmamnauummmlunumm .anun savanna: fl NNOZOOOO 186 MN OOO N NON NNOM MON O.NN OO N NO mmommm NNN NNON NN OOO O ONM NONN OOO O.NN OM N MO Odommm NNN NNON O ONM O MOO NOO OON M.ON NM N OO mOOmmm NON NNON N OON O NMO OON OO N.O ON N OO memmm NON NOON O NNN O OOM O O 0.0 ON N OO mOOEOO NO NOON M NNN M mMN O O 0.0 ON N OO mmommm NO NOON M ONN N MNN O O 0.0 ON O OO mmoamm NO NOON mach . an 5.8 . Hm . N.O . .N... mma aunt OHNO O> NA OONONO mo: OOO» .ou .oo .oo .oo use \oo one one one: AOOONOOO OOOSOOOA OOONZOO scoz aosNNOOO unsIsOsnsz O.OO .xsonN saNm sans: usoom .usNusmm xsonN suNm .noNuosnoum masouO one NO unessOsnsn uNINNIusesIsNo : .ome a-auaua t KNOZHOO< anoNuosnoum No use» NsNUNnN N .nONusuNNmsm mmdmmdmummmlflwmmu NOON o a .noe nears no nun-Ian .nnue nave-nan 187 ON ONO N OMN OOOO ONO O.MN OO N OO mOOhmm NNN NNON NN MNO M MNN OOOM OMO N.NN OO N OO mOommm NNN NNON ON MMM O NOM OOON ONM 0.0N MM N OO mOOEmm NON NNON N NON O OOO OON OO N.O ON N OO mOommm NON NOON O ONN O MOM O O O.m ON N NO mOommm NO NOON M ONN M NON O O O.N ON N NO mmommm NO NOON M OO N ONN O O N.O O O NO mOommm NO NOON Ozoa .Hh OOOO .Bh .HO .HO OOO OOOO OONO u>oa OOHOHO OOO OOO» .oo .oo .oo .oo n>¢ \co one may one: JOOONOOO OOO3ONOO OOOA3¢O scoz .osNNmO¢ unsleOsns: O.OO .xsonN suNm sans: usosm aesNosOm xsonN suNm .nONaosnouO avloum one NO vnsIsOsnsl UNINNIusasIsNo : .om¢ aaaoaua O KNOZHOOO .nONausnoum No use» NsNUNnN N .noNusuNNOsm OONmmNmmmuNIOmumm NOON o n .noo usage on nuns-so .nnue savanna: 188 ON OOO N NON OONO ONO O.NN OO o OO uuoaun NNN NNoN NN Non N OON NNon moo N.NN on 0 OO uncanu NNN NNoN a ONN O can OONN ONN o.NN Na 0 OO uuoaun NoN NNoN N oON O Non can Oo O.N ON N OO uzoauu NoN NooN O NNN O OON o o n.o ON N NO Oceans No NooN n NNN N ONN o o N.N ON N NO uzoaun Na NONN N on N ONN o o N.O a o OO uzoann No NONN uses .9; nzgu .NL .au .Nu an: u¢o¢ auNn n>Na mapsau uoO «ONN .au .50 .nn .ao n>¢ \On ¢ua :na are: a¢aaNaa¢ noozaaaa mooaz¢u NNON .NNON .NOON :N can. Nun-n no uoou ou¢9Uu OO 0» o>on¢ EONO .chH .GONNOGC unclonaulx 0.00 .uovaN OONO INna:,uao:O N.ONOOOO unvaN IUNO NO .om: N-NuNaN NNON .uoNaooNoua No use» N-NaNuN M .uONquNNaon umquuquuqqanuuquu .QON»00noua £03090 can uuolouaudl aNINNIuIOOIINv :O n .mox Ooqu no manila. .ONud xdvuoant < RHOZflhat 189 OO OOMN NN ONNN MONON NNON O.NN ONN ON NNN OOOOOO NNN NNON OO ONNN NN OONN OOOMN NOON O.NN OON O NNN OOOOOO NNN NNON OO OONN NN ONNN OMONN OONN O.NN OON O OON OOOOOO NON NNON NO NNON MN NONN OOONN NNON O.NN NON O OMN OOOOOO NON NOON OO OOON NN NONN NOOO NOON O.NN NON N OON OOOOOO NO NOON OO OOON ON NMON OOON ONON 0.0N ONN N MON OOOOOO NO NOON NO OOON NN OONN OOOO OONN 0.0N OON O OON OOOOOO NO NOON OZOH .Bh nxOO .ak .Bh .Hh ZOO Hunt aflNn n>Nq @5848“ ”0‘ m‘u» .90 .au .nn .ao n>¢ \NN mahdhm n04 zdna .DD .30 .an .50 fl>‘ \‘fl fink (OH 820: Atannmux OOOZLADO mooaxtm ocoz NGONNOOO uaOIOOauI: O.OO .uoqu ouNO ONOO: uaosm .OONOOOO nounN ouNO .uONu00noum no ado» N.NuNaN N .aONucuNNOox NO .001 NINUNlN NOON .uONuuonouO avaouO on. OOOIOOOOII Nouuuoo N .Oox OoNza «o nun-Isa .ONIO xNvuoam‘ d xHoznmmt 191 OO OONN NN NOON NMOON OOMN 0.0N NON O ONN OOOOOO NNN NNON NO NNNN NN ONON OOOMN MMNN 0.0N OON M ONN OOOOOO NNN NNON OO ONON ON NOON, OOONN OONN 0.0N NNN M NNN OOOOMO NON NNON NO ONON ON NOON NOONN NOON 0.0N OON M OMN NOOOOO NON NOON OO MOON ON OOON OONNN NOON O.MN OON M MMN OOOOOO NO NOON OO ONON NN MNMN OOOON ONNN M.MN OON N OMN OOOOOO NO NOON NO OOON ON ONNN NOOO OOON O.NN OMN O OMN HOOOOO NO NOON OZOH .9& GOOD .NL .Bh .Hh sun ndot OHNO fl>flu ODBCHO n04 z‘u» .50 .20 .nn .ao n>¢ Nan «ma ONN are: N‘DONmfld Ooozmdaa mooqxdm 0:02 NOONNOOO unaloucnl: O.OO .uovaN ouNO ONOO: NOOOO .OONOOOO noqu ouNO .u0N000noaO no an.» NINuNaN M .OONaIuNNOox No .ou¢ NONuNuN NONN .uONNOInoum avxouO can OBOIOOOQII Nouuaoo M .nox OONIH No Nhnllsn .NNI¢ quuoamd 4 xNoZHOOC APPENDIX B APPENDIX B Rog:osoion Equations usod to Calculoto bulb-r Yioldo fro. Grodod Hardwood Logs ””‘di‘ .- 1 a 192 BPPIIDI! B Rogroaaion oquationa for anon: NAPLI. Independent Variabloa Dopondont variablo : Dbhrx lumbar grad. Dbh2 Horchantablo Horohantabla height height TREE GRAD! 1 PBS -13.7 0.04397 -0.5012 0.003110 PASIF -S.1 .09560 .2343 -- Select. 8.7 -- -- -- No. 1c 64.4 -.08478 -2.S900 .009585 No. 2c -20.2 .17790 .6335 -.000416 No. 3A 17.9 -.02044 -.0217 .001842 No. 38 -49.2 .15884 1.7636 -.002061 TREE GRAD! 2 PAS -1.1 0.02014 -0.1342 0.001130 1381! -26.2 .10385 .5039 -- Select. 7.6 -- -- -- No. 1c -47.9 .23097 .2012 .002139 No. 2C -34.0 .14299 1.1578 .000284 No. 3A -9.6 .01312 .8717 .000694 No. 38 33.7 -.09204 -.3045 .003507 TREE GRADE 3 :35 3.6 -0.01538 -0.2051 0.001103 P381! -9.9 .06014 .1904 -.000541 Solocta -- .01602 -- -- No. 1C -32.4 .19398 .4660 -- No. 20 17.1 -.03144 -.8246 .006521 No. BA -40.0 .15978 1.7973 -.002380 No. 38 .4 -.01007 1.1382 .001620 193 APPENDIX I Appendix 8-2. legreeeion eguatione for nasswooofl Independent variables Dependent variable: Dbh1 x lumber grade Constant Dbh2 Merchantable Merchantable height height TRIS GRADE 1 PAS 109.2 -0.28506 -3.8748 0.013515 Selecte 46.2 -.05390 -.6958 .003807 No. 30 -94.0 .32304 1.6243 -.003752 TRI! GRADE 2 PAS 34.4 -0.14665 -1.2592 0.006328 Selecte 2.7 .02116 -.5684 .004110 No. 20 -31.4 .14176 2.4489 -.001547 No. 3c 37.8 -.06258 -l.7140 .006367 TREE GRAD! PAS -5.1 0.00443 0.1479 0.000396 Selecte 10.4 -.07589 -.3836 .003898 No. 10 -27.7 .11870 .8929 -- No. 20 -29.8 .23438 1.2059 .001412 No. 3C 14.8 -.01777 -.4513 .003582 * Theee equations also used for white aeh and American elm. APPOIldiX B- 3 e 194 APPINDII I Regression equations for YELLOW BIRCH. Independent variables Dependent variable: Dbh2 x lumber grade Constant Dbh2 Merchantable Merchantable height height TREE GRADE 1 PAS -81.6 0.25084 0.9478 -- PASlr 4.9 .06269 -- -- Selects .9 .00458 .1614 -- No. 1c 42.7 -.07439 -2.2l97 0.009403 No. 20 75.0 -.13688 -1.5430 .006463 No. 3A 29.6 -.07673 -.4814 .003595 No. 38 -l6.1 .06006 .8727 -.000553 TREE GRADE 2 PAS -31.3 0.10494 0.5159 -0.000243 3A51F -.15.9 .06511 .4257 -.000395 Selecte 3.7 .00242 .0495 -- No. 10 -52.7 .25423 1.7080 -.003374 No. 20 -5.2 .01883 .4377 .002703 No. 3A -16.0 .06754 1.1508 -.000963 No. 38 -4.8 .02837 .1856 .001751 TREE GRADE 3 PAS 0.9 -0.00197 -0.0494 0.000301 FASlP -.4 .00701 -.0677 .000735 Selects 1.3 .00229 -- -- No. 1c -35.8 .19206 .6504 -- No. 20 .9 -.06329 .4244 .005634 No. 3A —32.7 .15298 1.5680 —.002739 No. 38 .2 08238 .5935 —.001215 ”pndix 3“ e 195 APPENDIX 3 Regression equations for RED NAPLE. Independent variables Dependent variable : DbhI x lumber grade Constant Dbh2 Merchantable Merchantable height height TREE GRADE 1 PAS 107.0 -0.32016 -3.2156 0.013067 FASIF -55.7 .17503 1.2690 -.001008 Selecte 33.9 -.09531 -.6961 .002937 No. 1C -58.6 .26467 .5604 -.000106 No. 2A -59.0 .21877 1.7695 -.002140 No. 28 23.1 -- .6110 -- No. 3A -3.1 -- .5536 -- No. 38 -18.8 .02797 .4187 .000139 TREE GRADE 2 EAS 50.1 -0.19679 -1.9356 0.009389 FASIE 36.0 -.11245 -1.1578 .006010 Select.9 7.9 -.02157 -.0667 .001043 80. 1C -35.9 .21697 .4824 -.000330 No. 2A -62.8 .25951 2.5908 -.004669 No. 28 20.6 .00823 .6919 -— No. 3A -5.6 .03546 .3150 ~- No. 38 -31.3 .12511 .4-546 -.001094 TREE GRADE*3 PAS -3.6 0.03176 0.0160 -- EASlE 3.3 -.03651 -.2691 0.002968 Selecte -9.2 .03157 .3783 -.000581 No. 10 -21.0 .18789 .1210 ~- 80. 2A -17.1 .06125 .6576 .002633 No. 28 -10.6 .02622 1.6818 -- No. 3A -2.3 .04847 .1145 .000231 No. 38 -2.9 .04757 -- -- IIDLIOGRAPIY Adams, D.N.: Ek, A.R. 1974. Optimizing the management of uneven-aged forest stands. Canadian Journal of Forest Research 4(3): 274-287. Adams, D.N. 1976. A note on the interdependence of stand structure and best stocking in a selection forest. Forest Science. 22(2): 180-184. Anderson W.C. 1976. A method for appraising multiple-product sales of southern pine stumpage. Res. Pap. 80-126. New Orleans, LA: U.S. Department of Agriculture, Forest Service, Southern Forest Experiment Station. 9 p. Arbogast, C. Jr. 1957. Marking guides for northern hardwoods under the selection system. Sta. Pap. 56. St. Paul, MN: U.S. Department of Agriculture, Forest Service, Lake States Experiment Station. 20 p. Barnett, 8.J.; Norse, C. 1963. Scarcity and growth: the economics of natural resource availability. Baltimore, ND: Johns Hopkins University Press. 288 p. Belcher, D.N. 1981. The user's guide to STENS: stand and tree evaluation and modeling system. Gen. Tech. Rep. NC-70. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 49 p. Belcher, O.N. 1982. TWIGS: the woodsman's ideal growth projection system. In: Noser, J.W. Jr., ed. Microcomputers: a new tool for foresters. Proceedings of the workshop: Nay 18-20, 1982. West Lafayette, IN. SAF Publ. 82-05. Bethesda, ND: Society of American Foresters: 70-95. Belcher, D.N.: Holdaway, A.R.: Brand, G.J. 1982. A description of STENS: the stand and tree evaluation and modeling system. Gen. Tech. Rep. NC-79. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 18 p. Boyce, S.G.: Carpenter, R.D. 1968. Provisional grade specifications for hardwood growing stock trees. Res. Pap. NE-97. Upper Darby, PA: 0.8. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 12 p. Brand, G.J. 1981. GROW: a computer subroutine that projects the growth of trees in Lake States' forests. Res. Pap. NC-207. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 11 p. Brand, G.J.: Holdaway, N.R. 1989. Assessing the accuracy of TWIGS and STENS 85 volume predictions: a new approach. Northern Journal of Applied Forestry. 6(3): 109-114. Brisbin, R.L.; Dale, N.E. 1987. Estimating tree-quality potential in a managed white oak stand by Narkov Chain analysis. Canadian Journal of Forest Research. 17(1): 9-16. Brower J.E.: Ear, J.H. 1984. Field and laboratory methods for general ecology. Dubuque, IA: William C. Brown Co. 138 p. 196 197 Buell, 8. 1937. Some financial aspects of selective logging. Journal of Forestry. 35(1): 59-62. Buongiorno, J.: Nichie, 8.R. 1980. A matrix model of uneven-aged forest management. Forest Science. 26(4): 609-625. Buongiorno, J.) Young, T. 1984. Statistical appraisal of timber with an application to the Chequamegon National Forest. Northern Journal of Applied Forestry. 1(4): 72-76. Cardellichio, P.A.: Binkley, C.S. 1984. Hardwood lumber demand in the United States: 1950 to 1980. Forest Products Journal. 34(2): 15-22 e Cleland, D.T.: Scott, R.E. 1990. The U.S. Forest Service perspective on managing for biodiversity on federal lands. Nichigan Academician. 22(4): 355-363. Coffman, H.S.: Alyanak, 8.: Rotar, J.) Ferris, J.F. 1984. Field guide to habitat classification system for Upper Peninsula of Hichigan and northeast Wisconsin. Noughton, NI: Hichigan Technological University, Cooperative Research on Forest Soils. 82 p. Cooney, R. 1992. Are we mining our red oaks? In: College of Agricultural and Life Sciences Quarterly. Nadison, WI: University of Wisconsin. 11(2): 2. Crow, T.R. 1985. Diameter-limit cutting. In: Hutchinson, Jay, Northern Hardwood Notes 4.07. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 2 p. Crow, T.R.; Jacobs, R.D.; Oberg, R.R.; Tubbs, C.H. 1981. Stocking and structure for maximum growth in sugar maple selection stands. Res. Pap. NC-199. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 16 p. Dana, S.T. 1930. Timber growing and logging practice in the Northeast. Tech. Bull. 166. Washington, DC: U.S. Department of Agriculture, Forest Service. Davis, R.P. 1966. Forest management--regulation and valuation. 2d ed. New York: NcGraw-Hill. 519 p. Davis, L.S.; Johnson, N.R. 1987. Forest management. 3d ed. New York: NcGraw-Hill. 790 p. Della-Bianca, L.; Beck, D.E. 1985. Selection management in southern Appalachian hardwoods. Southern Journal of Applied Forestry. 9(3): 191-196. Dempsey, G.P.: Luppold, N. 1992. The state of hardwood markets. Northern Logger and Timber Processor. 40(9): 22-23, 30. Dennis, D.F. 1989. Trends in New Hampshire stumpage prices: a supply perspective. Northern Journal of Applied Forestry. 6(4): 189-190. Dennis, D.F.; Remington, 8.8. 1985. The influence of price expectations on forestry decisions. Northern Journal of Applied Forestry. 2(3): 81-83. 198 Dubois, N.R.) Straka, T.J.) Watson, W.F. 1991. A cost index for southern forest practices. Southern Journal of Applied Forestry. 15(3): 128-133. Duerr, W.A. 1960. Forest economics. New York, NcGraw-Hill Book Co., Inc. 579 p. Duerr, W.A.) Fedkiw, J.) Guttenberg, S. 1956. Financial maturity: a guide to profitable timber growing. Tech. Bull. 1146. Washington, DC: U.S. Dept. of Agriculture. U.S. Govt. Print. Off. 74 p. Ellefson, P.V. 1992. Forest resources policy--process, participants, and programs. New York: NcGraw-Hill. 504 p. Erdmann, G.G. 1983. Tending valuable young northern and central hardwoods in natural stands. In: Stier, Jeffrey, ed. Silviculture of established stands in North Central forests: Proceedings of a conference) 1983 September 14-16) Duluth, NN. SAF Publ. 83-11. Washington, DC: Society of American Foresters: 76-95. Erdmann, G.G. 1986. Developing quality in second growth stands. In: Nros, G.D.) Reed, D.D.) eds. The northern hardwood resource: management and potential: Proceedings of a conference) 1986 August 18-20) Houghton, NI. Noughton, NI: Nichigan Technological University: 206-222. Erdmann, G.G. 1987. Nethods of commercial thinning in even-aged northern hardwood stands. In: Nyland, R.D., ed. Nanaged northern hardwoods: Proceedings of the symposium) 1986 June 23-25) Syracuse, NY. Washington, DC: Society of American Foresters: 49-80. Erdmann, G.G.) Godman, R.N. 1981. Tending young northern hardwood stands. In: National silviculture workshop on hardwood management: Proceedings of a workshop) 1981 June 1-5) Roanoke, VA. Washington, DC: U.S. Department of Agriculture, Forest Service, Division of Timber Nanagement: 124-150. Erdmann, G.G.) Oberg, R.R. 1973. Fifteen-year results from six cutting methods in second-growth northern hardwoods. Res. Pap. NC-100. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 12 p. Erickson, N.D.) Reed, D.D.) Nroz, 6.0. 1990. Stand development and economic analysis of alternative cutting methods in northern hardwoods: 32-year results. Northern Journal of Applied Forestry. 7(4): 153-158. Eyre, F.H., ed. 1980. Forest cover types of the United States and Canada. Washington, DC: Society of American Foresters. 148 p. Eyre, F.H.) Sillgitt, W.N. 1953. Partial cuttings in northern hardwoods of the Lake States: twenty-year experimental results. Tech. Bull. 1076. Washington, DC: U.S. Dept. of Agriculture, Forest Service, Lake States Forest Experiment Station. 124 p. Gansner, D.A.) Herrick, 0.W. 1987. Using timber stand and tree characteristics to estimate rates of value change. In: Economics of eastern hardwood management: Proceedings of a conference) 1987 March 9-11) University Park, PA. University Park, PA: Pennsylvania State University: 53-60. 199 Gevorkiants, S.R. 1947. Guide for estimating defect in northern hardwoods. Journal of Forestry. 45(2): 128-130. Gevorkiants, S.R.) Olsen, L.P. 1955. Composite volume tables for timber and their application in the Lake States. Tech. Bull. 1104. Washington, DC: U.S. Department of Agriculture, Forest Service. 51 p. Gingrich, S.F. 1967. Neasuring and evaluating stocking and stand density in upland hardwood forests in the Central States. Forest Science. 13(1): 38-53. Godman, R.N. 1985. Applying the shelterwood system. In: Hutchinson, Jay) ed. Northern Hardwood Notes 2.04. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 3 p. Godman, R.R.) Books, D.J. 1971. Influence of stand density on stem quality in pole-sized northern hardwoods. Res. Pap. NC-54. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 7 p. Godman, R.N.) Erdmann, G.G. 1981. Production of quality hardwoods. In: National silvicultural workshop on hardwood management: Proceedings of the workshop) 1981 June 1-5) Roanoke, VA. Washington, DC: U.S. Department of Agriculture, Forest Service: 151-162. Godman, R.N.) Nendel, J.J. 1978. Economic values for growth and grade changes of sugar maple in the Lake States. Res. Pap. NC-155. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 16 p. Hanks, L.F. 1976. Hardwood tree grades for factory lumber. Res. Pap. NE-333. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 81 p. Hannah, P.R. 1988. The shelterwood method in northeastern forest types: a literature review. Northern Journal of Applied Forestry. 5(1): 70-77. Hannah, P.R. 1991. Regeneration of northern hardwoods in the Northeast with the shelterwood method. Northern Journal of Applied Forestry. 8(3): 99-104. Haynes, Richard W., coord. 1990. An analysis of the timber situation in the United States: 1989-2040: a technical document supporting the 1989 USDA Forest Service RPA assessment. Gen. Tech. Rep. RN-l99. Fort Collins, CO: 0.8. Department of Agriculture, Forest Service, Rocky Nountain Forest and Range Experiment Station. 268 p. Herrick, A.N. 1956. The quality index in hardwood sawtimber management. Sta. Bull. 632. Lafayette, IN: Purdue University, Agricultural Experiment Station. 26 p. Hilliker, R.L.) Webster, H.H.) Tritch, J.W. 1969. Cost relationships for evaluating major timber management opportunities in the Lake States. Journal of Forestry. 67(3): 170-174. Holdaway, N.R. 1985. Adjusting STENS growth model for Wisconsin forests. Res. Pap. NC-267. St. Paul, MN: U.S. Dept. of Agriculture, Forest Service, North Central Forest Experiment Station. 9 p. 200 Holdaway, R.R.) Brand, G.J. 1986. An evaluation of the STENS tree growth projection system. Res. Pap. NC-269. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 10 p. Holmes, T.) Bentley, W.) Broderick, S.) Hobson, T.) 1990. Hardwood stumpage trends and characteristics in Connecticut. Northern Journal of Applied Forestry. 7(1): 13-16. Hunter, N.L. Jr. 1990. Wildlife, forests and forestry: principles of managing forests for biological diversity. Englewood Cliffs, NJ: Prentice Hall. 370 p. Irland, L. 1974. Is timber scarce?--the economics of a renewable resource. Bull. No. 83. New Haven, CT: Yale University, School of Forestry and Environmental Studies. 97 p. Jacobs, R.D. 1974. Damage to northern hardwood reproduction during removal of shelterwood overstory. Journal of Forestry. 72(10): 654-656. Johnson, J.A. 1986. Evaluation of a 1978 thinning site. In: Hardwood thinning opportunities in the Lake States: Proceedings, 1984 April 20) Escanaba, NI. Gen. Tech. Rep. NC-113. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station: 69-73. Narnosky, D.F. 1986. Research needs for hardwoods in the Lake States. In: Hardwood thinning opportunities in the Lake States: Proceedings, 1984 April 20) Escanaba, NI. Gen. Tech. Report NC-113. St. Paul, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station: 136-138. Raya, I.) Buongiorno, J. 1987. Economic harvesting of uneven-aged northern hardwood stands under risk: a Markovian decision model. Forest Science. 33(4): 889-907. Kaya, I.) Buongiorno, J. 1989. A harvesting guide for uneven-aged northern hardwood stands. Northern Journal of Applied Forestry. 6(1): 9-14. Kelty, N.J.) Nyland, R.D. 1981. Regenerating Adirondack northern hardwoods by shelterwood cutting and control of deer density. Journal of Forestry. 79(1): 22-26. Kingsley, N.P.) DeBald, P.S. 1987. Hardwood timber and stumpage prices in two eastern hardwood markets: the real story. Res. Pap. NE-601. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 17 p. Rowalski, D.) Gertner, G. 1989. A validation of TWIGS for Illinois forests. Northern Journal of Applied Forestry. 6(4): 154-156. Leak, W.B., Filip, S.N. 1977. Thirty-eight years of group selection in New England northern hardwoods. Journal of Forestry. 75(10): 641-643. Leak, W.B.) Gottsacker, J.H. 1985. New approaches to uneven-age management in New England. Northern Journal of Applied Forestry. 2(1): 28-31. 201 Leak, W.B.) Solomon, D.S.) Filip, S.N. 1969. A silvicultural guide for northern hardwoods in the Northeast. Res. Pap. NE-143. Upper Darby, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 34 p. Leefers, L. 1990. Hanaging for biodiversity through public forest land management process. Hichigan Academician. 22(4): 365-370. Lemsky, A. 1971, 1981, 1991. Hardwood market report and lumber newsletter. Nemphis, TN. Northern hardwoods price lists for November, F.0.B. mills, northern area. 28 p. Lothner, D.C.) Rallio, E.) Davis, D.T. 1982. Hinnesota and Wisconsin forest product prices--a historical review, 1950-80. Duluth, NN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 114 p. In cooperation with: Arrowhead Regional Development Commission. Lyon, 0.W. 1986. Economics of all-aged hardwood stands. In: Nros, G.D.) Reed, 0.0., eds.) The northern hardwood resource: management and potential: Proceedings of a conference) 1986 August 18-20) Houghton, HI. Houghton, HI: Hichigan Technological University: 308-320. Hagurran, A.E. 1988. Ecological diversity and its measurement. Princeton, NJ: Princeton University Press. 179 p. Hanthy, H.S. 1978. Natural resource commodities--a century of statistics. Washington, DC: Resources for the Future. 240 p. Harquis, D.A. 1967. Clearcutting in northern hardwoods: results after 30 years. Res. Pap. NE-85. Upper Darby, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 13 p. Harquis, D.A.) Ernst, R.L.) Stout, S.L. 1984. Prescribing silvicultural treatments in hardwood stands of the Alleghenies. Gen. Tech. Rep. NE-96. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station: 9-10. Hartin, G.L. 1982. Investment-efficient stocking guides for all-aged northern hardwoods forests. Res. Pap. R3129. Nadison, WI: University of Wisconsin, College of Agriculture 8 Life Sciences. 12 p. NcCay, R.E.) DeBald, P.S. 1973. A probability approach to sawtimber tree-value projections. Res. Pap. NE-254. Upper Darby, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 9 p. NcLintock, T.F., tech. cons. 1987. Research priorities for eastern hardwoods. Hemphis, TN: Hardwood Research Council. 103 p. NcNeely, J.A.) Niller, R.R.) Reid, W.V.) Nittermeier, R.A.) Werner, T.8. 1990. Conserving the world's biological diversity. Washington, DC: World Resources Institute. 200 p. Hetsger, F.T.) Tubbs, C.H. 1971. Influence of cutting method on regeneration of second-growth northern hardwoods. Journal of Forestry. 69(9): 559-564. 202 Michie, R.R.) Eamets, T. 1987. Timber price reporting to non-industrial owners. Northern Journal of Applied Forestry. 4(3): 160-161. Miller, G.W. 1991. Economic residual stand structure goals for single- tree selection in central Appalachian hardwoods. In: Proceedings, 6th biennial Southern silvicultural research conference) 1990 October 30-November 1) Memphis, TN. Gen. Tech. Rep. SE-70. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southeastern Forest Experiment Station: 821-831. Miner, O.L.) Walters, N.R.) Belli, M.L. 1988. A guide to the TWIGS program for the North Central United States. Gen. Tech. Rep. NC-125. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 105 p. Moser, J.W. Jr.) Hall, 0.F. 1969. Deriving growth and yield functions for uneven-aged forest stands. Forest Science. 15(2): 183-188. Moser, J.W. Jr.) Tubbs, C.H.) Jacobs, R.D. 1979. Estimation of a growth projection system for unven-aged northern hardwoods. Journal of Forestry. 77(7): 421-423. Mroz, 0.) Reed, 0.) Prayer, W. 1990. An evaluation of bole form and microsite conditions for birdseye maple growing in the western Upper Peninsula of Hichigan. Northern Journal of Applied Forestry. 7(1): 44-45. Murphy, P.A.) Guldin, R.W. 1987. Financial maturity of trees in selection stands revisited. Res. Pap. 80-242. New Orleans, LA.) U.S. Department of Agriculture, Forest Service, Southern Forest Experiment Station. 7 p. Murray, R.L. 1989. New resources for funding intensive forest management. In: Proceedings, 17th annual hardwood symposium) May 7-10, 1989, Merrimack, WI. Memphis, TN: Hardwood Research Council: 69-83. Nautiyal, J.C. 1982. Stumpage price functions for hardwoods in the Niagara district of Ontario. Canadian Journal of Forest Research. 12(2): 210-214. Nautiyal, J.C. 1983. Towards a method of uneven-aged forest management based on the theory of financial maturity. Forest Science. 29(1): ‘7-58e Nelson, J.C.) Roach, B.A.) Frank, O.L.) Ward, W. 1975. Timber management. In: Clearcutting in Pennsylvania. State College, PA: Pennsylvania State University: 33-49. Chapter 4. Niese, J.N.) Strong, T.J. 1992. Economic and tree diversity trade-offs in a managed northern hardwood stand. Canadian Journal of Forest Research. 22(11): 1807-1813. Niese, J.N.) Vasievich, J.M. 1990. Information needs for comparing even- and uneven-aged management in mixed hardwoods. In: Hickman, C.A., ed. Southern forest economics workshop on evaluating even and all- aged timber management options for Southern forest lands: Proceedings of the workshop. 1990 March 28-30) Monroe, LA. Gen. Tech. Rep. 80-79. New Orleans, LA: U.S. Department of Agriculture, Forest Service, Southern Forest Experiment Station: 11-22. 203 Nyland, R.D. 1987. Selection system and its application to uneven-aged northern hardwoods. In: Nyland, Ralph 0., ed. Managing northern hardwoods: Proceedings of a symposium) 1986 June 23-25) Syracuse, NY. Washington, DC: Society of American Foresters: 49-80. Nyland, R.D. 1992. Exploitation and greed in Eastern hardwood forests. Journal of Forestry. 90(1): 33-37. Oliver, C.D. 1978. The development of northern red oak in mixed stands in central New England. Bull. 91. New Haven, CT: Yale University, School of Forestry and Environmental Studies. 63 p. Oliver, C.D. 1980. Even-aged development of mixed-species stands. Journal of Forestry. 78(4): 201-203. Olson, J.T.) Lundgren, A.L.) Rose, D.R. 1978. Equations for estimating stand establishment, release, and thinning costs in the Lake States. Res. Pap. NC-163. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 7 p. PARADOX Relational Database, v. 3.5 [computer program]. Copyright 1985, 1990. Scotts Valley, CA: Borland International. Patil, G.P.) Taillie, C. 1982. Diversity as a concept and its measurement. Journal of the American Statistical Association. 8(77): 548-561. Peterson, T. (various dates). Wisconsin forest products price review, stumpage editions. Madison, WI: University of Wisconsin, Cooperative Extension Service Programs. Peterson, T. 1981, 1991. Wisconsin forest products price review, lumber editions. Madison, WI: University of Wisconsin, Cooperative Extension Service Programs. Potter, N.) Christie, F. 1962. Trends in natural resource commodities: statistics of prices, output, consumption, foreign trade, and employment in the United States 1870-1957. Washington, DC: Resources for the Future. 568 p. Potter-Witter, R.L.) Lacksen, J.T. 1993. The status of the maple-birch forest type in Michigan. East Lansing, MI: Michigan State University, Agricultural Experiment Station Report RR-553. 11 p. Rails, O.N.) Smith, W.8. 1982. A simulated inventory update: Wisconsin's 1979 timber resource. Res. Pap. NC-209. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 40 p. Randall, A. 1981. Resource economics--an economic approach to natural resource and environmental policy. Columbus, OH: Grid Publishing, Inc. 415 p. Randall, A. 1988. What mainstream economists have to say about the value of biodiversity. In: Wilson, E.O., ed. Biodiversity. Washington, DC: National Academy Press: 217-223. Rast, E.D.) Sonderman, D.L.) Hilt, D.E. 1987. Recognizing hardwood quality-- key to increased profits? In: Multiple-use management of California's hardwood resources: proceedings of a symposium) 1986 Nov. 12-14) San Luis Obispo, CA. Gen. Tech. Rep. PSW-100. 204 Berkeley, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station: 273-278. Reed, 0.0.) Holmes, M.J.) Johnson, J.A. 1986. A 22-year study of stand development and financial returns in northern hardwoods. Northern Journal of Applied Forestry. 3(1): 35-38. Roach, B.A. 1974. Selection cutting and group selection. AFRI Misc. Rpt. 5. Syracuse, NY: State University of New York. 9 p. Rosen, 8. 1984. Price reporting of forest products to nonindustrial private forest owners. Journal of Forestry. 82(8): 491-495. Row, C.) Raiser, F) Sessions, J. 1981. Discount rate for long-term Forest Service investments. Journal of Forestry. 79(6): 367-369, 376. Salwasser, H. 1990. Gaining perspective: forestry for the future. Journal of Forestry. 88(11): 32-38. Sammi, J.C. 1961. deLioCourt's Method, modified. Journal of Forestry. 59(4): 294-295. Shifley, S.R. 1987. A generalized system of models forecasting Central States tree growth. Res. Pap. NC-279. St.Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Experiment Station. 10 p. Sizemore, W.R.) Sixemore, M.H.) Herrick, A.M.) Hargreaves, L.A. Jr. 1965. Estimating fair market value of southern forest lands. Athens, GA: Georgia Forest Research Council) Report No. 15. 120 p. Smith, D. 1976. The economics of hardwood plantations. In: Proceedings, Hardwood short course. Raleigh, NC: North Carolina State University, School of Forest Resources: 97-110. Smith, D.M. 1986. The practice of silviculture. 8th ed. New York: John Wiley 8 Sons. 578 p. Smith, D.M. 1987. Even-aged management: when is it appropriate and what does it reveal about stand development? In: New perspectives on silvicultural management of northern hardwoods. Gen. Tech. Rep. NE-124. Durham, NH: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 12 p. Smith, H.C.) DeBald, P.S. 1978. Economics of even-aged and uneven-aged silviculture and management in Eastern hardwoods. In: Uneven-aged silviculture and management in the United States: Proceedings of the conference) 1975 July 15-17) Morgantown, WV. Washington, DC: U.S. Department of Agriculture, Forest Service: 125-141. Smith, H.C.) Miller, 0.W. 1987. Managing Appalachian hardwood stands using four regeneration practices--34-year results. Northern Journal of Applied Forestry. 4(4): 180-185. Smith, P.M.) West, C.D. 1990. A cross-national investigation of competitive factors affecting the United States wood furniture industry. Forest Products Journal. 40(11/12): 39-48. Smith, W.B. 1983. Adjusting the STEMS regional forest growth model to improve local predictions. Res. Note NC-297. St. Paul, MN: U.S. 205 Department of Agriculture, Forest Service, North Central Forest Experiment Station. 4 p. Sonderman, D.L. 1986. Changes in stem quality on young thinned hardwoods. Res. Pap. NE-576. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 9 p. Sonderman, D.L. 1987. Stem-quality changes on young, mixed upland hardwoods after crop-tree release. Res. Pap. NE-597. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 7 p. Sonderman, D.L.) Brisbin, R.L. 1978. A quality classification system for young hardwood trees--the first step in predicting future products. Res. Pap. NE-419. Broomall, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 7 p. Spencer, J.S. 1983. Michigan's fourth forest inventory: area. Res. Bull. NC-68. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 39 p. Spencer, J.S.) Hahn, J.T. 1984. Michigan's fourth forest inventory: timber volumes and projections of timber supply. Res. Bull. NC-72. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 95 p. Spencer, J.S.) Smith, W.B.) Hahn) J.T.) Raile, O.N. 1988. Wisconsin's fourth forest inventory, 1983. Res. Bull. NC-107. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 158 p. Spurr, 8.8. 1954. Simplified computation of volume growth. Journal of Forestry. 52(12): 914-922. Stearns, F. 1986. Ecological view of the second hardwood forest and implications for the future. In: Mros, G.D.) Reed, D.D., eds. The northern hardwood resource: management and potential: Proceedings of a conference) 1986 August 18-20) Houghton, MI. Houghton, MI: Michigan Technological University: 51-66. Stier, J.C. 1990. The world eucapulp industry and its impact on the Northern U.S. pulp and paper industry. Northern Journal of Applied Forestry. 7(4): 158-163. Straka, T.J.) Watson, W.F.) DuBois, M.R. 1989. Costs and cost trends for forestry practices in the South. Forest Farmer. 27th manual ed. 48(1): 8-14. Sturos, J.A.) Thompson, M.A. 1986. Harvesting hardwood thinnings. In: Proceedings, Hardwood thinning opportunities in the Lake States) 1984 April 20) Escanaba, MI. Gen. Tech. Rep. NC-113. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station: 31-68. Thiede, G.J. 1986. Economics of thinning from the forest management standpoint. In: Hardwood thinning opportunities in the Lake States: Proceedings) 1984 April 20) Escanaba, MI. Gen. Tech. Rep. NC-113. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station: 102-107. 206 Tietenberg, T. 1988. Environmental and natural resource economics. 2d ed. Boston: Scott, Foresman and Co. 559 p. Tubbs, C.H. 1977a. Natural regeneration of northern hardwoods in the Great Lakes region. Res. Pap. NC-150. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 20 p. Tubbs, C.H. 1977b. Manager's handbook for northern hardwoods in the North Central states. Gen. Tech. Rep. NC-39. St. Paul, MN: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station. 29 p. Tubbs, C.H. 1987. Uneven-age management: when do conditions require this approach? In: New perspectives on silvicultural management of northern hardwoods. Gen. Tech. Rep. NE-124. Durham, NH: U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station: 27-30. U.S. Department of Agriculture, Forest Service. 1988. The South's fourth forest: alternatives for the future. For. Resour. Rep. 24. Washington, DC: U.S. Department of Agriculture, Forest Service. 512 p. Vasievich, J.M.) Frebis, R.) Wiethe, R.M. 1984. Quick-Silver v. 2.0: the forestry investment program [computer program). Florence, AL: The Forest Resources Systems Institute. Vasievich, J.N.) Rowalski, D.G.) Hobrla, S.L. 1988. FORSight: the timber price analysis system (version 1.0) [computer program]. Florence, AL: The Forest Resources Systems Institute. Walters, R.S.) Nyland, R.D. 1989. Clearcutting central New York northern hardwood stands. Northern Journal of Applied Forestry. 6(2): 75-78. Webster, H.H. 1986. Some strategic considerations in managing hardwood forests in the Lake States. In: Hardwood thinning opportunities in the Lake States: Proceedings) 1984 April 20) Escanaba, MI. Gen. Tech. Report NC-113. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station: 1-5. White, D.C.) Campbell, G.E. 1988. Optimal tree size for harvest in an uneven-aged stand of upland hardwoods. Rep. 88-8. Champaign, IL: University of Illinois, Department of Forestry, Illinois Agriculture Experiment Station. 30 p. Winebar, S.L.) Gunter, J.E. 1984. Costs of forest management practices in the Lake States. Res. Rep. 457. East Lansing, MI: Michigan State University, Agricultural Experiment Station. 8 p. Wonnacott, T.) Wennacott, R. 1984. Introductory statistics for business and economics. New York, John Wiley and Sons. 746 p. Woodruff, D. 1991. Herman Miller: how green is my factory. Business Week, Environment Column. September 19, 1991: 75-76. Young, R.A.) Reichenbach, M.R.) Perkuhn, F.H. 1985. PNIF management: a social-psychological study of owners in Illinois. Northern Journal of Applied Forestry. 2(3): 91-94. 207 Zillgitt, W.M.) Eyre, F.H. 1945. Perpetuation of yellow birch in the Lake States forests. Journal of Forestry. 43(9): 658-661. Zon, R.) Garver, R.D. 1930. Selective logging in the northern hardwoods of the Lake States. Tech. Bull. 164 (U.S. GPO Project no. 160-2). Washington, DC: U.S. Department of Agriculture. 46 p. "Ififlllllfifllfllfllflllfllflfiflias