mam... f — P... 5‘...“ a. in". a. .3 .nc, :. a... .. r. 3.12.; 1. ... at»??? fivum. . :1. I E... 31).? 193 £2.33. . at. 2v4inayianfi-Oa-u‘ .‘3 r§g$fimwfl \ Ii. 2 ¢ I ha... 3.3.. - 9 tblli‘.tnzfln : 2.1.1.. .5 5...: vflum. .ist; ix... .,. a 3W... a. f... ,Lb .. 1 #3.... .r .xl .izr.fl&$«|)‘ E a .1. .3. . tans“... a i. . #94114 i; .i .. 53-11:... 53.: .. 2 In... .4442: . a... y. .tl .¢ . II: a .. . 2.32%.“? Lwar‘ul a n. D: U.“ x v. .9 .: .- ula. . i. .... I II $§3§.L§£ 9.? l. {:1 1...! a. I). A .. ‘ .33.:z$n\nag.m.bia hm...“ LN}... ’33.. 7:... 1.3-(2. A?! :2!!! l .. . 5r: 1. 3.536.! «3‘ 5.5.5.5“. .(t e 2. W." huh“ . . .4. 271.4151 :11: . if: -, 41:. 3 512 . . . gram LIBRARY Michigan State University This is to certify that the thesis entitled Habitat Evaluation and Communication Strategies to Reduce Agricultural Crop Damage by White-tailed Deer presented by Kathryn B. Reis has been accepted towards fulfillment of the requirements for M.S. degreein Fish. & Wildl. Date August 7, 2000 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE IAN 2 1 2005 0311?? HABITAT EVALUATION AND COMMUNICATION STRATEGIES TO REDUCE AGRICULTURAL CROP DAMAGE BY WHITE-TAILED DEER By Kathryn B. Reis A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife 2000 ABSTRACT HABITAT EVALUATION AND COMMUNICATION STRATEGIES TO REDUCE AGRICULTURAL CROP DAMAGE BY WHITE-TAILED DEER By, Kathryn B. Reis Agricultural crop damage by white-tailed deer (Odocoileus virginianus) has occurred in localized areas throughout Michigan since the 19303. To address this issue proactively, a habitat evaluation procedure (PODD, Predictor of Deer Damage) was developed in 1998 and tested in 1999. The purpose of the PODD was to predict relative levels of deer damage to alfalfa fields and dry bean fields in central lower Michigan. In 1998, the model was reviewed by wildlife biologists and farmers, and all habitat variables previously identified as indicators of deer crop damage were sampled within 23.3 km2 areas for 8 alfalfa fields and 7 dry bean fields. Vegetation sampling results for 1998 and input from biologists and farmers were used to modify the PODD before its evaluation in 1999. A majority of the fields (73.3%) in 1998 had some crop loss and 46.7% of the sampled fields were correctly classified using the PODD’s preliminary version. In 1999, 38.5% of the fields sustained crop losses, but correct classification using the PODD’s final version occurred for only 23.1% of the fields. An interactive, workshop format also was developed to improve information exchange about crop damage issues among farmers, hunters, and wildlife biologists. The workshop format was piloted in the Upper and northern Lower Peninsulas in April 1999. Based on the organizers’ observations and questionnaire results, the workshop format can be used to improve communication among stakeholders and identify stakeholder concerns for a specific issue. ACKNOWLEDGMENTS Completion of this thesis would not have occurred without the support of several organizations and individuals. First, I thank Michigan State University, the Boone and Crockett Club, Michigan Farm Bureau, UP Whitetails Association, and The Wildlife Society for fimding my research and encouraging me to address a contentious issue with an untraditional approach. I also thank The Wildlife Society for selecting my interactive workshop as a pilot project for the Wildlife Information Network Program. The Society’s financial support helped the workshop develop as a forum in which deer management issues could be discussed objectively by often polarized stakeholder groups. Throughout my graduate experience, I received excellent guidance from the members of my advising committee. The expertise of Dr. Scott Winterstein (Department of Fisheries and Wildlife) in biological statistics played a critical role during the research design and data analysis stages of my project. Both Dr. Ben Peyton in the Department of Fisheries and Wildlife and Dr. Kirk Heinze in the Department of ANR Education and Communication Systems improved my understanding of the human dimension aspect in wildlife management and offered cogent advice as I developed my focus groups and workshops. I especially thank my major advisor, Dr. Henry Campa, ID, for believing in my knowledge and skills as a wildlife biologist and accepting me as his graduate student. His invaluable mentoring in ecological research and the graduate experience empowered me to pursue my educational goals and professional ambitions. Field-testing of the Predictor of Deer Damage would not have occurred without the cooperation of the 19 farmers in Isabella, Mecosta, and Montcalm Counties and iii research assistance of Ali Beck and Matthew Stersic. The amiable nature of each farmer promoted an enjoyable working relationship and our discussions greatly improved my understanding of agricultural production. Ali and Matt were patient, industrious workers who helped me juggle the demands of vegetation and crop sampling, but more importantly lifted my spirits with mindless games and witty songwriting. Drs. Shari Dann (Department of Fisheries and Wildlife), Larry Leefers (Department of Forestry), and Dean Beyer (Michigan Department of Natural Resources) served as key resources during the workshop development and implementation stage. Through their assistance, along with that of my advising committee, the workshop evolved into a structured but dynamic format that can be replicated for a range of human- wildlife conflicts. Thanks are extended also to the Kettuncn Center in Tustin, Michigan, Bay dc Noc College in Escanaba, Michigan, the MSU Extension Office in Roscommon County, and Michigan Department of Natural Resources (MDNR) for the use of their facilities when I conducted the workshops and focus groups. Lastly, I offer immense gratitude to the unconditional support of my family and friends. On a weekly basis, my parents, Bob and Linda Reis, remained as my closest confidants despite the vast, geographical distance between us. The refreshing humor and energetic disposition of my sister, Karen Harned, helped me regain perspective on what is most important to me — to live for the present. And my fellow graduate students in the Department of Fisheries and Wildlife and close friends around the world were always eager to lend a listening ear and help me escape the monotony of research and studying. iv TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES. . GENERAL INTRODUCTION. CHAPTER 1: Predicting Deer Crop Damage. INTRODUCTION. STUDY AREA. METHODS. Biologist focus group. Selection of farmers. . Farmer interviews. Measurements of crop loss. The PODD model. Evaluating the PODD model. . RESULTS. Deer crop damage. Habitat evaluation. Spring food. F all/winter food and security cover. . Thermal cover. Multiple stepwise regression analysis. Page viii — xi 14-16 17-19 19-23 24 -50 24-27 27-38 30 30-35 36 36-39 PODD evaluation. . Wildlife biologist focus group. Farmer interviews. Multiple stepwise regression analysis of the PODD. . Field classifications using the PODD. DISCUSSION. Observed crop losses. . Woodlot variables influencing alfalfa and dry bean losses. . Agricultural and opening variables influencing alfalfa and beanlosses. PODD evaluation. . CHAPTER 2: Communicating with Farmers and Hunters. . INTRODUCTION. METHODS. Focus groups. . Workshops. RESULTS. Focus groups. . Workshops. DISCUSSION. Focus groups. . Workshops. CONCLUSION AND RECOMMENDATIONS. Page 39-50 39-42 42-44 44-47 47-50 51-63 51 -52 52-53 54-55 55-63 64-84 65-68 69-73 69-70 70-73 74-79 74-77 77-79 80-84 80-82 83-84 85-91 LITERATURE CITED. APPENDICES. APPENDIX A — The Predictor of Deer Damage. APPENDIX B — Multiple Stepwise Regression Analysis Tables. I. Results of the multiple stepwise regression analysis (F -values and P-values) for the model containing all habitat variables influencing deer crop damage in alfalfa fields of central lower Michigan in 1998 and 1999. . II. Results of the stepwise regression analysis (F -values and P-values) for the model containing all habitat variables influencing deer crop damage in dry bean fields of central lower Michigan in 1998 and 1999. III. Results of the multiple stepwise regression analysis (F-values and P-values) for the PODD model based on its application in landscapes adjacent to alfalfa and dry bean fields in central lower Michigan in 1998.. IV. Results of the multiple stepwise regression analysis (F-values and P-values for the PODD model based on its application in landscapes adjacent to alfalfa and dry bean fields in central lower Michigan in 1999.. . APPENDIX C — Agenda for Educational Workshop. APPENDIX D — Questionnaire for Workshop Participants. . vii Page 92-96 . 97-123 . 98-108 109-112 109 110 111 112 113 - 122 123 Table LIST OF TABLES Page Climatic conditions for central lower Michigan in 1998 and 1999 based on data collected at Kent Airport in Grand Rapids (National Oceanic and Atmospheric Administration, 1998 and l999)...................10 List of habitat variables evaluated to describe deer habitat quality around alfalfa and dry bean fields of central lower Michigan in 1998 and 1999. Surveyed area was 23. 3 kmz. . . . . . . . . 20 Percent and perimeter of woodlots (coniferous and deciduous), agricultural fields, openings, and development adjacent to evaluated alfalfa and dry bean fields in central lower Michigan for 1998 and 1999. Refer to Table 5 for results of crop weight differences.................25 Percent differences in crop weight between exclosures and areas exposed to deer foraging for alfalfa and dry bean fields (field numbers 1 through 15) in central lower Michigan in 1998 and 1999. Negative values represent a crop loss. . . . . . . . 26 Mean crop weight differences in kg/ha (SE) relative to each cover type surrounding alfalfa and dry bean fields in central lower Michigan during 1998 and 1999. Negative values represent a croploss..................28 Mean values for total area (ha), mean stand size (ha), and percent of each cover type within a 23.3 km2 evaluation area in central lower Michigan in 1998 and 1999. . . . . . . . . . . 29 Mean values (SE) for habitat variables evaluated in agricultural areas to describe deer habitat attributes around in alfalfa and dry bean fields of central lower Michigan in 1998. Surveyed area was 23.3m2.................31-32 Mean values (SE) for habitat variables evaluated in agricultural areas to describe deer habitat attributes around alfalfa and dry bean fields of central lower Michigan in 1999. Surveyed area was 23.3m2.................33-34 viii Table 10 11 Page Comparison of PODD index values (V) and percentages of crop loss (P) for individual alfalfa and dry bean fields in 1998 in central lower Michigan. Negative values represent a crop loss. . . . . 48 Comparison of PODD index values (V) and percentages of crop loss (P) for individual alfalfa and dry bean fields in 1999 in central lower Michigan. Negative values represent a crop loss. . . . . 50 Comparison of public involvement techniques (Kathlene and Martin 1991, Moote and McClaran 1997, Steelman and Ascher 1997).. . . . . . . . . . . . . . . . . .66 ix LIST OF FIGURES Figure Page 1 Model of herbivore-plant interactions as shown in Morrison et al. (1998). Arrows represent negative and/or positive pathways forchange. . . . . . . . . . . . . . . . . 60 GENERAL INTRODUCTION Crop damage by white-tailed deer (Odocoileus virginianus) emerged in Michigan as a localized problem during the 19308. Agricultural landowners in Allegan County, primarily row crop producers (e. g., corn, dry beans), were the first individuals to document deer crop damage and receive a special deer harvesting permit from the Michigan Department of Natural Resources (MDNR). In 1941, the MDNR allowed Allegan landowners to bypass the lottery system and receive a deer permit (buck or antlerless), but only if they could document significant damage to row crops. During the 1950’s, the MDNR extended the program to fruit tree growers in Allegan County and in 1961 to agricultural landowners in other regions of the Lower Peninsula (Dudderar et a1. 1989) Through the 1970’s the state’s deer population continued to increase and more farmers expressed problems with depredating deer (Dudderar et al. 1989). Accordingly, the MDNR adopted its Summer Shooting Permit Program, in which only farmers documenting serious crop yield losses can participate (Nelson and Reis 1991). Because the program allows farmers to shoot antlerless deer out-of-season, the MDNR must approve all designated shooters and collect harvested deer for soup kitchen donations (D. Smith, MDNR District Wildlife Biologist, personal communication 1998). Forty-two summer shooting permits were distributed among eligible landowners in 1976. But by 1989, the estimated Michigan deer population had reached 2 million (4 times its 1971 size), despite the issuance of 1,406 summer shooting permits. The number of reported deer-vehicle collisions also increased. In 1989, there was a total of 46,784 deer-vehicle accidents, which was 4 times greater than the number of accidents reported in 1971 (approximately 11,696) (Nelson and Reis 1992). Since hunters, farmers, and neighboring landowners were not strong proponents of the summer shooting permit program, Michigan's Natural Resource Commission (NRC) decided to pilot a Block Permit Program in 1989. This program allowed farmers to purchase a group of antlerless permits if they documented a history of deer crop damage during 2 of the previous 5 years. The permits were valid during the archery and firearm seasons, could be distributed to shooters without MDNR approval, and did not count against the hunter’s bag limit for the regular season. The MDNR sold over 5,300 block permits to eligible landowners, of which 3,960 were filled (Nelson and Reis 1992). In 1990, the NRC extended the Block Permit Program as a 3-year experimental study to evaluate its effectiveness for reducing crop damage problems and improve hunter access to private lands. First year data revealed that the Block Permit Program accounted for 8% of the state’s antlerless deer harvested in 1990 (Nelson and Reis 1992). Compared to 1988 figures reported by the MDNR, an average of 33% more hunters used participating farms during the regular season, whereas non-participating farms had only 11% more hunters. Furthermore, by 1990 harvest of antlerless deer increased 138% over 1988 levels on participating farms, compared to only 75% on non-participating farms. Judged a success by the NRC, the Block Permit Program was continued after the study’s termination in 1993. Although deer crop damage has evolved into a statewide problem, only 2,000 among 51,172 farms and orchards in Michigan attributed serious economic losses to deer damage in 1989. That subgroup represented 4% of the state's private landowners (Dudderar et al. 1989). However, the economic burden of deer crop damage may depend on crop type. Campa et a1. (1997) studied Michigan’s deer — agricultural crop damage issue between 1993 and 1995 and characterized serious economic loss as a harvest loss valued above $20 per acre. The research team surveyed alfalfa (n=157), grain corn (n=246), soybean (n=106), and table bean (n=29) farmers in the Lower Peninsula and found that 20% of the alfalfa, 25% of the grain com, 30% of the soybean, and 55% of the table bean farmers had significant losses. Michigan's history of deer crop damage is not unique. The US. Department of Agriculture's National Agricultural Statistics Service (NASS) conducted surveys in 1989 and 1994 and found that farm bureaus, state agencies, and extension agents regarded deer as the main problematic wildlife species. In 1989, NASS estimated wildlife crop damage at $237 million, and over 41% of agricultural producers in northeastern and northcentral states attributed their economic losses to deer. Five years later, the estimated economic loss rose to $316 million with deer again noted as the main problem species (Fagerstone and Clay 1997). Closer to home, the Wisconsin Department of Natural Resources (WDNR) reported a record high for their white-tailed deer population in 1991 (1.35 million deer) and estimated deer cropdamage at nearly $1.5 million (Horton and Craven 1997). Horton and Craven (1997) believed this estimate may be higher since a 1984 survey implied a $36.7 million loss among farmers reporting deer crop damage. Moreover, farms with severe crop damage in the mid-19903 had deer densities that were twice as high as Deer Management Unit goals of lO-14 deer/kmz. Farmers, on average, removed <11 deer from their farms, but the WDNR predicted a 70% increase in the state’s deer population. Thus, Horton and Craven (1997) emphasized the importance of a liberal, antlerless harvest in the fall to minimize agricultural problems in the summer. Most likely, deer crop damage will remain a problem in specific regions of Michigan and other states since deer are mobile animals and agricultural production always will be a social need. Ideally, wildlife managers should address this issue proactively when managing deer populations and habitat and communicating with farmers, hunters, and other interest groups linked to Michigan’s deer crop damage problem. Therefore, my master’s research pursued 2 objectives: 1. Develop and test a habitat evaluation procedure to predict relative levels of deer damage to alfalfa and dry bean fields. 2. Develop and implement a workshop format that enhances communication among farmers, hunters, and wildlife biologists. CHAPTER 1: Predicting Deer Crop Damage INTRODUCTION The literature abounds with information detailing how white-tailed deer in the midwestem states have adjusted favorably to fiagmented, agricultural landscapes (e.g., Gladfelter 1984, Nixon et al. 1988). In Nebraska, deer residing in and near the DeSoto National Wildlife Refuge altered their home range size to reflect agricultural development and shifted the center of their home range to be closer to corn fields during the growing season (VerCauteren and Hygnstrom 1998). Similarly, during spring months and the firearrn-hunting season in Wisconsin, deer used non-irrigated crop fields near the Buena Vista Marsh for food and/or cover in proportion to the crop fields’ availability (Murphy et al. 1985). Additional studies have documented crop losses due to locally abundant deer (e.g., Palmer et a1. 1982, Vecellio et al. 1994). Deer crop damage in Michigan poses two interesting, ecological questions. Why are deer locally abundant in select regions, resulting in economic losses for some farmers but not others? What components of deer habitat entice deer in certain places to forage on agricultural crops? Between 1993 and 1996, Braun (1996) and Sitar (1996) tackled these questions as part of a multi-disciplinary study of deer crop damage in northern lower Michigan. Their investigations indicated that the movement and habitat use patterns of deer, combined with the interaction of crop field characteristics (e. g., cropping techniques, field size) and the composition and quality of deer habitat in areas adjoining agricultural fields, contributed to localized crop damage problems (Campa et al. 1997). Specifically, Sitar (1996) observed that approximately 42% of the 73 radio-collared deer in northern lower Michigan were permanent residents of agricultural areas, while 44% of the collared, migratory deer used agricultural areas as their summer range. Not surprisingly, all deer foraged in croplands and openings mostly at night, but they also used these areas more during the severe winter of 1994 than during that year’s summer. Non-migratory deer and all females used croplands and openings more frequently in comparison to migratory deer and males. Habitat features of agricultural regions influenced the deer distributions Sitar (1996) observed in northern lower Michigan. According to Braun (1996), a correlation existed between crop loss and the distance separating agricultural fields and deer habitat. A correlation also existed between crop loss and specific attributes of deer habitat quality (seasonal food sources and protection from predators and winter storms) (Braun 1996). For instance, relatively greater crop losses were observed in red kidney bean fields that were surrounded mostly by woodlots. Deer use of these bean fields was especially high if the adjacent woodlot(s) had highly preferred browse (e.g., aspen, Populus spp., and oak, Quercus spp.), a high percentage of grass and forbs for spring grazing, and a favorable wintering area to protect deer from wind and snow. Similarly, research on tart cherry trees documented higher levels of browsing when adjacent lands contained more herbaceous openings, higher quality spring foods offered by other agricultural crops, and more upland forests relative to agricultural fields (Braun 1996). These results correspond with other research findings. When surveying agricultural landowners in Ohio and Wisconsin, Stoll and Mountz (1983) and Horton and Craven (1997) respectively observed higher claims of deer damage as the percent of woodland surrounding a crop field increased. The above information suggests that wildlife biologists may improve control of locally abundant, depredating deer if they could help farmers: (I) manage deer populations; and (2) understand which habitat features may influence crop loss at varying scales, either within a geographical area or by crop type. When making management decisions, wildlife biologists could use a tool that helps them, as well as interested farmers, assess deer habitat quickly and identify the relative level of crop loss a field may sustain due to deer foraging. In pursuit of this goal, 3 objectives were developed: 1. Develop the Predictor of Deer Damage (PODD) to assess the potential of agricultural fields to sustain crop losses. 2. Evaluate the utility of the PODD with wildlife biologists and farmers. 3. Implement the PODD to assess the relative potential of alfalfa and dry bean fields to receive crop losses. STUDY AREA Field work for this project occurred in the tri-county area of Isabella, Mecosta, and Montcahn Counties in central lower Michigan between May and October 1998 and 1999. Agriculture is the dominant enterprise in this region and is supported by fertile soil conditions and a favorable climate. Corn, hay, dry beans, wheat, oats, and potatoes represent the principal crops (Kerr and Trull 1928, Schneider 1960, and Corder 1984). Gently rolling plains with a sandy clay soil typify Isabella’s western belt, but valleys with steep walls along rivers and prominent, isolated hills occasionally occur. Additionally, streams are bounded by short, steep slopes, and adjoining depressions regularly collect and hold drainage from surrounding areas to form lakes, ponds, or muck areas (Kerr and Trull 1928). In Mecosta County, undulating to steep moraines with a loamy sand base dominate the landscape although flat to gently rolling outwash plains occur along major rivers. Numerous lakes and streams also exist (Corder 1984). The majority of Montcalm County exists as nearly level to undulating outwash plains and bears a loamy sand base (Schneider 1960). According to weather data at Kent Airport in Grand Rapids, both years of the study featured a relatively mild winter although twice as much snow fell in 1999 as did in 1998 (National Oceanic and Atmospheric Administration 1998 and 1999). In 1998, the average daily temperature during the winter was 07°C with a total snowfall of 107.4 cm whereas in 1999 the average daily temperature was —1.9°C with a total snowfall of 222 cm (Table 1). During the spring and summer, the average daily temperature was 19.7°C amt mod 3.3 3.3 $6 mo.m we. Wm- ad ad 89888 o c o 36 Sam Ed N._ v." 5: wd 89552 o o o c 3N SK QM we v.2 Q3 c3800 o c o o m _ .w de NS 5.3 QMN fimu £38m o o o o w _ .m Nmé 03 ”3 m.m~ wsm «mama? o c o o N2. 36 _.w_ v.2 van Own 32. o o o o wed end w.3 Wm fl men can 0:2. o o o o mmd N56 wd v; _ mdu vém ~32 o o c o q: 23 an 2 N2 3: E3. v~.m_ cw.- 8.0m moan Sam ~m.~_ 31 m. 7 ad fie £0.32 No.5 o _ .2 mmdm 54 E...“ Ed we. 9N- m.m wé sebum wdm dem Qw: on. _ m 3.x and— vd- Qm- ed- NA 5:3 33 :2 32 wee 23— 33 33 wmfl 33 wag 5me seem. £3.55»: -§xe=m Efieéaum 83% E335: $25 :8: $35.38: $25 :8: 5.52 Es sausage G. V 588:3 .82: e5 was €533.55. ofiaeo§< e5 cease .88an $52 2.5 a 8&2 seem a 38:8 5% 8 Bag 32 2a Mae a 5322 532 38° 3.. 8888 2330 ._ 23. 10 in 1998 and 193°C in 1999 yet substantially more rain fell in 1999 (56.6 cm) than in 1998 (34.0 cm). Prior to European settlement, the tri-county area was heavily forested with hardwoods on heavier soils and pines (Pinus spp.) on sandy soils. Virgin hardwood stands contained mixes of either sugar maple (Acer saccharum), beech (Fagus grandifolia), and yellow birch (Betula alleghaniensis) or oak (Quercus spp.) and hickory (Carya spp.). Stands of white and red pine (Pinus strobus and P. resinosa) with oak and aspen species (Populus spp.) also were common. However, after the extensive timber harvesting and burns of the late 1800’s through the 1920’s, the forests of Isabella, Mecosta, and Montcalm Counties assumed a different species composition. Aspen and red maple (Acer rubrum) now join sugar maple-beech/birch stands, and aspen, for the most part, has replaced hickory in the oak-hickory stands. Reproduction among white and red pines is low, thereby allowing for secondary growth of oak, aspen, and red maple. Finally, on poorly drained soil elm (Ulmus spp.), ash species (Fraxinus spp.), red maple, and swamp conifers proliferate (Kerr and Trull 1928, Schneider 1960, and Corder 1984). Throughout the tri-county area, estimated deer density ranged mostly between 13.5 and 19.3 deer per kmz. However, in southwest Isabella, pockets of 31 deer per km2 are not uncommon (J. Greene and D. Reeves, MDNR Wildlife Biologists, personal communication 2000). 11 METHODS Completion of this project involved 3 stages - development, evaluation, and testing. Development of the PODD occurred in 1998 based on a literature review of deer ecology. MDNR Wildlife Biologists and farmers evaluated the assessment procedure in 1998, and their input was used to modify the PODD before testing occurred in 1999. Modifications of the PODD also were based on the suitability of deer habitat in central lower Michigan. The habitat was quantified in 1998 (and 1999). When testing the PODD’s ability to classify crop fields correctly (i.e., predict relative levels of deer damage), the PODD was implemented around each alfalfa and dry bean field that was sampled in 1999. The predicted results were then compared to observed crop losses. Since the PODD was developed using the habitat variables recognized as influencing alfalfa hay and red kidney bean losses in the northern Lower Peninsula in 1994 to 1995 (Braun 1996), the same crop types were sampled in central lower Michigan. Both crop types sustain relatively high levels of crop damage by white-tailed deer (Dudderar et al. 1989). Biologist focus group Nine MDNR, District Wildlife Biologists were invited by phone to participate in a focus group discussion regarding their managerial experience with deer crop damage. Collectively, the invitees represented all regions of Michigan. The full day discussion occurred in June 1998 and addressed: 1. Distinctions between injury and damage for given cash cr0p. 2. How biologists currently assess deer crop damage within their management districts. 12 3. Training procedure for assistants. 4. Level of interaction or cooperation among MDNR biologists and with agricultural agencies/ groups. 5. Time of the year farmers seek assistance. 6. Receptivity of farmers to suggestions on farming practices that benefit wildlife and/or alleviates wildlife damage problems. 7. Changing patterns of the agricultural landscape due to farm sales. 8. Applicability of a standardized approach to deer crop damage assessments. Selection of farmers During separate meetings with 2 MDNR biologists in central lower Michigan, 31 agricultural landowners were identified as potential partners in this study. All farmers produced either alfalfa or dry beans, participated in the Block Permit Program for 2 2 years, and received 2 20 block permits each year. Initially, the third criterion stipulated 2 4O block permits to indicate relatively serious problems with depredating deer, yet a minimum of 20 block permits was later adopted to minimize driving distance between participating farms. In May 1998, letters explaining the purpose of the study were mailed to each farmer, and farm visits were conducted in late May both years to learn about the farmers’ experiences with deer crop damage and their willingness to participate in the study. All farm visits occurred before the farmers completed their first alfalfa harvest and planted fields with dry beans. As a result of these meetings, partnerships were forged with 15 landowners in 1998 and 10 in 1999. Collectively, these landowners represented 14 different townships in the tri-county area of Isabella, Mecosta, and Montcalm. One crop 13 field per farm was selected each year, but in 1999, 2 crop fields (1 alfalfa and 1 dry bean) were selected per farm for 3 landowners. Thus a total of 8 alfalfa and 7 dry bean fields were used in 1998, and 6 alfalfa and 7 dry bean fields were used in 1999. Individual crop fields were selected randomly. Farmer interviews Interviews were scheduled with 14 out of 15 farmers in 1998 and with each of the 4 new landowners participating in the study for 1999. During these interviews, the farmers discussed: 1. How they distinguish between deer crop injury and damage. 2. The agency(s) with which they work to control deer crop damage problems. 3. Time of year they seek assistance. 4. The community’s perception of the deer crop damage issue. 5. Explanation of and reactions to the PODD, including the farmers’ willingness to learn how to use the procedure on their own if the research concludes the PODD is a useful management tool. 6. Other ideas for addressing the deer crop damage issues. 7. Questions on page 1 of the PODD (See Appendix A). Measurements of crop loss Fields used to assess possible agricultural crop damage by deer were adjacent to woodlots, openings, other agricultural fields, and/or developed areas (i.e., home, road, assorted farm buildings). To measure relative levels of crop loss, the installation of deer exclosures (2.0 m x 2.0 m x 1.5 m) was stratified by cover type, and the proportion of an adjacent cover type was identified with aerial photographs. If a field had a maximum 14 length and width > 180 m, it had a core area that would receive minimal or no deer damage since deer generally forage within 90 m of a field’s edge (Bender and Haufler 1987). All exclosures were installed within this 90 m zone, and exclosure sites were identified by a random number of paces from the field’s edge. If no core area existed, however, a random number of paces along the cover type’s edge also helped identify exclosure sites. For control areas exposed to deer browsing, sampling plots were located 5 m away from the exclosure towards the field’s edge. Depending on the size of the crop field, 4 to 7 deer exclosures were installed in each field. The exclosures were constructed with four 1.5 m steel t-posts and 8.0 m of Tenax C-flex polypropylene fencing (Construction Supply, Inc., Highland, Michigan) that had either a 5.63 cm x 6.89 cm or 5.08 cm x 2.54 cm mesh opening (Braun 1996). Fields < 12.1 ha received 4 exclosures in 1998 and 5 in 1999; fields that were 12.1 to 24.3 ha received 5 exclosures in 1998 and 6 in 1999; and fields > 24 ha received 6 fences in 1998 and 7 in 1999. Exclosures were installed within 4 days of an alfalfa cutting or dry bean planting and removed temporarily when farmers cultivated their dry bean fields for weed control purposes (reinstalled within 24 hours if not immediately after cultivation). Paired, 1 m2 crop samples were collected within 4 days prior to an alfalfa cutting or dry bean harvest. Alfalfa samples were collected for the farmers’ second and third cuttings since the study began at the time of the first alfalfa cutting each year. The second cutting occurred between 1 and 20 July in 1998 and 25 June and 25 July in 1999. The third cutting occurred between 13 August and 15 September in 1998 and 18 and 31 August in 1999. However, in 1999, higher rainfall levels in spring and early summer 15 (Table 1) produced more alfalfa hay than the farmers could sell. Thus, only 2 farmers had performed a third cutting in late August. Dry bean samples were collected 8 to 26 September in 1998 and 13 September to 16 October in 1999. To collect alfalfa samples, alfalfa stems were cut with gardening shears 10 cm from the ground, the height at which the farmers harvested the fields. Samples were placed in paper bags and stored until oven-drying began in September. Dry bean samples included only the bean pods, which were later shelled and stored in paper bags. For both crop types, samples were oven-dried at 378°C for 48 hours to obtain their 100% dry weight in kg/ha. Paired samples were pooled by adjacent cover type across all fields to represent varying levels of deer density (Braun 1996 and Sitar 1996). Paired t-tests (or = 0.10) (Sokal and Rohlf 1981) were used to identify significant weight differences between exclosures and areas exposed to deer foraging. The percent difference between paired samples (exposed sample — exclosure sample) also was found for each evaluated field. Differences of -1 to -10% represented relatively low levels of crop loss, -11 to -20% differences represented relatively medium levels of crop loss, and > -20% differences represented relatively high levels of crop loss. Positive differences, of course, indicated no crop loss. These categories were based on the survey responses of 790 Michigan farmers regarding their attitudes to deer depredation (Fritzell 1998). Fritzell (1998) found that among 50% of the respondents, crop losses of approximately 4% were deemed problematic but tolerable whereas losses of approximately 11% were considered intolerable. These ranges of percent difference were used to evaluate the accuracy of the PODD. l6 The PODD model Initial development of the PODD was based on the results of Braun’s (1996) and Sitar’s (1996) deer crop damage research in northern lower Michigan and the white-tailed deer Habitat Suitability Index model for the Upper Great Lakes Region (Bender and Haufler 1987). The first draft contained all of the habitat variables that Braun (1996) identified as significant indicators (P 1 m in height) and sapling stems per ha; (2) oak diarneter-at-breast height (DBH) using a biltrnore stick (1998) and DBH tape (1999); and (3) basal area for oak and coniferous species. Basal area was calculated using the DBH measurements of S 10 trees, and only trees with a DBH 2 10 cm were counted. Only canopy cover and basal area were stratified by type of woodlot (coniferous versus deciduous) to highlight the thermal cover attributes for deer habitat. The number of sampling plots for each woodlot and opening depended on the variability of shrubs for each cover type (Higgins et al. 1994). In general, 5 plots were sampled in each hardwood stand, and 3 plots were sampled in each coniferous stand and opening. However, more plots were used if the number of shrubs plotted against the number of vegetation plots sampled did not plateau by the fifth or third sampling area. All plots were located randomly within the sampled woodlots and openings. Pastures were recognized as agricultural fields and, therefore, not sampled. Because vegetation sampling within openings was initially overlooked, designated openings that contained no shrubs or saplings were presumed to have 100% availability of spring forage (this does not include clearings within sampled woodlots). Horizontal coverage, however, was measured for all openings. Measurements for all crop field characteristics, including distance to conifer stands and percent fall/winter food and security cover, were measured from aerial photographs provided by the Farm Service Agency’s county offices. Percent fall/winter 21 food represented openings and wooded areas whereas security cover represented only wooded areas. These photographs were taken in 1992 for Mecosta and Montcalm Counties and in 1988 for Isabella County. However, the only discrepancies observed between the photographs and field observations were associated with agricultural practices (e.g., fields turned fallow or fields no longer under a conservation easement). Vegetation sampling was necessary for 2 reasons. First, it created a data base of vegetation characteristics specific to the study area. In 1998, this information was used to reconfigure PODD categories based on potential ecological conditions for each listed habitat variable. Secondly, the sampling results were used for a multiple stepwise regression analysis (a = 0.15) (SAS System, Cary, North Carolina) to identify which habitat variables significantly influenced deer crop damage for alfalfa and dry bean fields in the study’s tri-county area (Sokal and Rohlf 1981). A second stepwise regression (or = 0.15) was performed using variables found in the PODD model. Results of both regression analyses were compared to determine which, if any, attributes of habitat are common to both deer crop damage models. In 1999, crop data for the second and third alfalfa cuttings were combined to describe deer damage since only 2 farmers performed a late summer harvest. Regarding modifications of the PODD after the first field season (1998), the description of understory vegetation in woodlots and the type of openings found next to a crop field is more specific in the final version. In the first draft, the percent of vegetation found in a woodlot’s understory or an opening and the type of identified vegetation (i.e., grass/forbs and shrubs) were recorded separately. In the final version, however, the wildlife biologist using the PODD must identify the percent of each vegetation type 22 identified when evaluating an opening or the understory of a woodlot. Consequently, the final version has 4 descriptive categories instead of 3 for these 2 variables. Other modifications of the PODD were focused on landscape variables. First, in the original draft the distance between each cover type and crop field had only 2 descriptive categories: < 180 m or 2 180 m. In the final version, 4 categories were created to describe the distance variable: 5 10 m, 11 to 25, 26 to 50, and > 50 m. The descriptive categories for percentage of woodlots adjoining the evaluated crop field and within the 23.3 km2 evaluation area were slightly modified (e.g., instead of 25 to 50% the second category is now 30 to 60%). These modifications created a better alignment between the PODD and the Habitat Suitability Index for white-tailed deer developed by Bender and Haufler (1987). Lastly, because very few crop fields evaluated in 1998 adjoined a conifer stand, the variable “number of conifers within the evaluation area” was added to the PODD’s final version to insure wildlife biologists can identify at least one attribute of thermal cover for deer. For a majority of the habitat variables, the index values for descriptive categories were altered to reflect more appropriately the ecological relationship between deer and a specific habitat component. These value changes were based solely on deer ecology literature. 23 RESULTS Deer crop damage Most crop fields in the study area were relatively small to medium in size (8 to 14 ha) and had fairly high problems with depredating deer (Table 3). On average, 45 block permits were issued to field owners during the regular hunting season in 1997. In 1998, a majority (73.3%) of all sampled fields had crop losses attributed to deer (Table 4): 42.9% of the alfalfa fields for Harvest 2, 62.5% of the alfalfa fields for Harvest 3, and 71.4% of the dry bean fields. Yet only 1 alfalfa field during the second harvest, 2 alfalfa fields during the third harvest, and 2 dry bean fields incurred relatively high crop losses that ranged between -28 and -48% (Table 4). The percentage of observed losses among all crop fields dropped substantially in 1999 to 38.5%. Relatively high levels of crop loss were observed for only 1 of 2 sampled alfalfa fields during the third harvest (-30%) as well as 1 dry bean field (-25%). No crop losses were observed among alfalfa fields in 1999 during the second harvest (Table 4). However, each year the largest percentage of crop loss was observed among dry bean fields; 71.4% of the fields in 1998 and 42.3% in 1999. Because a third alfalfa cutting occurred for only 2 out of 7 alfalfa fields in 1999 due to the spring and early summer’s wet conditions, its crop loss observations were not included in this crop type comparison. In 1998, the evaluated alfalfa and dry bean fields adjoined mostly large woodlots (perimeter approximately 8,000 m) and other agricultural fields (Table 3). In the following year, the landscape surrounding evaluated crop fields was more varied. Agricultural fields and developed sites comprised the main cover types adjoining alfalfa 24 92.2 a 95.3: 3.35 -- $38 3.8: :55. - E 882:8 Ed 2 SC Get Ex: 4.8 2: on 3». - 388m mung $23 an. as 5.89 -- 23a $.82 “22 - «é reassess 3.5 am. a 83: $5.3 gm 3. S 2 - amuse gay... 82 an: a 225 @633 -- 93.: Son: ”28» - as essence 39$ £9 a; $5 ....3 2 E4 3:“ .828 “avg 3.38 SEN- $5.35 -- 3.2a $.22 $.88 - E immerse awe 33V 63: A2 .3 a: as men a? - :82 3% «a: 68 GE £8 £9 EoEQBoSQ wficono _83_:o_.&< $2953 no.5 80> accuses Ems; gee ease HO 2.32 é m 2%... a £3. .82 9a :2 as 5322 .532 3on 5 man :39 be 98 «rate Baa—go 8 uncommon “55%—26v 98 53:88 £23 5839:? .Amaozfloov 28 980353 32303 we 835.com v5 Enos; .m 39$. 25 m5mfi£ Bow 8 some 398 was 825288 50253 Emmi, no.8 E unaccoEE $3.. v 8 magmas.“ Bow 8 none 283 wen 858—38 :8an 8388 new Emma? not. 5 oodoBbE $8. 3 2- :— mamas.“ Bow 8 none 88¢ 98 magma—axe 502509 8383 new Ems? mob E ooeoSbmv .xbm- A "a 832223.82 5.x: 5.2 22.2- $5.2- 3.x.” 63.3 6.22 6.2: $.22 3.x... 5.»? 33.2: @322 232...- 3.5.2.2- €me 63.8 5.2.8 6v 8% as .\.w~ A__V.am~- 2v $3. a: .\.2- 5 .\.w~- 85:21 682. $22. $32. $.82... 5.2V? .3qu .DQ m .83: 52.2.. m first .22... 82 33% 2Q m ES: .22.. E .\._N E .2: 5 .2- 5 .2 2- 6 .22- N .85: 52.2.. $2 commaum nomemQ $3225 33 .5282 .2: 8.0 so.» .32 mob a Bowen».— .8? .8202 .22 E. 2.2 s 5322 .252 32.8 a 5 .253. 2 888.... 2.8 .28 9.3 be new £33 new wfiwfla but 8 @8093 288 can 8.3835 5053 «ammo? mob E 8058b? Beacon .v 033. 26 fields but a more even distribution of woodlots, agricultural fields, openings, and development surrounded dry bean fields. For both years, openings comprised the smallest amount of adjacent land for each crop type (<11%) (Table 3). Deer depredation was most common in field sections adjoining woodlots and openings for both crop types (Table 5). Yet, it was only in 1998 that statistically significant crop losses occurred (P S 0.05 for alfalfa’s third harvest and dry bean; P _<_ 0.10 for alfalfa’s second harvest). Although crop losses relative to openings were observed each year, these losses were not statistically significant (P 2 0.10). Other non- significant crop losses were observed in sections of: (l) alfalfa fields adjoining developed areas during the third harvest in 1998; (2) alfalfa fields adjacent to other agricultural fields in 1999; and (3) dry bean fields next to developed areas in 1999. In several instances, crop production was significantly different (P S 0.10) between paired samples but reflected more crop production outside the exclosure (Table 5). This was observed among alfalfa fields next to agricultural fields (third cutting) in 1998 (P S 0.10) and woodlots (P S 0.10) in 1999 and dry bean fields adjacent to other agricultural fields (P S 0.05) in 1999. Habitat evaluation Hardwood stands, agricultural fields, and developed areas comprised a majority of the evaluation area (23.3 10112) around alfalfa and dry bean fields evaluated in central lower Michigan (Table 6). Dry bean fields averaged slightly more agricultural land per evaluation area (33.8 to 35.9%) than alfalfa fields (27%) whereas alfalfa fields averaged slightly more developed [and (31.7 to 35.2%) compared to dry bean fields (23%). The size of individual agricultural fields and openings in each evaluation area ranged between 27 .wamwfiom Hoot 2 3898 30.8 93 858—05 5233 smog-“ 69:3 .2 .owmv moococobmc agomwhwmm... .wfiwfiou Hoot 8 some muons was 85835 5050: O83 3:3 .36me mooaoHo-ce “533mm: 2&8: $3.2 332 A38: 33. 2%- :23: one... 2.8m on 23.25 2 SE 20% 28.3 m a m 285: 3w. 6 m2- 3.:- bnfi gs? aaa~ an. 3 G33 :39 $5.2 : 2.3 5.2 _- 8.2 .338- 23% be $32 82 2 c 638 2.3 m .35: 8.2. w. SN- .32. :2 :- as? :32 23.2 9.32 9&3 N acre: 3.. 2.3- mg .35.- 62...? aaa~ Ammo 62 62 Ammo EoEgEQSQ wannO _§~=otw< 8:503 no.5 30> .32 nob a Eomoaou mos—a.» 962.2 .22 2.. .32 was. 3222 .252 was. 5 flow 8.5 .5 2. 2.2. gages... cab B>oo some 3 gum—o.— mcobo Ewen-8m 98 59 atmo— E abacus-amt “ammo? nob :32 .m 033- 28 am: am: Gm: 8m: 8:: SN 3 a: 0.2 2 - “822 s: 32: 8:: $2: -- 3 2: e. :V 3” - E: as 2% GEN: 55: 89:: 8:: 33: 03% NE «.5 ed: :3 - E: 32 88m a5 A8,: A3. : 5...: an: 2 2: n. _ m 3 EN 92 3 - “:82 39: $0: 5.2: 5.8 -- v3 2 3m 3 - 3: Ba 2% 23m: $3: and: $33: 32: 32. 5.2 $8 $2 a: - E: 32 QB? .33 33.: 8a.: 5.: 2 n: 83: 08 3 man 3% 3 - Boos: El: 89: an: 2 2: -- q: 9: NS 3 - E: as 25... GS: : Ea: :3: 5.5 8?: 3% 3.: 3M: 28 c3 - E: 33 38m a5 23.: 93.: £3: 3.: 82: NR 2 EN SN 2 - “Bea a2: 32: GE: 8?: -- ”.8 we 5% S - E: as “as... 8.3: 93.3: 5.8: 3.2: 53: 3% m2 3.3 38 98 - E: 39 gas..- 83 am: am: am: am: am: 550 wafinof 3330:»... 33m Bose“: €55 £28 £5 mam 53> .32 98 :3 E 5w£22 832 35:3 E 33 song—«>0 ”.5— m.m~ a 565 25 350 some .8 “coupon :5 A5: 3mm 2.86 :38 .35 8.3 _88 Ba 839» :32 .o 033- 29 7 and 11 ha. Individual hardwood stands ranged between 38 and 63 ha, but the average size of conifer stands each year was substantially smaller (< 7 ha). Spring food Sources of spring food included herbaceous plants, seedlings, and shrubs < 1m in height. Not surprisingly, openings had higher levels of available spring forage (89 to 100% cover) than woodlots (46 to 60% cover) each year (Tables 7 and 8). Annual differences were not observed for percent of spring forage among woodlots surrounding alfalfa fields but did occur among openings (89.9% in 1999 and 100% in 1999). Among dry bean fields, annual differences in spring forage levels occurred among woodlots (46.8% in 1998 and 60.7% in 1999) but not openings (approximately 98% each year). F all/winter food and security cover Measurements of fall/winter food included the density (stems/ha) of shrubs and saplings and the basal area of oaks. When viewing aerial photographs, herbaceous openings were included since deer often graze on grasses and forbs until the ground is covered with snow (Gladfelter 1984). According to the photographs, the mean percentage of fall and winter food in evaluation areas surrounding the alfalfa and dry bean fields sampled in 1998 was 35.6% and 42.7%, respectively (Tables 7). Approximately 40% of the evaluation areas surrounding both crop types sampled in 1999 contained a fall and winter food source (Table 8). Measurements of security cover included the stem density of shrubs and saplings per ha, basal area of oaks and coniferous species, and horizontal cover of herbaceous plants and shrubs. Relative to fall and winter foods, security cover occurred at lower 30 Egon E 8 ~ v 335. was .me—woom which .3an .«o 830 “cop—on "a 5.3 5.83 :3: a 2 33¢ 2. 32: S; m 2:: gas: mwcsnam 2.33 8:38 @383 $.38 €35. 33 2mm: 3.62 a. _ RN 632:5: as: c is :33 8:28 Go. :3 2m 2:: 038 2:2 gags: Beam 33K: 33% o 5.3.: o 8.8: :5 £88 9 893:5 3.8 62: o N o ~ $253.: 88on 830 an: 2 m. c o _3 o 2.: 98 was mmn 2&8 €08 o :3 o as :8 awe and 5.3 one 3.2 o 3.2 2:38 35.5 3.8 38m $3 83: o :3: o 2 .3 €832§ ant Gas o 3.; o 2 .3 2%:8 gov .850 3280 so. 3 5.8 :3 23V ~33 :3: :2: 3% +5 5:8 owes msam 5.38:0 Bayes: «3::qu 3:39: :55 to £32 05:5, .NE: 2: as :2: 8:33 :2: a 5:222 532 35:8 co flow 53 5 :5 are: :58: 83.5; “853 Soc 258% 8 82m Bea—35¢ E Bums—SS 852.23 “SB“: :8 59 mos—g :82 .h 053. 31 mosh—m 55 35:23 .3288? 8 maniac can mac—woo? macaw $an 93 .mwczaam .335 x. 2.: 5.8 8 m: «:8 REEEKEB £2: . $93 :8: :3: 3:3 2:52: :83 82:8 828 8.8: 25% E 5:3 3:55: 82:8 83:8 32: ”N. _ m: E $35.5: and 8a. 3 2.3 8.: «:8 8.: Bum nob Bums—gm 3.8 $5 3.2. :3: 9:8 :38 328% 3.3 3.3 8.? 3.2 :38 88 3858:: 8 3.8 G _ .8 A38 o a 3 :N :2 23. E8 :28 58 A28 N 3 3 8 23 22 8N8 A38 2.8 :28 : mm o R :: 3 e 3 8:8 3:8 3:8 £8 2 8. :N :N :: 3 e 3 A88 8:8 58 8:8 8 S E S a 3 e a a8 3:8 382$ $52880 nommmmmm $5326 Sbgemm :53me :52 ”5:55 :83 : use. 32 fiwwo: E E ~ v 33% can .mmczgom .38“ .3me mo .550 “coupon "a 33 825 8:89 c 2.5 o 338 gas»: mwczawm 8.3 3.35 c _. 3 E32 3 233% 8 38m 2 53m $3323: .55. a: A 55¢ awe $2: a 2 3: _ 3 it; assesg Beam $.88 a3§ o 3.8.: o 2.3: 95 £38 3 8525 $2: 2 me o _ o _ $250: ”28% x5 God chi o N3 o a; 95 38V ES 308 ewe o Ed o Ed «8 5.9 Gav o 32 o 3.2 Shae AS555 «Bu 33m 55 :99 SE m: 03: o :2 €8§§ 3.3 8.3: o :8 o 8.8 2%..8 $3 560 3230 5. s :3; 2: 23v 2.? £8 2: 8.3 .§E>8 ease macaw MfifimQO “338% @5325 3:83: ”Samba «£52 azure, .NEx m.mm mm? 8.8 vozogm .32 E 53:22 .532 _abcoo mo 33¢ :3; S 93 £33 9598 3:553 “85a: bow oar—80v 3 mafia Rad—norm“ E Bums—go 33st? 35m: 8m £8 82? :82 .w 035. 35m 55 mmficoao .3230? 3 mwficvao 98 3230? 958m 38» EB imam—mam .335 ”a and 8 3w 2: $8 3Ea§~§83 8&8 AS .8 3%. 2.3 333 8.53.883 23.8 $38 ".32 2.2a «E 53.: 2:235 5.88 avg: :83 ENE «é Lamstmm and 608 2.2 2.2 88 3.: Eva nee Buns—gm 5.8 $8.8 3.: 8.? .38 :38 bfism $8 88.8 R? 5.8. .28 Be Esaaam g .8 $8 958 _ a :8 o mm s 2 s 3. 898 so. 8 5.8 N 2 :8 o ,3 s: e 3 G _ .8 a .8 5.8 N 8 8 o 8 s: 23 a _ .8 5 .8 828 M: «N A38 : _m s 3 s 3 9 8 .8 c «.8 5.8 8 3 A28 3 3 s 3 e e Ao\ov .8>oo _Scoduom 358 8883 A 388 8883 3an ED 053:5 A888 m 29¢ 34 percentages within most evaluation areas. Based on the aerial photographs, the mean percentage of each evaluation area that provided security cover was similar between cr0p types in 1998 (34.9% and 36.4% surrounding alfalfa and dry bean fields, respectively) but dissimilar in 1999 (43.2% and 34.7% surrounding alfalfa and dry bean fields, respectively) (Tables 7 and 8). As expected, stem densities of shrubs and saplings were much greater in the woodlots than openings that surrounded agricultural fields of both crop types (Tables 7 and 8). In woodlots, the average shrub stem density per hectare was 2,374.9 in 1998 and 2,613.6 in 1999. However, in 1998, the density of shrubs per ha was greater in Openings surrounding alfalfa fields (33,6133) compared to woodlots (2,231.9) (Table 7). The average sapling stem density in woodlots was relatively smaller than the stem density count of shrubs: 2,020.6 in 1998 and 1,934.6 in 1999. Oak trees represent another popular source of fall food for deer but occurred sporadically throughout the tri-county area. Rarely did a woodlot have more than 1 oak species, nor did oaks occur as the dominant tree species, demonstrated by its small, mean basal area of 0.65 cmZ/ha (Tables 7 and 8). Herbaceous and woody vegetation covered 44 to 63% of the profile board’s first height stratum (0 to 0.5 m) in woodlots and openings (Tables 7 and 8). In woodlots, vegetation covered 21 to 31% of each height stratum between 0.5 and 2.5 m. But in openings, the average percent cover for each stratum was between 1.0 and 2.5 m was 3.3%. 35 Thermal cover Not many cedar swamps, or any type of conifer stand, were found in central lower Michigan. As shown in Table 3, conifer stands were < 7 ha in size and comprised < 3% of the evaluation areas for all alfalfa and dry bean fields. Most stands contained small trees with a diameter-at-breast height between 5 and 23 cm. Consequently, a low number of trees dominated each stand, as indicated by the mean basal area of 13 mZ/ha (Tables 7 and 8). Around dry bean fields in 1998 the conifer stands had a slightly larger basal area of 16.9 cmzlha. The percent of canopy cover, however, within the conifer stands ranged between 60.7 and 81.9%, which, according to Bender and Haufler (1987), allows for high levels of snow interception. The average distance between conifer stands and evaluated crop fields was between 1,107.6 and 1,388.8 m in 1998 and between 1,129.3 and 1,207.0 min 1999 (Tables 7 and 8). Multiple stepwise regression analysis A multiple stepwise regression was conducted using the entire 17 habitat and 23 landscape variables that were sampled in the study area. However, the number of block permits issued and filled were not a part of the model when conducting the stepwise regression analysis for the third harvest of alfalfa in 1998. Only the variables that were significant at a=0.15 were identified as statistically significant indicators of alfalfa and dry bean losses. Also, each indicator remained in the crop loss model if it made ecological sense. During the second harvest of alfalfa in 1998, the size of conifer stands (P=0.0009, R2=0.9793) and total area of conifer stands within a 23.3 km2 evaluation area (P=0.0037, R2=O.7794) influenced observed losses (Appendix B-I). The area of hardwood stands 36 (P<0.0001, R2=1.000) and size of agricultural fields (P <0.0001, R2=0.9999) were other influential variables. Among these indicators, however, the total area of conifer stands within the evaluation area was the only variable with a notable trend due to the small sample size of alfalfa fields. Alfalfa losses occurred within evaluation areas that had at least 94 ha of conifer stands. During the third harvest, no conifer attribute influenced alfalfa loss. Instead, the perimeter (i.e., shape) of the alfalfa field (P<0.0001; R2=0.9809) and adjacent woodlots (P<0.0001; R2=1.000) , size of agricultural fields (P=0.0035; R2=0.7827), and total area of agricultural land within an evaluation area (P < 0.0001; R2=0.9999) influenced observed losses (Appendix B-I). Two notable trends were observed among the sampled fields. First, no alfalfa losses were observed among fields (n=3) that did not adjoin a woodlot. Among the fields that adjoined a woodlot, alfalfa losses were 2 -16%, and the smallest perimeter for the adjacent woodlot was 1,122.6 m. Secondly, the perimeter of alfalfa field number 8 (2,596.5 m), which had the highest percentage of alfalfa loss (-44%) (Table 4), was substantially larger than the remaining fields. Its perimeter was 1.5 times larger than the alfalfa field number 6, which had the next largest field perimeter (1,753.2) but no crop loss. In 1999, the regression analysis for alfalfa losses was performed on Harvests 2 and 3 combined because only 2 agricultural fields had a second cutting. Three variables were identified as influencing observed losses, none of which were found as significant indicators of crop loss in 1998: (1) the perimeter of adjacent agricultural fields (P=0. 1004, R2=0.6741); (2) diameter-at-breast height for oak trees (P=0.0063, R2=0.9587); and (3) the number of shrubs per ha (P=0.0716, R2=0.9882) (Appendix B-I). 37 Among dry bean fields, the same 3 variables influenced observed losses each year: (1) perimeter of the evaluated bean field (P=0.0004, R2=0.9877 in 1998; P<0.0001, R2=0.9999 in 1999); (2) perimeter of adjacent agricultural fields (P=0.0020, R2=0.9984 in 1998; P=0.0009, R2=0.9918 in 1999); and (3) perimeter of adjacent openings (P=0.0002, R2=0.8845 in 1998; P=0.0002, R2=0.9213 in 1999) (Appendix B-II). Different attributes of conifer stands also described bean losses. In 1998, the average distance between an evaluated field and conifer stand (P=0.0002, R2=1.000) was an influential variable whereas in 1999 it was the total area of conifer stands within the evaluation area (P<0.0001, R2=1.000). The area of each evaluated dry bean field (P<0.0001, R2=1.000) was the final variable to influence observed losses in 1998. The most interesting trends between observed crop losses and the variables indicating deer depredation occurred for the dry bean fields. First, in 1998 bean fields that were within 895 m of a conifer stand sustained relatively high crop losses (248%). During the study’s second year, no bean fields were within 895 m of a conifer stand. In fact, field number 8 was the closest to a conifer stand at a distance of 1,253 m, but it did not have any bean losses (Table 4). Secondly, the amount of conifers within the evaluation area surrounding each bean field that was sampled in 1999 varied across all fields. However, field number 11 had the smallest amount of conifer stands in its evaluation area (7.3 ha) and incurred the greatest bean loss (-25%) (Table 4). Only 2 dry bean fields in 1998 and 3 dry bean fields in 1999 adjoined an opening, but no trend was observed among the sampled fields within a year. In fact, in 1998, 3 fields that did not adjoin an Opening sustained dry bean losses. Between years, however, a comparison of openings among the dry bean fields with crop losses revealed that 38 relatively high dry bean losses occurred when the perimeter of an adjacent opening was _>_762.5 m. PODD evaluation Wildlife biologist focus group Three of the 9 invited biologists attended the meeting, and each individual addressed (or used to until recent reassignment) deer crop damage complaints within the tri-county area of Isabella, Mecosta, and Montcalm. The participants were familiar with the management approaches of other districts outside the study area that were not represented at the meeting. The biologists expressed a critical need for a standardized assessment tool because the MDNR did not have a set of guidelines that would permit consistent crop damage assessments across the state. For instance, one biologist asked farmers routine questions and mapped the areas of greatest damage for each agricultural field. Another biologist created an investigation form that he completed with each farmer. The third biologist sent farmers a deer damage inspection form along with the block permit(s) (not recognized as a standard approach within the MDNR). Despite the need for consistency, the biologists stipulated that the system should only influence the assessment process and not the determination of how many block permits a farmer should receive. These individuals stressed the need for biologists to continue weighing a farmer’s tolerance level for deer crop damage against the level of observed damage during a permitting process. Depending on the type of crop being produced, farmers contacted biologists at different times of the year to report deer damage complaints and to identify methods for 39 curtailing future crop losses. Farmers approached the biologists in January if growing apples, during the spring to early summer months if growing cherries, and in June or July if growing row crops. Corn producers, however, submitted deer damage complaints from June through August since deer forage on corn plants from the emergence stage through the tassel formation stage. For all other row crops (i.e., dry beans, table beans), deer damage was most problematic during only the plant emergence stage. Distinction between deer injury and damage was straightforward according to the biologists. Injury to dry beans and table beans occurred when deer foraged on the leaves of a top shoot. But if all the leaves of a bean plant are removed during the emerging stage, then damage is noted. Deer damage of corn includes leaf removal of an emerging plant as well as silk and tassel foraging. Deer exclosures are needed to identify intensity of deer grazing in alfalfa fields. And among fruit trees, deer damage is problematic for young trees only. While leaf removal is considered deer injury, browsing on a young tree’s terminal leader, resulting in lateral branching, is recognized as deer damage. Only one of the biologists trained assistants to conduct crop damage assessments each summer. The assistant’s training lasted 1 week, during which s/he accompanied the biologist in the field to learn the assessment process and damage markings of different animals. A second biologist did depend on an assistant once but found it too difficult to teach the assistant “the art form” of weighing observed crop damage levels against the farmer’s tolerance level for deer damage. Minor differences existed across the state regarding the issuance of out-of-season shooting permits. For instance, in one district farmers who received an out-of-season shooting permit were required to donate their deer meat to a soup kitchen. Yet, in 40 another district, depending on the circumstances, a farmer could keep the harvested deer for his/her consumption. Also, not all MDNR biologists required farmers to designate which shooters would use the issued shooting permits. Interaction between MDNR biologists and MSU Agricultural Extension Specialists or with other agricultural professionals occurred intermittently. According to 2 participants, Agricultural Specialists tended to be confrontational regarding the deer crop damage issue instead of acting as a mediator. However, the third biologist identified some good relationships with officials in an agricultural extension office. For example, representatives of the office often attended the MDNR’s deer management planning meetings. The biologists recognized changing land use patterns in their districts. More farms were being sold to recreational landowners for hunting club purposes, yielding mixed results for deer habitat quality. In some regions of the state, winter cover decreased but overall deer habitat quality increased because land remained idle. In other regions, an increase in winter cover lowered the quality of deer habitat, but the parceling off of corn and hay fields was creating more openings for deer use. In any case, the biologists commented that in recent years more farmers and neighboring landowners were partnering to implement Quality Deer Management as a way to reduce local deer densities and increase the size of older bucks. During the meeting’s second half the group discussed which attributes of deer habitat may contribute to deer damage. At this time, the PODD was introduced as a potential tool that biologists could use to evaluate cover types immediately surrounding a crop field and to identify a crop field’s potential for relatively high, medium, and low 41 levels of crop loss. The biologists supported this management approach but offered recommendations for strengthening the model’s landscape component: identifying how adjacent lands are used and what types of land exist within a 23.3 krrr2 area. Other recommendations included noting the farm’s size (owned and rented acreage and number of land tracts), and reducing the form’s length. Farmer interviews Eighteen of the 19 farmers were interviewed. All farmers offered the same distinctions between deer crop injury and damage that the biologists gave during their focus group and, except for one farmer who also worked with his county’s Soil Conservation District, said they worked with only MDNR biologists to address deer crop damage problems. Eighteen of the farmers relied on the Block Permit Program as their sole tool for controlling deer damage since other tactics were deemed ineffective or economically impractical for their farms (e.g., guard dogs, fences, repellents). Furthermore, a majority (n=16) of the farmers participating in this study dismissed the idea of letting strangers hunt on their property for one or more reasons: (1) fear of liability for any hunting accident that may occur; (2) distrust of a stranger’s hunting ethic and/or treatment of the farmer’s property; and (3) satisfied with the number of family members and fiiends hunting deer on his property. Not surprisingly, a majority of the farmers (n=12) commented that community members who also farmed acknowledged deer crop damage as a problem, whereas non- farrning residents did not understand or were indifferent to the deer problems of farmers. Four individuals explained that the focus of nearby hunt clubs on buck hunting posed the most problems for farmers who wished to eliminate a substantial number of antlerless 42 deer fiom the local population. Yet, 3 farmers along the Isabella-Mecosta County border stated that their communities were embracing the idea of harvesting only antlerless deer during the regular hunting season. For one farming community, it was a group of non- farming, recreational landowners that initiated the no-buck philosophy. When explaining the PODD to each farmer, a modified version of the model was presented to focus the discussion on the model’s underlying concept and management implications. This version showed the categories describing each habitat attribute without each category’s associated index value. All farmers understood how the PODD would be used and the relationship between the quality of local deer habitat and the size of local deer populations. Among the 18 farmers interviewed, 55.6% (n=10) said they would consider learning how to use the procedure themselves. However, 2 of these individuals (plus 1 person who is not interested in learning how to use the PODD) pointed out that they as a renter did not have management authority for untillable land. The remaining 8 farmers opposed using the PODD because they: 1) were concerned that application of the PODD would transfer his deer problem to another landowner; 2) preferred using a cost-effective and persistent technique, such as a spray that lasts throughout the growing season but does not endanger the deer’s health; 3) strongly believed deer harvesting was the only answer; 4) did not own the untillable land; or 5) were disinterested in using the tool. Several of the farmers offered suggestions for either the PODD or deer management, in general. Regarding the PODD, 2 individuals recommended noting how many deer-vehicle accidents were reported within the farmer’s township during the previous year. As for deer management, suggestions included: 1) more research in the 43 area of crop production (e. g., development of better sprays or crop seeds that contain a chemical to deter wildlife feeding); 2) a deer feeding ban; 3) issuing only antlerless tags during the regular hunting season; and 4) initiating a compensation program for farmers with deer crop damage. Taking a more political approach, 1 farmer commented that deer (and other wildlife species damaging crops) should be recognized as property of the private landowner and not the State of Michigan. Multiple stepwise regression analysis of the PODD For this regression analysis, 14 (13 in 1999) habitat and 16 (17 in 1999) landscape variables were used to identify which variables significantly influenced alfalfa losses at 0t=0.15. In 1998, 10 variables were identified as significant indicators of deer crop damage during the second harvest of alfalfa (Appendix B-III). Four of the variables characterized hardwood stands adjoining evaluated alfalfa fields. Tree species dominant in the overstory (P=0.0497, R2=0.9998) was one significant variable. Among the 8 sampled fields, maple was commonly the dominant tree species within hardwood stands. The percent of understory grth (i.e., grass, forbs and shrubs; P=0.0081, R2=0.9935) also was significant. Grasses and forbs occupied the understory of all woodlots, and shrubs also were sampled in woodlots surrounding 2 alfalfa fields. The basal area of the tree species dominant in the overstory (P<0.0001, R2=0.9992) was the third variable influencing observed crop losses. Six of the alfalfa fields were adjacent to mature woodlots (basal area _>_ 8 mZ/ha). The final variable describing hardwood stands and influencing crop losses was the distance between each hardwood stand and the alfalfa field (<177 m) (P=0.0004). The remaining variables influencing alfalfa losses during the second harvest were percent of field edge adjacent to woodlots (P=0.0080, R2=0.9965), openings (P=0.0006, R2=0.8323) and development (P<0.0001, R2=0.9697) as well as the percent of woodlots (P<0.0001, R2=l .000), agricultural land (P=0.0003, R’=0.5102), and openings (P<0.0001, R2=0.9177) within the 23.3 lam2 evaluation area. During the third harvest of alfalfa, the percent of field edge adjacent to woodlots (P=0.0008, R2=0.4550), agricultural fields (P<0.0001. R2=0.8012), openings (P=0.0021, R2=0.9603), and development (P=0.0002, R2=0.9995) influenced alfalfa losses. The percentage of woodlots (P<0.0001, R2=1.000) and development 1998 (P <0.0001, R2=O.9985) within the evaluation area also significantly influenced alfalfa losses (Appendix B-III). Agricultural land was the predominant cover type (>50% of the field edge) adjoining half of the sampled alfalfa fields in 1998. The percent of field edge adjacent to woodlots was 230% (n=6), and 75% (n=6) of the sampled fields had <10% of their edge adjoining openings. Within the 23.3 km2 area surrounding 6 sampled alfalfa fields, woodlots comprised 30 to 60% of the area, agricultural fields comprised 20 to 50%, openings comprised 10 to 30%, and development comprised >25% of the area (n=6). Similar results were observed for dry bean fields in 1998. Percent of field edge adjacent to woodlots (P<0.0001, R2=1.000), agricultural fields (P=0.0020, R2=0.9930), and openings (P=0.0012, R2=0.7264) were indicative of deer depredation. The percent of woodlots (P=0.0232, R2=0.9809), openings (P<0.0001, R2=O.9925), and development (P=0.0059, R2=0.4068) within the evaluation area and basal area of hardwood species dominating the overstory of adjacent woodlots (P=0.0012, R2=O.9689) also influenced dry bean losses (Appendix B-III). A majority of the dry bean fields (n=6) had fairly 45 mature trees in the adjacent hardwood stands (basal area 28m2/ha), and 230% of the field edge was adjacent to woodlots (n=6). Agricultural land was the predominant cover type adjoining the sampled fields (>50% of the field edge, n=4). Openings comprised <10% of the field edge for 6 fields, and <25% of the field edge was adjacent to development (n=5) . Before field-testing of the PODD began in 1999, the model was recalibrated using the regression analysis results for 1998’s vegetation data. Changes to the model entailed the development of new index values for descriptive categories in each cover type and more specific descriptions of understory growth for deciduous and coniferous woodlots and distances separating the evaluated crop field and surrounding cover types. A variable was added to identify how many conifer stands occur within a 23.3 km2 evaluation area. Under the new PODD model, percent of field edge and cover types within a 23.3 km2 area remained as significant variables influencing alfalfa and dry bean losses. No characteristics of hardwood forests were identified as influential variables of deer crop damage, but the number of conifer stands within the evaluation area was influential for dry bean losses (P=0.0048, R2=0.8015) (Appendix B-IV). A maximum of 27 conifer stands were identified within the evaluation area surrounding 5 dry bean fields opposed to a maximum of 21 conifer stands among 4 alfalfa fields. Among the alfalfa fields, the indicators of crop loss were percent field edge adjacent to other agricultural fields (P<0.0001, R2=0.9687) and openings (P=0.0423, R2=0.7281) and percent of openings (P=0.0085, R2=0.6641) and developed lands (P=0.1218, R2=0.7611) within the evaluation area (Appendix B-IV). In 1999, fewer woodlots adjoined the sampled alfalfa fields (525% of the field edge, n=5). The percent 46 of field edge adjacent to agricultural land varied between 10 and 50% (n=5), and development comprised 26 to 50% of the field edge for 5 dry bean fields. Openings were either <10% or >50% for 5 fields. For dry bean fields, the remaining indicators of crop loss were percent of field edge adjacent to woodlots (P<0.0001, R2=1.000), agriculture (P=0.0016, R2=0.S772), and openings (P=0.0647, R2=0.9199) and the percent of agricultural land (P<0.0001, R2=0.9999) and development (P=0.0279, R2=O.8805) within the 23.3 ion2 evaluation area (Appendix B-IV). For 4 of the sampled fields, woodlots comprised 26 to 50% of the field edge. The percent of field edge adjacent to agricultural land and openings was more varied, but for most sampled fields (n=5) development was adjacent to 2 25% of the field’s edge. Less than 60% and 50% of the evaluation area respectively contained woodlots (n=4) and agricultural land (n=5). Only 1 dry bean field had no development within its evaluation area, and openings comprised <10% or >50% of the evaluation area for a majority of the fields (n=6). Field classifications using the PODD Crop loss percentages and total PODD index values were compared qualitatively to determine how many alfalfa and dry bean fields were correctly classified using the PODD (Table 9). Nearly half (46.7%) of all crop fields sampled in 1998 (n=15) were classified correctly: 28.6% of the alfalfa fields (n=2) during the second harvest, 62.5% (n=5) during the third harvest, and 28.6% (n=2) of the dry bean fields. Refinement of the PODD model did not increase the number of correct field classifications in 1999; in fact, the percentage of correct classifications dropped. Among all sampled fields (n=13), only 3 fields (23.1%) were classified correctly: 33.3% (n=2) of 47 e\oN: 2 2 $2. 2 v— o\om_ o\c~- $2- QN 2 o\owv. NN g N— : nwE 6m 8 _N ”85:58 .om 9 I ”one: 8 33 .3 8 c u 33? x33 DQOA 3 Emma o\o~N- .Ar AEEBEV o\oo~- 3 3- Assoc o\oc~- 8 T 3285 o 2 won—g + u 3.ch owned—u 33030 "a 2 o\owN- ~N o\om _ r am A 3 $9.1 $3.. o\o_N cm 2 2 o w a 58:2 22% o\oMN N~ .xba Nx 0 Ram o\oom S 2 a o\o£- $3- 2 o\ov¢r ~N o\ow_ o\om- 3 S Q o\o_ T e\ewm- 3N <24 n v m N _ ”6v .6: 3.3 on ”on .6: 3.3 33 ”6V .6: 3.3 5333: ”6v ..3 as Es 33m in ”5v .6: 3.3 6: ”6V .6: 3.3 23 ”6V .6: 3.3 .5585 ”SC .6: as «m3 «dam ”8V .6: 3.3 0: ”8v .6: 3.3 33 “8V .6: 3.3 .5535 “5v ..9: as «M3 an «:32 980 ..32 no.6 a 88058 839» guawoz .wao— E 53:22 533 35:00 E v.20: 53 s 98 £33 32332: 8.. EV m8. nob he 893:3qu v5 9V «2:? x09: COCA .«o :85930 .a 039—. 48 the alfalfa fields during the second harvest, none during the third harvest, and 14.3% (n=1) of the dry bean fields (Table 10). Large misclassifications occurred when the PODD index value was 3 or more points away fiom the numerical border categorizing observed crop losses for a specific field. Deviations of 3 or more were selected to insure liberal description of potential crop losses. Because annual variations in climate and vegetation growth will occur, interpretation of the PODD index value for a particular field should account for this variability. Large discrepancies existed for 33.3% (n=5) of the misclassified fields in 1998 and 61.5% (n=8) of the misclassified fields in 1999. For example, during the first alfalfa harvest in 1998, field number 5 had relatively high levels of crop loss (-38%), but according to the PODD deer would use this field at low levels (index value=10). Likewise, dry bean field number 13 in 1999 had relatively low levels of crop loss (-3%), but according to the PODD deer would use the field at medium levels (index value=17). Slight misclassifications (i.e., the PODD index value was 52 points away from the nrunerical border categorizing observed crop losses) for 33.3% (n=5) of the sampled fields in 1998 and 30.8% (n=4) of the sampled fields in 1999. For instance, in 1998 alfalfa field number 8 had relatively high crop losses (-40%) during the third harvest. The PODD index value for this field was 20, but a value of 21 represented a prediction of high crop losses. Similarly, dry bean field number 12 had no crop losses (18%) in 1999, and its PODD index value was 13 which predicted medium crop losses. However, this index value is also the lowest value possible for a medium crop loss prediction. 49 nwE .3 9 cm ”8362: .2 9 2 ”one: 8 33 .fi 9 c u $29, .865 QQOA x— EwEv o\o_N- N. ABE—oofiv o\oom- 8 3- 230: o\oo_- 9 _. x285 o 8 839, + u mowefl 0mg voZomnO "a o\om_ o\oNN o\o: $0 o\oMr o\o_- : RN .2 2 2 Q3 3 o\omN- o\oom 2 2 : 2 o w a o eoaeazsora o\ovr 2 m o\ew 2 o\o~_ Room 2 2 v m 2 o\oCM- o\oo_ AVon N3 2 N _ ”on .6: .33 o: ”on .. 6: 3.3 33 ”8V .. 6: 3.3 .5535 x9 .6: as .33 382. in ”SC .. 6: 3.3 0: ”6V .6: 3.3 33 ~on .6: 3.3 £3.35 “on as: as 3.3 «an ”5V .6: 3.3 on ”on .6: 3.3 33 ”on .6: 3.3 53.38.: ”5V .6: as Es an £32 980 ..32 nob a Suwanee mos—g guawoz .32 5 52:32 832 15:3 8 vac—ow 53 be can 33? 363%.: no.“ Ev 32 nob mo mama—820a v5 A>v 83? x33 QQOm we 5389.80 .3 oBah 50 DISCUSSION In this study, select features of deciduous and coniferous woodlots, openings, and agricultural fields influenced how much deer damage the adjacent alfalfa and dry bean fields sustained. Furthermore, the PODD emerged as a potential tool that wildlife biologists and farmers could use to foresee problems with depredating deer and to identify management techniques that would alter deer habitat and populations accordingly. Observed crop losses In both years, alfalfa losses occurred at higher levels during the late summer harvest (Table 4). At this time of year, deer typically forage on a mixture of herbaceous plants and leaves of shrubs and saplings to support the impending energy needs of the fall rut and winter season (Gladfelter 1984, Beier and McCullough 1990). Yet, the summer of 1998 was relatively dry and hot (Table l), perhaps forcing deer to supplement natural forage in woodlots and openings with the nutritious alfalfa The same rationale can be applied to the comparison of dry bean losses between years. Dry conditions accompanied the planting of bean fields in spring 1998, a time when deer forage predominately on grasses and forbs to regain strength and restore fat reserves lost during the winter months (Beier and McCullough 1990, Bender and Haufler 1987). Yet the lack of green growth in March through May 1998 may have encouraged deer to forage in dry bean fields, thereby producing the differences observed in crop loss between 1998 and 1999 (Tables 4). Minor nibbling on the leaves of a bean plant when it begins to sprout automatically strmts the plant’s growth. However, complete leaf 51 removal from the bean plant’s first shoot terminates the plant’s growth for the season (Glenn Dudderar, former MSU Wildlife Extension Specialist, personal communication 1998). Woodlot variables influencing alfalfa and dry bean losses In 1998, highly significant levels of crop loss were observed in the sections of alfalfa and dry bean fields adjoining woodlots (Table 5). Based on the results of the regression analyses, these losses were attributed to different components of conifer stands found in the evaluation areas. Among alfalfa fields in 1998, the total area of conifer stands in the evaluation area and the mean size of each stand were significant indicators of crop loss (Appendix B-I). Among dry bean fields, the average distance between each field and conifer stand (1998) and the total area of conifer stands in the evaluation area (1999) significantly influenced crop loss observations (Appendix B-II). Collectively, these variables describe the tendency of deer to rely on conifer stands for thermal cover in late spring. During the study’s first year, night temperatures averaged below 14°C in May and June, but in 1999 the average low fluctuated between 9.5 and 15°C (Table 1). Once the nights grew warmer, the availability of thermal cover within close proximity of foraging sites was no longer a concern for deer, as demonstrated by the lack of conifer variables influencing alfalfa losses during the third harvest. Also, deer may have used the canopy cover and understory growth of conifer stands to seek respective refuge from the summer sun and predators. The perimeter of woodlots adjoining evaluated crop fields (mean=8,020.6 m; Table 3) also influenced alfalfa losses in 1998 (Appendix B-I). As the perimeter of an adjacent woodlot increased, the ratio of woodlot edge to woodlot area also increased, 52 thereby increasing the amount of security cover lining the edge of an alfalfa field. Also representing the value of security cover to depredating deer was the mean size of hardwood stands throughout the evaluation area (62.7 ha) (Tables 6 and Appendix B-I). By nature, deer feed within close proximity of cover to seek immediate protection from predators and human disturbances (Prior 1983, Beier and McCullough 1990). Deer also use the cover of trees and shrubs and saplings to rest and ruminate between feeding periods. According to Bender and Haufler (1987), 23,000 shrub/sapling stems per hectare represent optimal security cover for deer. In 1998, sampled woodlots within the evaluation area for each alfalfa field averaged nearly 4,000 shrub/sapling stems per hectare (Table 7). Thus, throughout the evaluation area ample security cover existed for deer. Two among 3 of the variables influencing alfalfa losses in 1999 were the number of shrubs per ha within woodlots and the diameter-at-breast height of oak trees (Appendix B-I). Collectively, these variables represented the value of fall/winter food to deer. An average of 1,475 shrubs per ha with approximately 3,250 stems per ha (Table 8) occupied the woodlots surrounding alfalfa fields. According to Bender and Haufler (1987), woodlots and openings should contain at least 3,000 shrub stems per ha to satisfy the browsing needs of deer. The acorns of oak trees represent another fall/winter delicacy for deer (Gladfelter 1984). Although the woodlots surrounding alfalfa fields in 1999 contained relatively small oak trees (mean DBH per sampled plot was 9.7 cm; Table 8), the diameter-at-breast height of oak trees in 1998 was much smaller (3.3 cm; Table 7). 53 Agricultural and opening variables influencing alfalfa and dry bean losses According to the stepwise regression analysis, the perimeter of agricultural fields adjacent to dry bean fields (1998 mean=1830.0 m; 1999 mean=1368.4 m) and the area of the dry bean fields (1999 only; mean=13.2 ha) influenced crop losses (Appendix B-II). According to the Pearson correlation analysis, as the perimeter of adjacent agricultural fields decreased in 1999, the likelihood of dry bean losses increased (Pearson coefficient-$0.684, P=0.0202). Typically deer do not venture >90 m from their covered resting spots when searching for food. Thus, agricultural fields with a maximum width 5180 m are completely available to deer foraging (Bender and Haufler 1987). In fact, no significant losses (P>0.10) were observed in alfalfa or dry bean fields of northern lower Michigan when the sampled fields had a length and width >180 m (Braun 1996). The perimeter of openings adjacent to dry bean fields each year (1998 mean=174.0 m; 1999 mean=627.9 m) (Appendix B-II) also influenced crop losses, even though no crop losses relative to openings were deemed significant (Table 5). Unlike the perimeter of adjacent agricultural fields, this variable was negatively correlated with observed crop losses, which indicated an increase in dry losses as the perimeter of adjacent openings increased. A significant correlation occurred for dry beans in 1998 (Pearson coefficient=-0.623, P=0.0520). Interestingly, the perimeter of openings adjacent to all alfalfa and dry bean fields was small in comparison to the perimeter of sampled agricultural fields (Table 3). Moreover, herbaceous openings comprised <11% of the sampled fields’ edge (Table 3) and 59.5% of the evaluation area for alfalfa and dry bean fields (Table 6). According to Bender and Haufler (1987), herbaceous openings should comprise 10 to 30% of an 54 evaluation area to offer deer optimal spring forage. Yet, if openings comprise <10% of the evaluation area, adequate levels of spring forage may still exist. In central lower Michigan, it is plausible that deer used the alfalfa and dry bean fields as an extension of adjacent openings to maximize their intake of green plants, especially in 1998 when the precipitation was relatively low (Table 1). Also, deer may have used agricultural fields adjoining dry bean fields for the same reason. In the Habitat Suitability Index model for white-tailed deer, corn and alfalfa are identified as optimal spring food sources (Bender and Haufler 1987). Corn adjoined 5 dry bean fields in 1998 and 4 dry bean fields in 1999 whereas alfalfa bordered 2 dry bean fields in 1998 and 3 bean fields in 1999. The perimeter of the alfalfa (second harvest, 1998) and dry bean fields served as the final indicator of crop losses due to deer foraging (Appendix B-I and H). This variable described the shape, or edgezarea ratio, of an agricultural field and was negatively correlated with observed losses. As the shape of a field becomes more convoluted, the field has increasingly more edge touching an adjacent cover type(s) and, therefore, becomes increasingly more vulnerable to deer depredation (assuming the adjacent cover type is an opening or woodlot). Prior (1983) acknowledged this shape effect when discussing variables influencing deer browsing in tree plantations. Like plantations, crop fields provide less security cover for deer when they are large and square and have a perimeter that is proportionally less than the field’s total area. PODD evaluation The 19 farmers participating in this study contacted MDNR wildlife biologists for assistance after observing signs of deer damage in late spring or early summer. At that time, a wildlife biologist or trained technician visited the farmers’ property to assess the 55 extent of deer damage in vulnerable crop fields and determine how many block permits each farmer should receive during the regular hunting season. However, wildlife biologists attending the focus group meeting commented that the MDNR needs a standardized approach for assessing deer damage. The PODD could help wildlife biologists meet that need. The purpose of the PODD is to help farmers plan their planting scheme for a future growing season (e.g., plant a less vulnerable crop next to a woodlot, no field cultivations) and help wildlife biologists determine what the antlerless deer harvesting quota should be for a specific Deer Management Unit. By evaluating deer habitat next to an agricultural field, wildlife biologists and farmers can determine which habitat features may influence predicted levels of damage. Also, completion of the PODD for a cohort of crop fields within a given area (e.g., 23.3 kmz) can help wildlife biologists and farmers identify which habitat features may influence relative densities of deer. Thus, 1 wildlife biologist could use the PODD to increase her/his interaction with multiple farmers and enhance community efforts in managing the deer crop damage issue flour a habitat and population perspective. All wildlife biologists attending the focus group meeting and 55.6% (n=10) of the interviewed farmers expressed a willingness to learn how to use the PODD. However, successful implementation depends on community level support. A general trend throughout Michigan is for farmers to perceive the deer density as too high and for hunters to view the density as too low (Mirmis 1996). Yet, in central lower Michigan, 3 farmers residing along the Isabella-Mecosta County border commented that farmers and non-farming, recreational landowners were forging partnerships to address the problem 56 of locally abundant deer. Although this is a step in the right direction, the problem of increasing hunter access on private, agricultural land remains. Only 10.5% (n=2) of the farmers in this study allowed strangers to hunt on their property. In Ohio, Stoll and Mountz (1983) and Scott and Townsend (1985) observed a similar aversion among farmers to let strangers hunt on their property, mostly because the farmers feared an inability to control hunter behavior. Improved efforts in education and outreach at the community level may overcome the above farmer-hunter conflicts (see Chapter 2). Although the success rate of the PODD’s final version was low in 1999, continued refinement of the procedure’s structure and content may still make it a valuable management tool for wildlife biologists and farmers. During the first harvest of alfalfa, 33.3% of the fields were classified correctly opposed to none during the second harvest (Table 13). Correct classifications occurred for only 14.3% of the dry bean fields. Most likely, changes in the PODD before field-testing fostered this low success rate. After the first field season (1998) the PODD was modified to include more detailed descriptions of deer habitat (e. g., percentage of grass and shrubs in woodlots and openings). The index values for descriptive categories also were altered to reflect more appropriately the ecological relationship between deer and a specific habitat component. Perhaps these changes made the categories too narrow. Perhaps results of the PODD would be more accurate if only broad characterizations of the different cover types adjoining a crop field are noted. Thus, future studies on Michigan’s deer crop damage issue should incorporate continued refinement of the PODD’s descriptive categories before the procedure is used as a management tool. 57 Other features of the PODD should be modified as well. First, the accuracy of the PODD may be improved by integrating a climate variable. Currently, the PODD does not account for annual variations in weather because Braun (1996) did not observe a significant correlation between climate and deer crop damage in northern lower Michigan. However, in my study the level of precipitation (rain and snow) varied substantially between 1998 and 1999 (Table 1). As mentioned earlier, it is possible that the improved growing conditions of 1999 diminished the need for deer to forage in alfalfa and dry bean fields in spring and early summer. In 1999, woodlots adjacent to dry bean fields had higher percentages of spring forage compared to the woodlots adjoining dry bean fields in 1998 (Tables 7 and 8). If the above supposition is true (i.e., dry conditions and consequently less availability of spring forage in natural areas force deer to forage more in crop fields), then the ecological framework of the PODD must change. I initiated this research project assuming a direct relationship between habitat quality and deer depredation among agricultural fields. As previously discussed, the work of Braun (1996) indicated that as the quality of deer habitat surrounding a crop field increases, more deer would occupy the cover types adjoining the evaluated crop field, thereby increasing the crop field’s likelihood to sustain deer damage. Yet my results suggest the opposite, at least for central lower Michigan. Then again, perhaps deer are more prone to forage in agricultural fields if adjacent openings and woodlots have either relatively high or low habitat quality ratings. If such is the case, then farmers who have openings and woodlots of medium quality immediately surrounding their agricultural fields would have the least amount of problems with depredating deer. 58 Inherent to this discussion of which habitat characteristics promote deer use of agricultural fields is the role of population dynamics and animal movements and behavior. Habitat quality alone cannot account for how many deer use a crop field. Herbivore-plant interactions are complex and involve 6 variables that relate community composition (including soil characteristics) and habitat quality to the behavior and distribution of herbivores (Figure 1) (Morrison et al. 1998). Thus, when predicting foraging pressure by deer for a single crop field, wildlife biologists must consider the population dynarrrics of deer at the farm and regional scales. Natality and mortality factors along with daily and seasonal movement patterns of deer collectively determine what the density of deer will be at each spatial scale. Thus, just as the density of deer will vary among farms within a single region, so will the quantity and quality of deer food and cover. Future studies must be conducted to determine how the PODD can effectively integrate population dynamic variables. Three farmers recommended that the number of deer-vehicle accidents should be added to the PODD model because it may help wildlife biologists estimate the size of local deer densities, particularly at the township level. According to Corteville et al. (1999), however, the number of deer-vehicle accidents adjusted for traffic volume is not a strong predictor of total deer harvests (antlerless and buck) in Michigan. Only the buck harvest could be predicted accurately by reviewing deer-vehicle accident reports. Furthermore, Manson-Hansen (1998) discovered regional inconsistencies in the ability of deer-vehicle accidents to predict total harvest levels during the firearm season and combined season of firearms and archery. Therefore, significantly different results in the 59 Figure 1: Model of herbivore-plant interactions as shown in Morrison et al. (1988). Arrows represent negative and/or positive causal pathways for change. fi Plant demography and community composition Behavior and physiology of herbivores Soil and landscape characteristics Quantity and quality —> of plant food and cover Distribution and ‘—— abundance of herbivores 60 Performance of individual herbivores efficacy of the PODD model may not occur if the model included the number of reported deer-vehicle accidents for a local community. The percent of misclassifications for agricultural fields in 1998 (53.3%) and 1999 (76.9%) (Table 13) may indicate high variability for at least one habitat attribute. Each year the number of shrub and sapling stems per hectare varied a great deal within woodlots and openings for both crop types (Tables 7 and 8). For instance, in 1998 the standard error for shrub stems in woodlots surrounding alfalfa fields was 849.9, and in 1999 it was 1,129.7. Better accounting for this type of variability would occur if the size of the evaluation area was expanded to reflect daily movements of deer (Sitar 1996). Consequently, a wildlife biologist would have to gain access to neighboring crop fields. Regression of the PODD variables yielded different results in comparison to the regression on all habitat variables evaluated to describe deer damage. Although both models contained variables representing the same attributes of food and cover, the manner in which measurements were recorded differed. Under the PODD model, the collection of detailed, habitat information was sacrificed to facilitate easy habitat assessment by biologists and farmers (i.e., the biologist selects the best descriptive category for a habitat attribute). Secondly, the PODD model was applied to only the cover types immediately next to the alfalfa and dry bean fields whereas the evaluation area during vegetation sampling was much larger (23.3 kmz). Hence, compared to vegetation sampling techniques, the PODD model is less sophisticated when characterizing habitat attributes. In 1998, several characteristics of fall/winter and spring food within hardwood stands significantly influenced alfalfa losses during the first harvest and, to a lesser 61 extent, dry bean losses: (1) dominant tree species in the overstory; (2) percent of understory vegetation; (3) basal area of the dominant tree species; and (4) distance between the hardwood stand and alfalfa field (Appendix B-III). Basal area was the only variable common to both alfalfa and dry bean fields. The number of conifer stands within a 23.3 km2 area was the only significant woodlot variable in 1999 (Appendix B- IV). Because few oak trees were sampled in the study area’s hardwood stands, deer depended on twigs and grass as sources of fall/winter food. But the availability of spring forage within woodlots was sub-optimal each year: 552% in 1998 and 560% in 1999 (Tables 7 and 8). According to Bender and Haufler ( 1987), 25 to 50% ground cover (i.e., strictly herbaceous plants) yields a Habitat Suitability Index value of 0.5 for white-tailed deer in the Upper Great Lakes Region. As with the regression analysis on all variables evaluated to describe deer damage, the availability of thermal cover remained an important habitat attribute for deer during the spring. The percent of each cover type adjoining the alfalfa and dry bean fields and the percent of each cover type within a 23.3 km2 area served as the remaining significant indicators of deer damage under the PODD model. Fragmented, agricultural landscapes dominated central lower Michigan in 1998 and 1999 (Table 6). Therefore, deer most likely established relatively large home ranges to fulfill annual habitat requirements. Sitar (1996) documented this tendency when following the movements of 73 radio- collared deer in northern lower Michigan in 1994 and 1995. During the first year, 32.5% of the radio-collared deer occupied an agricultural area year round, and in 1995 51.2% of the deer were permanent residents of agricultural communities. For these deer, the 62 annual home range averaged 389.9 ha whereas migratory deer occupied an average home range size of 319.7 ha. Although the difference was not statistically significant, it does emphasize the need to expand the evaluation area when using the PODD. 63 CHAPTER 2: Communicating with Farmers and Hunters INTRODUCTION Since the 1970’s, public interest in natural resources has diversified and played a more prominent role in natural resources management. First, traditional users are more specialized in either the animals they prefer to harvest (e. g., deer and elk only) and/or the style in which they prefer to hunt or fish (e. g., bowhunting versus firearm). Secondly, new interest groups represent historical values of preservation, and citizens throughout the United States often demand that management decisions reflect their interests and concerns (Decker et a1. 1996). Federal and state agencies have tried to meet this demand through traditional public involvement techniques, but conflicts still arise. Citizen surveys, public hearings, comment periods for a management plan or policy, and to some extent citizen advisory committees represent the primary forms of traditional public involvement in natural resources planning (Kathlene and Martin 1991, Moote and McClaran 1997, Steehnan and Ascher 1997). Mostly, these approaches are unsuccessful because they produce unequal representation of public interests, no direct stakeholder interaction, and inadequate exchange of information among stakeholders (Table 14). If these techniques do succeed, however, it is typically for a short period of time either because the public’s attitude changes or public values remain the same (Rasmussen and Brunson 1996). To succeed at conflict resolution, natural resource managers need to better address the public’s perception, attitudes, and values that influence management issues (Peyton 1984, Rasmussen and Brunson 1996, Fritzell 1998). Increasing stakeholder interaction may help stakeholders view a human-wildlife conflict critically and identify the 65 Table 11. Comparison of public involvement techniques (Kathlene and Martin 1991, Moote and McClaran 1997, Steelman and Ascher 1997). Public involvement Strength Weakness technique Mailed survey 0 Recipients randomly One-way selected. communication. 0 Identifies public concerns Static approach. and attitudes on a given issue. Public hearing Open to the public. One-way Identifies public concerns conununication. and attitudes towards a Low attendance; oflen management plan. polarized. Limited commenting time. Public less inclined to accept final decision. Referendum 0 Registered voters have No information authority to make rules. exchange. Citizen advisory committee Public more inclined to accept final decision. Two-way communication. Higher potential for comprorrrise. Time intensive. Few participants; often appointed. 66 consequences of each management alternative. Interactive processes that emphasize voluntary participation and free-flowing dialogue encourage participants to share information, values, and other perspectives in a non-polarizing environment (Selin and Chavez 1995, Stout et al. 1996). If designed properly, interactive processes can help stakeholders understand the scientific principles and differing beliefs and values that underlie a resource management issue (Wondolleck 1985). In “appreciative planning,” natural resources stakeholders seeking a proactive solution to management issues unite only to share information and identify common values for a natural resource (Selin and Chavez 1995). Extension specialists represent one group of professionals that could implement this design since their long-term focus is fostering land stewardship and seeking “common ground” among private landowners (Wotecki 1997). I applied the goals of appreciative planning to Michigan’s white-tailed deer crop damage issue by developing an interactive workshop format. In collaboration with faculty members in the Departments of Fisheries and Wildlife and Forestry at Michigan State University, I developed a format that would accommodate farmers, hunters, and wildlife biologists. The workshops were developed in direct response to the recommendations of Campa et al. (1997) and Minnis (1996). Through survey research in the Lower Peninsula, Minnis (1996) identified a need for improved communications about this issue. Surveys were distributed to 2,134 farmers (49% adjusted response rate) and 792 hunters (65.2 % adjusted response rate) in selected counties and revealed attitudinal differences between and within the 2 groups regarding the crop damage issue and effectiveness of deer management techniques. For 67 instance, most farmers experiencing deer damage regarded 1994 as their worst year; however, there were always a few farmers who reported other years as the most severe. Furthermore, misperceptions about the summer shooting and block permit programs emerged as a critical factor in the controversy. Thus, Campa et al. (1997) and Minnis (1996) recognized a need to increase understanding and tolerance for different values and beliefs held by affected farmers, hunters, and the biologists responsible for recommending deer harvest levels. To address this communication need I pursued the following objective: to develop and pilot a workshop format that unites wildlife biologists, farmers, and hunters in an open discussion about Michigan’s deer management issue. The workshop format was designed to complete, hopefully, 3 goals: 1. Help identify stakeholder concerns that influence emerging or existing issues. 2. Inform stakeholders about the biological and ecological principles underlying management issues. 3. Help stakeholders understand that wildlife management programs address diverse interests and values. 68 METHODS Focus groups Two focus groups were held in J anuary-February 1999, one with hunters in the Upper Peninsula and the other with farmers in the northern Lower Peninsula. These meetings were held to identify how each stakeholder collects information on wildlife management issues, perceives identified sources of information, and prefers to learn about wildlife management issues. I began each meeting with a discussion of the focus group’s purpose and established the ground rules (e. g., one person spoke at a time, I maintained the group’s discussion with prepared questions). Additionally, I obtained permission from each group to tape record its discussion. The focus groups lasted 2 hours, and the same questions were presented in each one. For the farmer’s focus group, I generated a list of 252 individuals through 9 county offices of MSU Extension: Alcona, Arenac, Gladwin, Grand Traverse, Iosco, Kalkaska, Lake, Osceola, and Oscoda. MSU Extension representatives were asked to provide a list of all dry bean and alfalfa producers in their county and not simply the ones with perceived deer problems. A list of 90 hunters living in the Upper Peninsula was received from the Chief Executive Officer for the U. P. Whitetails Association, a Sportsman’s group in the Upper Peninsula. My desired focus group size for each meeting was between 8 and 12 people. Therefore, I randomly selected 20 farmers and 20 hunters from each master list and mailed an invitation letter explaining the focus group’s purpose. Enclosed was a pre- stamped, self-addressed postcard that listed 2 possible meeting dates. Invitees were 69 asked to mark on which date(s) they could participate and return the postcard by the specified date. Letters were mailed 3 weeks in advance for the farmers, 6 weeks in advance for the first 10 hunter invitees and 2.5 weeks for the last 10 hunter invitees. Workshops Two full-day workshops were held in April 1999, one in the Upper Peninsula (UP) and one in the northern Lower Peninsula (NLP). Date selection was based on the schedule of the MDNR Wildlife Biologists and workshop planning team. The UP workshop targeted 3 counties in the Upper Peninsula’s central region: Delta, Iron, and Menominee. The NLP workshop targeted 13 counties in the Lower Peninsula’s northwest region: Antrim, Benzie, Charlevoix, Gladwin, Grand Traverse, Kalkaska, Lake, Leelanau, Manistee, Mason, Missaukee, Osceola, and Wexford. These targeted areas have a history with deer crop damage problems and reflect different cultures. Consequently, each workshop would have a unique dynamic. Development of the workshop format was an evolutionary process. The first, more technical version was to begin with a brief discussion of deer ecology in Michigan’s agricultural region. Next, heterogeneous groups of farmers, hunters, and wildlife biologists would review the same deer crop damage scenario and identify potential management solutions that accommodate the interests of each stakeholder. After each group simulated its management ideas in POP-II, a computer program that depicts population dynamics over time, participants would present their management strategies. This approach could benefit wildlife managers in 2 ways. First, it emphasizes hands-on learning about deer ecology. Secondly, it would allow workshop facilitators to collect information for potential use during future management planning efforts. Despite 7O these strengths, the workshop planning team and I pursued a different format for 2 reasons I 1. My main objective was to develop a workshop format that enhanced communication among wildlife management stakeholders and not to collect data to evaluate how the workshop impacted the attitudes and belief system of the workshop participants. Although POP-II creates life tables using only deterministic variables (e.g., rates of natality and mortality, number of does and bucks legally harvested), group facilitators must be well versed in the program. Furthermore, computer simulations will fail if natality and mortality factors are not altered in a specific manner. Thus workshop participants would have little opportunity to manipulate the data and explore a particular deer management scheme. Lastly, the program has many assumptions that farmers and hunters may not firlly comprehend without prior reading and/or training. The adopted format contained 2 sessions that would accommodate 20 farmers, 20 hunters, and 10 wildlife biologists. In the morning, workshop presenters were to lead an interactive discussion of factors controlling deer population growth, the role of computer models in management planning, and ecological factors influencing crop damage. The afiemoon session was planned as a structured communication dialogue designed to help resolve conflicts (Peyton 1990). The original workshop agenda was to have 5 stages: Stage 1: For 30 minutes, homogeneous stakeholder groups (e.g., farmers and hunters) list their management concerns while wildlife biologists observe and answer technical ~ questions when asked. Workshop presenters record the discussion. Stage 2: Participants reunite for 30-45 minutes to present and clarify their concerns within a facilitated environment. 71 0 Stage 3: Mixed groups of 10 farmers, 10 hunters, and 5 biologists separately discuss and prioritize potential management strategies for each concern. Workshop presenters facilitate and record the 45-minute discussion. 0 Stage 4: Participants reunite to present their strategies. For 30 minutes, the facilitator uses questions and feedback to determine when all participants understand the presented strategies. 0 Stage 5: The facilitator leads a 45-minute discussion about the strategies to find consensus on as many recommendations as possible. A workshop presenter records the list of accepted and non-accepted strategies. Experienced facilitators were to lead and debrief each session to help participants comprehend the meaning and utility of ideas presented. Workshop organizers would use the participant-observation method to gather information from and about the participants and to identify how the discussions impacted the participants (Weiss 1998). Fifty farmers and 50 hunters were randomly selected for each workshop. The farmer’s master list included 353 individuals for the UP workshop and 546 individuals for the NLP workshop. The list was generated in the manner described previously for the focus groups. From the MDNR, a list of 2,768 persons that resided and hunted within each targeted region during the 1997 deer hunting season was collected (845 for UP, 1853 for NLP). This list was supplemented with 70 members of the UP. Whitetails Association. For both workshops, individuals that had been invited to the focus group were removed from the farmer and hunter master list before workshop invitees were randomly selected. In contrast, the wildlife biologists were not randomly selected. All 72 33 invitees (13 for UP, 20 for NLP) were members of the Michigan Chapter of TWS and were experienced with deer management issues. Letters that explained the purpose and organization of the workshop were mailed to all invitees 5 weeks before their respective workshop was scheduled. Pre-stamped, self-addressed postcards were included as the form of RSVP for farmers and hunters whereas the biologists were asked to RSVP by phone or electronic mail. A letter reminding the non-respondents of the workshop’s logistics was mailed 3 weeks later. Only 17% of the farmers and hunters said they could attend the program. Therefore, 2 media sources publicized the workshops among farmers and hunters in the targeted regions. The ANR Communication Services, located within the Department of Agriculture and Natural Resources at Michigan State University, sent a press release to all daily and weekly newspapers and radio service offices in northern Michigan. Likewise, Michigan Farm Radio Network featured a news clip about the workshop several times one week before the NLP workshop. Representatives with Michigan Farm Bureau and UP. Whitetails Association also helped recruit participants through personal contacts. Before the focus groups and workshops were conducted, the project methods were reviewed and approved by Michigan State University’s Committee on Research Involving Human Subjects in June 1998 and was re-approved in February 1999. The project IRB number was 98365, Category 1-A,C. 73 RESULTS Focus groups Attendance was low for both focus groups. Six individuals attended the hunter focus group, and 5 attended the farmer focus group. The hunters maintained a free- flowing dialogue throughout the 2 hours and were eager to share their success stories in community education about deer biology and management. Dissirnilarly, the farmers’ discussion on communication efforts in wildlife biology and management developed slowly. The participants were more willing to discuss their fi'ustrations with deer management than how they gathered information on the topic. Yet, after one participant explained to the group my intentions, the farmers entered a dynamic discussion on communication programs. Both groups learned about wildlife biology and management mostly through hunting experiences and/or working on their property and fmthered their understanding about current issues when reading sportsmen magazines (e. g., Field and Stream) and newspapers. Among farmers, the MSU Extension’s Bulletin, a publication in each county office, and personal discussions with Extension specialists also were popular credible sources of natural resources information. Additionally, 2 farmers and 1 hunter acquired information from MDNR biologists on a regular basis due to their proximity to a Wildlife Bureau office or voluntary involvement with agency projects. Among hunters, videos were perceived as a strong teaching aid in wildlife biology and management for people of all ages. Examples were amateur productions that documented a deer population study and organizational productions that discussed basic 74 principles of wildlife biology. The hunters also expressed a keen interest in radio shows on which local residents, paired with an expert, presented information about a wildlife management issue. Dissirnilarly, farmers were indifferent to TV and radio talk shows and favored print media as their source of information. In particular, farmers regarded MSU Extension Bulletin as the most objective source of wildlife management information. Not all focus group participants believed the media presented credible information on wildlife management issues. As mentioned above, the hunters viewed periodicals or TV/radio shows as credible and realistic only when its topic pertained to a backyard experience. These hunters agreed with the statement of one individual: “If I can’t afford to go out and hunt a Montana mule deer, or whatever it is, and this guy’s hunting it every day all over the country, writing stories about it and showing pictures — that is, in my way of thinking, a commercial endeavor. It’s not a wild endeavor.” When judging the credibility of a communication piece, farmers also relied on the story’s relationship to their own lives as well as the joumalist’s writing history. Thus, when reviewing information printed in an article, farmers weighted the information against the quality of articles that the journalist wrote in the past. Most focus group participants characterized the media as selling news and not reporting the facts. Except for one individual, participants in the farmer’s focus group believed the media provided biased information. In their opinion, the community is geared towards the interests of hunters. One farmer commented, “The media doesn’t state information given by the farmer. They care less whether we survive or not.” Similarly, the hunters noted a bias in the media. In their opinion, Lower Peninsula journalists were out of touch with the UP. culture and promoted the interests of urban 75 residents. The hunters viewed MDNR officials in the same light. In the words of one hunter, “They’re not going to advertise a herd as down in the UP. because they’re going to lose all these sales from down state people buying licenses.” All participants noted a preference for active learning experiences. They emphasized that state biologists must meet with local community members if firture management programs are to address local concerns. Specifically, one member of each group proposed a local meeting in which all stakeholders for a management issue participate to discuss their concerns and search for solutions. One hunter commented, “The thing that you’re doing -— you’re talking to the farmers; you’re talking to the hunters. If you can get that group together, from whatever areas they come from,. . .you’re going to have a very diversified talk from everybody across the UP. And that’s what you need; people from the Upper Peninsula. You can’t bring somebody in from the other side of the bridge to tell me what’s happening in the UP. It’s impossible.” Among the farmers, an individual claimed that 99% of the hunters did not comprehend the economic hardships farmers incurred from deer. A second farmer responded with the statement: “If we could both sit down from two different viewpoints and discuss our problems and concerns, I think we can arrive at a conclusion. . .or at least a compromise that we could all live with. But right now we’re yelling at the DNR, the DNR’s going back and forth, and they’re the middlemen. They’re taking all the rocks.” Lastly, one hunter called for a new system of local representation: “Increase local representation across the state by looking at it from a different scale; land area, not population density. And have more communication between those representatives. That would be the only way the UP. could get equal representation, money, or anything. Information would be shared automatically by word of mouth.” The farmers agreed with this hunter in the sense that land characteristics (e. g., growing seasons, soil fertility, and topography) should determine area boundaries for local 76 discussions. But in their opinion, all farmers, regardless of where they live, would share the same views for a wildlife management issue. Some members of each group expressed doubt that interaction among stakeholders would resolve human-wildlife conflicts. For instance, one hunter participated in a meeting in which hunters, farmers, and biologists discussed a deer crop damage problem in Marquette County: “In my opinion, the meetings didn’t accomplish anything. You got a chance to voice your opinion and that’s about where it laid. Whether it did any good on the other end, I don’t think so.” A second hunter said that the sharing of opposing perspectives leads to controversy and not resolution. But if the MDNR organized a structured meeting in which only focused discussions occurred, the same man thought the meeting might be more successful. Regardless of their views on a specific approach to conflict resolution, all participants agreed that the MDNR’s traditional public involvement techniques (i.e., public comment periods and town meetings) were unsuccessful and fostered distrust towards the agency’s decisions. According to one hunter, “They placate people and make them think they’re involved in the decision-making process when they’re really not.” Workshops Total attendance for the workshops included 6 farmers (5 at UP, 1 at NLP), 7 hunters (4 at UP, 3 at NLP), and 3 wildlife biologists (1 at UP, 2 at NLP). Three biologists led each morning session, and 1 biologist facilitated each workshop. Because of the low turnout, participants worked as a single, interactive group throughout the day instead of forming homogeneous stakeholder groups and small, mixed stakeholder groups 77 to initiate the aftemoon’s discussion of deer management. Therefore, my partners and I did not conduct the structured communication dialogue as described by Peyton (1990). Participants contributed to morning discussions on deer ecology (i.e., factors controlling deer population growth, role of computer models in management planning, and ecological factors influencing crop damage) by asking questions or offering information. During the aftemoon sessions, UP participants were congenial and talkative, but did not address a diversity of issues. Their concerns were limited to high deer density problems and boundary designations for Deer Management Units. In contrast, NLP participants were opinionated, but eager to tackle many issues (e. g., hunter recruitment, landownership changes). Rather than ask hunters, farmers, and biologists to prioritize their suggested strategies, we allowed the groups to freely discuss implementation challenges for management ideas. Facilitators maintained focused discussions in the afternoon by setting boundaries for acceptable comments. When listing concerns, participants were asked to complete the sentence, “I am concerned about deer management in [workshop’s county site] because... .” Examples of the responses among the Upper Peninsula participants were to maintain deer numbers within the land’s ecological carrying capacity and to continue feeding deer in the winter. Among the northern Lower Peninsula participants, concerns included: (1) hunter access in lieu of landownership changes; (2) hunter recruitment; and (3) finding consensus for deer herd objectives. Participants also identified specific stakeholder responsibilities when responding to the statement, “Recommendation strategies for each stakeholder.” Between the 2 workshops, suggestions for landowners included leasing their land to hunters and 78 improving their understanding of deer biology. For wildlife biologists, suggestions included managing deer at local scales, which may entail changing Deer Management Unit boundaries, and implementing a doe permit preference program for youth. One hunter strategy was to improve their hunting ethics. For all questions, the facilitators recorded comments on a flip chart using different colors to distinguish the respective concerns and responsibilities of each stakeholder. To encourage the sharing of information among the participants, the facilitators posted summaries of comments on the walls. Six out of 9 UP participants and 5 out of 6 NLP participants completed an open- ended questionnaire (Appendix D). All respondents said they had opportunities to share personal views on deer management, felt others valued their opinions, and appreciated the chance to express their management concerns and ideas. Over half of the respondents said the workshop changed how they thought about deer management issues. Most respondents said they would prefer more representation among farmers and hunters and to have the workshop conducted during winter. 79 DISCUSSION Focus groups Although too few focus groups were held to collect representative viewpoints of Michigan farmers and hunters, it does yield useful information for wildlife biologists. Three points resonated in both focus groups: (1) the media presents biased information; (2) farmers and hunters desire locally-based management decisions; and (3) farmers and hunters want to interact with each other when solving human-wildlife conflicts (specifically deer management problems). Communication researchers argue that credibility for a news report exists at several scales (Gaziano and McGrath 1986). Austin and Dong (1995) hypothesized that receivers of a news story judged the report’s credibility on 3 different scales: (1) representativeness - how well the report covered all aspects of the story; (2) accuracy — trustworthiness of the story; and (3) personal perspective — degree to which the receiver believed the reporter placed equal emphasis on each aspect of the story. Further investigation of the personal perspective stage may reveal an explanation of the farmers and hunters’ interpretation of the media. When a person is involved with a reported issue at a low level, s/he focuses on the news report’s peripheral qualities, such as the source’s credibility and likability, instead of the report’s content. Dissimilarly, when an individual is highly involved with a reported issue, s/he judges how well the report’s content aligns with her/his viewpoints. In both situations, the person is inclined to reject messages conveyed in a news report (Gunther 1986). 80 Such results were apparent in both focus groups; yet a more thorough investigation is necessary to prove the theory set forth in Gunther (1986). All participants of the hunter focus group were members of UP. Whitetails Association, an organization known for its active involvement in deer management projects. Consequently, these participants have a good, if not strong, understanding about various principles underlying deer management decisions. Hence, it is reasonable to assume the focus group participants accepted media content only if the news report reflected their own deer management beliefs. And although a representation of opinions on Michigan’s deer crop damage issue was sought among NLP farmers during the invitation process, their determination to discuss personal problems with deer also reflected their propensity to judge media credibility against personal beliefs. The farmers’ and hunters’ fervent interest in stakeholder interaction and increased public participation in land management planning mirrored the sentiments of focus group participants in the Northern Lower Michigan Ecosystem Management Project (NLMEMP) (Smith et al. 1999). Between 1995 and 1998, 53 focus groups were conducted by natural resource professionals to identify natural resources interests and concerns among NLP residents. These participants represented traditional (e.g., recreational, industry) and non-traditional (e.g., community service, land development) natural resource interests. When asked their views of public participation, the participants emphasized a need for greater representation among citizens and more local community and government involvement in public land decisions. Despite the plea among farmers and hunters in my study for more involvement in wildlife management decisions, several individuals expressed doubt that increased public 81 participation would be beneficial. Indeed, Moote et al. (1997) documented the difficulty of pursuing consensual agreement in conflict resolution when they reviewed a natural resource issue in southern Arizona In 1992, the Bureau of Land Management (BLM) initiated the collaborative resource management planning process to address a land acquisition controversy along the San Pedro River. The planning team existed for nearly 2 years until relational problems (e.g., arguing increasingly more, rising distrust in each other as committees made increasingly more decisions on their own) disbanded the group. Moote et al. (1997) attributed these problems, in no particular order of importance, to: (1) lack of confirmation among the citizens that the BLM would implement agreed upon alternatives; (2) no common understanding of the group’s goals for managing the San Pedro riparian system; (3) conflicting, deeply held values; and (4) the BLM making a decision without consulting the planning team. Although the San Pedro River case was unsuccessful in resolving a land-based conflict, it highlighted 3 key ingredients for any process that increases public participation. First, if deeply held values are at stake, natural resources professionals should consider resolution techniques that involve a third party, such as trained negotiator or judge. Secondly, a collaborative discussion demands structure to facilitate trust among participants as well as decision-making rules that preclude circular arguments and allow discussions to move forward. Lastly, all participants must understand how collected information will be used and who retains the authority to make final decisions (e.g., MDNR) (Moote et al. 1997). This final point should be clarified at the beginning of a collaborative meeting when the participants are establishing goals and objectives. 82 Workshops The workshop’s format allowed and encouraged participants to voice their Opinions. This may be attributed to the presence Of experienced, neutral facilitators and the workshop’s focus on information exchange, learning, and voluntary and inclusive participation. Facilitators assume a critical role in interactive processes, especially when natural resource professionals address value-laden conflicts (e. g., deer crop damage) (Svejcar 1996). They can maintain focused discussions within a neutral atmosphere and enhance participation. Consequently, stakeholders learn to trust one another and become willing to share personal interests, beliefs, and values. The workshops associated with this project demonstrated these points based on questionnaire results. For example, a respondent commented, “I thought the small group got to know each other quickly. . .participants quickly Opened up to the facilitators.” Voluntary participation also builds trust among stakeholders because it reflects an interest among natural resource professionals to hear stakeholder concerns (Steelman and Ascher 1997, Smith et al. 1999). Several participants demonstrated a personal desire to attend the workshops by driving long distances (e. g., 1 farmer at the UP meeting drove approximately 145 km). Since the workshops were interactive, participants were able to raise questions intermittently and solicit information, Often generating group discussions that lasted into the afternoon. Also, the aftemoon session was driven by the participants’ discussion about their deer management concerns. Indeed, the UP participants were reluctant to end the workshop. Based on the organizers’ Observations and questionnaire results, participants accepted the diversity of opinions while exchanging information. However, 83 focus group participants in NLMEMP (Smith et al. 1999) did not recognize the importance of information exchange and learning to the extent the planning tearn’s professional members did. Thus, a formal evaluation of the contributions of information exchange and learning may clarify this discrepancy. A limitation of some traditional public involvement techniques, such as surveys and public hearings, is the lack of two-way interaction in natural resources planning. Wondolleck (1985) referenced these techniques as the “black box” because government agencies retrieve the public’s input and incorporate data into natural resources policies and programs on their own. Therefore, citizens cannot discern how their comments influenced the decision-making process, nor learn why their concerns may have been ignored (Wondolleck 1985, Smith et al. 1999). Through an interactive workshop that emphasizes information sharing (i.e., Opposed to public surveys and commenting periods), stakeholders may identify more ways to resolve a management issue and have a greater incentive to compromise their interests in search of a mutually beneficial situation (Steelman and Ascher 1997). A weakness of the interactive workshops was that I did not reach my desired participant goal. This can be attributed to not querying farmers and hunters prior to sending invitations, the size of the target area for each workshop, and the timing of the workshops (i.e., early spring — potential conflict with farming and recreational activities). 84 CONCLUSION AND RECOMMENDATIONS 85 Inevitably, deer crop damage will persist as a common problem for Michigan farmers wherever the ecological and agricultural conditions are just right. In some instances, agricultural fields will attract deer because their adjacent woodlots and Openings contain excellent sources Of food and/or cover to accommodate the year round needs of deer. At other times, deer will forage in agricultural fields when dry spring and summer conditions prevent sufficient grth of herbaceous plants. Whatever the unique situation may be for a particular farmer, various management options exist. Since deer crop damage emerged as a serious issue in the 1970’s, wildlife biologists have encouraged farmers to border vulnerable, high cash crop fields (e.g., dry beans, corn) with 1.8 to 2.4 meter high fences that prevent deer from entering the field. Yet fence construction becomes cost-prohibitive as the size of the crop field increases. F arrners also have been advised to place various repellents along a crop field’s periphery (e. g., human hair, meatrneal, and commercial products) or use scare tactics, like explosive guns, to discourage deer foraging. However, these latter techniques produce short-term effects and address only the symptom of high, local deer densities (Scott and Townsend 1985, Dudderar and Marlatt 1989). To improve their understanding of and ability to reduce deer crop damage, wildlife biologists and farmers need a tool that identifies why deer use agricultural fields at varying levels across a geographical area. Dudderar and Marlatt (1989) emphasized the value of planting grasses and legumes along a crop field’s edge, maintaining Optimal levels Of woody browse in adjacent woodlots, and planting crops less vulnerable to deer damage (e. g., hay, barley) next to vegetation types that provide suitable deer habitat attributes. And if farmers are forced to plant a vulnerable field with a high value crOp, 86 they could minimize their losses by mixing habitat management suggestions with legal deer harvesting options (e.g., availability and use of special shooting permits) and the application of repellents. The PODD is a new tool that can help wildlife biologists and farmers manage deer crop damage at localized spatial scales, such as a particular crop field or farm. Additionally, the strength of this procedure abides in its ability to identify which components of a woodlot or opening encourages deer use of an agricultural crop field. Hence, wildlife biologists can use the PODD to standardize their assessment Of deer crop damage from a biological and landowner perspective. In this study, only 28.6% (n=2) and 14.3% (n=1) Of the dry bean fields in 1998 and 1999, respectively, were correctly classified while the correct classification Of alfalfa fields (by harvest) ranged from 0 to 62.5%. Also, 13.3% (n=2) of the worst cases for deer crop damage were correctly classified in 1998 compared to none in 1999. However, 33.3% (n=5) of the misclassifications in 1998 and 30.8% (n=4) Of the misclassifications in 1999 were slight (i.e., the PODD index value was 52 points away from the numerical border categorizing Observed crop losses). To improve the accuracy of the PODD, wildlife biologists should focus attention on the assessment procedure in 2 ways. First, the descriptive categories for each habitat attribute listed in the PODD should be reviewed to determine whether or not the categories need to be redefined. As mentioned in the Discussion, the descriptive categories for at least the understory growth variables in woodlots and the vegetation variables for Openings may be too narrow, thereby ignoring important variability within each cover type. 87 Separate from this procedural change, however, wildlife biologists should initiate additional projects to explore frrrther the ecological nuances Of Michigan’s deer crop damage issue. The results Of my study generated 3 interesting research questions. First, a shortfall of my project was the exclusion Of deer population dynamics when attempting to predict relative crop losses attributed to deer foraging. Future projects should identify how many deer cause significant crop losses and examine which characteristics of deer population dynamics could help wildlife biologists describe or predict relative levels of crop loss. Possible characteristics of deer population dynamics that wildlife biologists could test in the PODD are the number of deer beds counted in alfalfa fields or Openings, number of does observed during winter versus spring, and how many deer-vehicle accidents occur within a county or township. These future projects should also explore how each characteristic should be weighted when computing the final PODD value. Secondly, if deer concentrate in areas with relatively high quality habitat (i.e., year round food is available with thermal and security cover nearby), how long does it take for use of the area at a high deer density to lower the habitat’s quality? Consequently, how long does it take for deer to establish new movement patterns and to occupy new home ranges that Offer higher quality habitat? The answers to these questions may shed new light on the relationship between habitat quality and deer crop damage. Lastly, a long-term habitat study should be implemented to determine if variations in the climate do indeed yield significant differences in deer use of agricultural fields. According to my data, the relatively dry spring and summer months Of 1998 led to higher levels of crop loss when compared to wetter growing season of 1999. If climatic 88 variation does indeed influence deer use of agriculttu'al fields, then the PODD should include seasonal or annual precipitation as a weighted variable. Regarding the human dimension component of Michigan’s deer crop damage issue, wildlife biologists often encourage farmers to increase hunter access to their properties with the goal Of lowering deer densities on individual farms. But what if many farmers within a community repel participation in a hunter access program? Their disinterest may stem from a satisfaction with the number of friends and family members already hunting on their property, liability concerns, and/or a distrust Of strangers. Such was the case for 17 of the 19 landowners participating in this study’s deer habitat evaluation component. Suppose further that farmers and hunters within a particular township or county have little Opportunity to discuss their deer management concerns with one another and clarify their misunderstandings? Lastly, suppose the same group of individuals does not have accurate biological and ecological information about its community’s deer population — information that may help reduce conflicts between farmers, hunters, and deer? Wildlife biologists may be able to address these problems proactively by using the interactive workshop format presented in Chapter 2. The workshop could be organized for a local community to unite stakeholders and foster information exchange and learning so that common management concerns are identified. The purpose of the interactive workshop is to let wildlife biologists simultaneously disseminate information and identify public concerns central tO a wildlife management issue. However, it may be unnecessary to include an educational session if the audience is familiar with many biological principles planned for discussion. In that case, during the planning stage workshop organizers need to identify the questions that 89 the participants may ask but use only the structured communication dialogue when implementing the workshop. As questions arise, the organizers can use prepared materials to discuss less understood principles. Regardless, it is important that workshop organizers identify which issue(s) workshop participants may want to discuss, which biological principles the participants may and may not understand, and what comments participants may make. Many different agencies and organizations within a community can help share information between wildlife biologists and stakeholders. Yet, it may be difficult for professional organizations to involve some members if the members cannot dissociate themselves from their employer or if they must devote extracurricular time to a workshOp that addresses a contentious issue. TO avoid such dilemmas, professional organizations should form a special committee to discuss these concerns and explore ideas that motivate members to represent their organization in a community-oriented program. Also, when planning an interactive workshop, professional organizations should seek the support of administrative leaders in the government agencies that have a stake in the wildlife management issue and employ members of the organization. If these agency leaders perceive how the interactive workshop may help their agency better manage a wildlife issue, the workshop organizers can work through the agency administrators to encourage members to participate in the workshop. When piloting the interactive workshop, I had sub-Optimal participation among farmers, hunters, and wildlife biologists. Such results can be minimized if workshop organizers: (1) visit with representatives of each invited stakeholder group to identify the best time of the year for scheduling the program; (2) create a master list for each 90 stakeholder group (each list should contain individuals who do not professionally represent the interests Of their respective stakeholder group), query everyone about their interest in the workshop, and randomly submit invitations to only those with a stated interest; and (3) non-randomly invite leaders and employers of the various organizations that represent the interest of each stakeholder (e.g., Michigan Farm Bureau, UP Whitetails Association). Lastly, wildlife biologists and other natural resource professionals confront the daily challenge of understanding the diverse beliefs and values of stakeholders and communicating sound, scientific management principles used in decision making. The interactive workshop format I piloted in Michigan may be a useful mechanism that helps all natural resource professionals meet this challenge. However, the workshop should be tested first to identify to what extent participants: (1) improve their understanding of the human-wildlife conflict at hand; and (2) change their behaviors to initiate partnerships with other stakeholders (e. g., number of farmers leasing their land to hunters before and after the workshop). 91 LITERATURE CITED 92 Austin, E. W. and Q. Dong. 1994. Source versus content effects on judgments Of news believability. Journalism Quarterly 71: 973-983. Beier, P. and D. R. McCullough. 1990. Factors influencing white-tailed deer activity patterns and habitat use. Wildlife Monograph 109: 1-51. Bender, L. C. and J. B. Haufler. 1987. A white-tailed deer HSI for the upper Great Lakes region. Michigan State University, Department of Fisheries and Wildlife, unpublished. Braun, K. F. 1996. Ecological factors influencing white-tailed deer damage to agricultural crops in northern Michigan. M. S. thesis, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 258 pp. Campa, H., 111, S. R. Winterstein, R. B. Peyton, G. Dudderar, and L. A. Leefers. 1997. An evaluation of a multidisciplinary problem: Ecological and sociological factors influencing white-tailed deer damage to agricultural crops in Michigan. Transactions Of the North American Wildlife and Natural Resource Conference 62: 431-440. Corder, P. G. 1984. Soil Survey of Mecosta County, Michigan/U. S. Department of Agriculture, Soil Conservation Service in cooperation with Michigan Agricultural Experiment Station. 155 pp. Corteville, L., K. Mattson-Hansen, H. Hill, B. Moritz, and S. Winterstein. 1999. A new look at an old index: Predicting white-tailed deer harvest from deer-vehicle accidents. Abstract for Poster Session at the 61St Midwest Fish and Wildlife Conference in Chicago, Illinois. Decker, D. J ., T. L. Brown and B. A. Knuth. 1996. Human Dimensions Research: Its importance in natural resource management. Pages 29-47 in A. W. Ewert, ed Natural resources management: The human dimension. Westview Press: Boulder, Colorado. Dudderar, G., J. Hanson, J. B. Haufler, R. B. Peyton, H. H. Prince, and S. R. Winterstein. Michigan’s deer damage problems: An analysis of the problems with recommendations for future research and communication. Deer Damage Committee. Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 97 pp. Dudderar, G. and S. Marlatt. 1989. Controlling deer damage. Michigan Farmer. May 6: 21-23. F agerstone, K. A. and W. H. Clay. 1997. Overview of USDA Animal Damage Control efforts to manage overabundant deer. Wildlife Society Bulletin 25: 413-417. 93 Fritzell, P. A. 1998. A survey of Michigan agricultural producers’ attitudes, perceptions, and behaviors regarding deer crop depredation to fruit, vegetables, and field crops. M. S. thesis, Department Of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 233 pp. Gaziano, C. and K. McGrath. 1986. Measuring the concept of credibility. Journalism Quarterly 63 (3): 451-462. Gladfelter, H. L. 1984. Midwest agricultural region. Pages 427-440 in L. K. Halls, ed. White-tailed deer: Ecology and management. Stackpole Books: Harrisburg, Pennsylvania. Gunther, A. C. 1986. Extremity Of attitude and trust in media. Journalism Quarterly 65: 279-287. Higgins, K. F ., J. L. Oldemeyer, K. L. Jenkins, G. K. Clambey and R. F. Harlow. 1994. Vegetation sampling and measurement. Pages 567-591 in T. A. Bookhout, ed. Research and management techniques for wildlife and habitats. The Wildlife Society: Bethesda, Maryland. Horton, R. R. and S. R. Craven. 1997. Perceptions Of shooting-permit use for deer damage abatement in Wisconsin. Wildlife Society Bulletin 25: 330-336. Kathlene, L. and J. A. Martin. 1991. Enhancing citizen participation: Panel designs, perspectives, and policy formation. Journal of Policy Analysis and Management 10(1): 46-63. Kerr, J. A. and F. W. Trull. 1928. Soil Survey of Isabella County, Michigan/U. S. Department of Agriculture, Bureau of Chemistry and Soils in cooperation with Michigan Agricultural Experiment Station. 1183-1202 pp. Manson-Hansen, K. 1998. Integration of archery white-tailed deer (Odocoileus virginianus) harvest data into a sex-age-kill population model. M.S. thesis, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 95 pp. Minnis, D. L. 1996. Cultural carrying capacity and stakeholders’ attitudes associated with deer crop damage issue in Michigan. Ph.D. thesis, Department Of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 386 pp. Moote, M. A., and M. P. McClaran. 1997. Viewpoint: Implications of participatory democracy for public land planning. Journal Of Range Management 50: 473-481. Moote, M. A., M. P. McClaran, and D. K. Chickening. 1997. Theory in practice: Applying participatory democracy theory to public land planning. Environmental Management 21: 877-889. 94 Morrison, M. L., B. G. Marcot, and R. W. Mannan. 1998. Wildlife-habitat relationships: Concepts and applications (Second Edition). The University of Wisconsin Press, Madison. 435 pp. Murphy, R. K., N. F. Payne, and R. K. Anderson. 1985. White-tailed deer use of an irrigated agriculture-grassland complex in central Wisconsin. Journal Of Wildlife Management 49 (1): 125-128. National Oceanic and Atmospheric Administration Climatological Data: Michigan annual summary. 1998. National Climatic Data Center, Asheville, North Carolina. National Oceanic and Atmospheric Administration Climatological Data: Michigan annual summary. 1999. National Climatic Data Center, Asheville, North Carolina. Nelson, C. M. and T. F. Reis. 1992. Michigan deer crop damage block permit study. Transactions of the 57’11 North American Wildlife and Natural Resources Conference. 89-95. Nixon, C. M., L. P. Hansen, and P. A. Brewer. 1988. Characteristics of winter habitats used by deer in Illinois. Journal Of Wildlife Management 52 (3): 552-555. Palmer, W. L., G. M. Kelly, and J. L. George. 1982. Alfalfa losses to white-tailed deer in Michigan. Wildlife Society Bulletin 10: 259-261. Peyton, R. B. 1984. A typology Of natural resource issues with implications for resource management and education. Michigan Academician 17: 49-58. Peyton, R. B. 1990. Leader’s Guide: Communication and dispute resolution for fisheries and wildlife managers. Responsive Management Project of the Western Association of Fisheries and Wildlife Agencies, Tallahassee, Florida. Prior, R. 1983. Trees and deer: How to cope with deer in forest, field, and garden. B. T. Batsford Ltd., London. 544 pp. Rasmussen, G. A. and M. W. Brunson. 1996. Strategies to manage conflicts among multiple users. Weed Technology 10: 447-450. Schneider, 1. 1960. Soil Survey of Montcahn County, Michigan/ U. S. Department of Agriculture, Soil Conservation Service in cooperation with Michigan Agricultural Experiment Station. 60 pp. Scott, J. D. and T. W. Townsend. 1985. Methods used by Ohio growers to control damage by deer. Wildlife Society Bulletin 13: 234-240. 95 Selin, S. and D. Chavez. 1995. Developing a collaborative model for environmental plamring and management. Environmental Management 19: 189-195. Sitar, K. L. 1996. Seasonal movements, habitat use patterns and population dynamics of white-tailed deer in an agricultural region of northern lower Michigan. M. S. thesis, Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 130 pp. Smith, P. D., M. H. McDonough, and M. T. Mang. 1999. Ecosystem management and public participation: Lessons fi'om the field. Journal Of Forestry 97 (10): 32-38. Sokal, R. R. and F. J. Rohlf. 1981. Biometry: The principles and practice of Statistics in biological research (Second Edition). W. H. Freeman and Company: New York. 859 pp. Steelman, T. A. and W. Ascher. 1997. Public involvement methods in natural resource policy making: Advantages, disadvantages and trade-Offs. Policy Sciences 30: 71-90. Stoll Jr., R. J. and G. L. Mountz. 1983. Rural landowner attitudes towards deer and deer populations in Ohio. Ohio Fish and Wildlife Report 10. 18pp. Stout, R. J ., D. J. Decker, B. A. Knuth, J. C. Proud, and D. H. Nelson. 1996. Comparison of three public involvement approaches for stakeholder input into deer management decisions: A case study. Wildlife Society Bulletin 24: 3 12-3 1 7. Svejcar, T. J. 1996. What are working groups and why should scientists be involved? Weed Technology 10: 451-454. Vecellio, G. M., R. H. Yahner, and G. L. Storm. 1994. Crop damage by deer at Gettysburg Park. Wildlife Society Bulletin 22: 89-93. VerCauteren, K. C. and S. E. Hygnstrom. 1998. Effects of agricultural activities and hunting on hoe ranges of female white-tailed deer. Journal Of Wildlife Management 62: 280-285. Weiss, C. H. 1998. Evaluation (Second edition). Prentice Hall, Upper Saddle River, New Jersey. 372 pp. Wondolleck, J. 1985. The importance of process in resolving environmental disputes. Environmental Impact Assessment Review 5: 341-356. Wotecki, C. E. 1997. Surveying the road ahead for Extension. Transactions of the North American Wildlife and Natural Resources Conference 62: 22-29. 96 APPENDICES 97 APPENDIX A The Predictor of Deer Damage The instructions detailed below how to use the PODD when evaluating cover types immediately surrounding a crop field. The first page is used to collect information about the farmer’s land use Objectives and basic information about the crop field selected for evaluation. On pages 2 through 5, several attributes Of deer habitat are listed in the left column, and each attribute has 3 or 4 descriptive categories. When evaluating the quality of deer habitat in each adjacent woodlot and Opening, select only 1 category that best describes the entire area. Consequently, several plots or ocular estimates (by professional biologists) will have to be used to assess the overall characteristics of each deciduous and coniferous woodlot and Opening. Use the worksheet on page 8 to record information for each plot or stopping point for an ocular estimate. Also, refer to the vegetation chart on page 6 to estimate the percentage Of herbaceous and shrub cover in woodlots and openings. The underlying premise of the PODD is that as the quality of deer habitat increases, the possibility of deer using the adjacent crop field also increases. Therefore, each descriptive category has its own index value to represent its ecological significance to deer as a source of food (spring and fall/winter) and cover (thermal and security). Each category was developed using literature on deer ecology (Gladfelter 1985, Bender and Haufler 1987) and vegetation data for agricultural regions of northern lower Michigan (Braun 1996) and central lower Michigan. The index values for each 98 descriptive category are listed on page 7 according to the numerical column associated with each descriptive category. On pages 2 through 5, column 1 (Cl) is farthest on the left and represents deer habitat of relatively high quality. Column 4 (C4) is farthest on the right and represents deer habitat Of relatively poor quality. After evaluating each cover type, find the index value for the selected descriptive categories. Sum the index values across all cover types evaluated and report the final PODD value at the right-hand bottom of page 7. To predict at what level deer may forage in the evaluated crop field, compare the final PODD value to the range Of scores representing low, medium, and high deer damage (page 7). Instructions 1. Visit with the farmer to complete all questions on page 1. 0 Development equates paved and dirt roads intended for public travel (one-lane dirt roads intended for farm use do not count), homes, assorted farm buildings, and businesses. 0 Defining land cover types surrounding a crop field: If a public road borders the crop field, do not count land features (e. g., agricultural field, home) found across the road. However, if a windbreak or fencerow <5 m in width borders the crop field, do count the cover type located on the other side. Note, this question has associated index values. See page 6. 2. Deciduous and coniferous woodlots adjacent to crop field (pages 2 and 3). 0 Randomly select at least 3 plots or stopping points for ocular estimates to assess deer habitat in each adjacent woodlot. More than 3 plots or ocular points may be 99 necessary if differences among plots make it difficult to select 1 descriptive category for at least 1 habitat attribute. o For each plot or ocular point, identify the type of tree dominant in the overstory. All other habitat attributes must be measured according to the selected tree species. For example, if maple (and beech) is the tree dominant in the overstory, basal area, canopy cover and size class should be assessed for only maple (and beech) trees. 0 Use a basal area prism to select the best descriptive category for basal area. 0 Understory vegetation: Using the vegetation chart on page 6, select the letter (A, B, C, or D) that best estimates the percentage of herbaceous and shrub cover for each plot or ocular point. 3. Agricultural fields adjacent to the crop field (page 4). o For each agricultural field that is identified next to the crop field, determine its individual percentage as an adjacent cover type. However, if at least 2 fields are planted to the same crop, report the percentage sum. For example, if 2 of 3 adjacent agricultural fields are planted to corn and 1 field adjoins 10% of the farmer’s field and the other cornfield adjoins 20% of the farmer’s field, report 30% as the percentage sum for corn. 4. Openings adjacent to the crop field (page 4). 0 Use the vegetation chart to select the letter that best represents herbaceous and shrub cover for each Opening. Rarely are more than 3 plots or ocular points necessary to define each type of Opening. 5. Development adjacent to the crop field (page 5). 100 0 Recall that one-lane roads intended for farm use are not recorded. A side note is optional. 0 For each development site that is identified next to the crop field, determine its individual percentage as an adjacent cover type. But, as with adjacent agricultural fields, report the percentage sum if the same development type is selected at least twice. 6. Cover types within 10.5 km of the crop field (page 5). 0 With the aid of the farmer and/or aerial photographs or by driving through the community, identify the percentage of each cover type within 10.5 km of the field. 101 K. B. Reis, 2000 Predictor of Deer Damage (PODD) Farmer: City: Township & Section: DM U: Deer management activities during previous year 1. ins-95s Did you allow deer hunting? Y N If yes, list number of hunters: Did you lease your land for hunting? Y N Number of summer shooting permits issued (filled): Number of block permits issued (filled): Did you use a lure crop to deter deer depredation? Y N If yes, name crop type(s): DO you use any other techniques to reduce deer crop damage? Y N If yes, name technique(s): Did you conduct deer baiting during the fall? Y N . Did you provide supplemental feed during winter or throughout the entire year? Y N If yes, name type Of feed: Distance to adjacent landowner's supplemental feed (if applicable): Farm and field characteristics 1. 2. Farm size (acres): Owned Rented Number of land parcels in farm: Soil type: 3. Crop type: 4. Approximate field shape and size (acres): 99089:" Circular Square Rectangle Oval Convoluted 1-25 26-50 51-100 >100 Number of cultivations (or cuttings for alfalfa): Do you fertilize? Y N DO you irrigate? Y N Management practice: Conservation or no tillage Tillage Define land types immediately surrounding crop field: All woodlands All agriculture All Openings Mixture: Woodland >50% 26—50% 525% Agricultural 10-30% 31-50% <10% or >50% Opening 10-30% 31-50% <10% or >50% Development 0-10% 1 1-25% 26-50% >50% 102 E OmA Es mvv :oumuocow0m .xoomv asses new E omtcm E 3.: 23.0 233303; 003. o m 35 3-3 a8 «NO moo; =~Emh3om Bnfimgmm $3-8 $2-8 «ENE m.m~- ”N3 226 one can: use Eooom ens can: Em 5.6 «ENE m.wm . 9mm 82w Conggam 83:25 0:532 e\oOBA asses swam “@5285: 8:503 ”20w mob can 8:503 503309 00535 A0 ”Amvbnfisa 8:083 ”coufiowg b05895 c0 Amronfise «0:083 ”333 mum—0 03m A0 Amiga—E 8:083 ”0030 .3280 3 ”Gross—S: 3:083 ”~03 Baum Am numb .003 macawmuuu 33.35% .3 @035 “3309: \e 030:3” .N 3.5 :0&< .3. ”860% 3828 £0.30 Ange—=5: 8:083 ”sameness 3 00.: naeafiucn 233333 830% .~ 20¢ no.6 3 2.00.53 85:02:» nae—.2009 "Gm—O.— ccoN £3— .m d— 103 E omA E oméa E 3-: E o. W .333 333.3%? 3.. Q U m < A80 mvv 3.0 3-3 A30 mmAv €583.33 mug—05v sougcowom moo; _EEmho—om 8.25.38 23.22 Axbmv $318 32.9. $5? «5.8 Nww «SN... m.£-m.w «ENE w.vm-c.£ «ENE «..va ”Boa 83 23° .8. 2: E53» 3.86 3.338 mo .3832 c “arenas: 6683 ”Bum no... 35. 3:303 5953 353me 8 ”wrong: .2383 32.8%? 33803.5 3 ”$538.... 3:303 ”293 as. 35 6 ”wrong: 3:303 ”..o>oo .3280 3 Antenna: 3:303 ”no.3 imam Am 3%.... 3.... 2333.333 33......33 .3 33.33 333... x... 3333...“. .N gunfight 323m 3335.3 2.3 3832 .6283. Ba 333 .o 8. £3. E 8.2... 380 323 Eofioz 628% 328—8 286 Aaron—Ed: 8:303 .3829... .... 3... 3333.833 33......33 .33». .~ 29: 3.3 3 33.5... 3:23.33 389:.30 “DA—O.— oocN £3— .m 2 104 E OmA E OmA 550 E omém 82-: SSW 23% =23Em§ m3. 8 cmém amaze c\comA .8 o\oo~v 88m 2%? 280m E m < E mm;— 82W EH 02 own—E. soumEomnoU $3; m $3-3 32:5 08H :95 £32 500 Amronfisc mango ..EmKREQ ~33 Mthmmc ammémo muzfiflQ .N Antenna: mfiaonO ..MS§&Q.\° 2%“ .~ Soc :93 3 2.3.5:. «ME—:50 ”arena? 22m ”Suwanee $3398 ~23 Sufimu 383%: =mm>cme mucuBQ .V ”@382 Eva ..«ziofixt 8.85% “mmEmMuzu: .m. ”arenas: 22m “@qu Vuuofiuu 53 VS: “zmucxnuka 38km .N Aaron—SE 20E ”not Nuxsfiautwaxo MQQ .~ Sou 30.5 3 2.35.23 nus—ow Eng—_SEMA‘ "GA—C.— ooom .30.: .m .M 105 a\omNA o\oomA .5 o\oo_v o\..OmA o\eomw E omA E Omém °\comA Xomém mmofimam waxy—:5 8.3 3283‘ o\omm-_ £009; o\oOm-~N $8-0m 83-: $3-: 552QO .xb ..EmERSQSQ .V $8-2 ..wszumo .m §~v ”353.2% .m go? ..Euficok .N 23. 9:9 ... .5. 2: as? 8.5 2.3 Aaron—SE EoEQEoEQ ..EmKQoLu ~23 32: E 2 W #39th $623. 2833 8§3Q .m ”Anton—8:: 3088:33— fiNmKQoLu 8 383563 £0 _ -o EmERSQSu hfiufimuflo “5.2mm .N ”Gianna: EoEQBoEQ 93M ..EmEmGESfiKQ 85 83mm. . N 22.. 998 3 2.3.5:. 2.25.2969 "an—O.— 88 £3. .m .M 106 A05 .550 9880 googfiom 873. mhém omém mNé < m U Q m m U Q mmé cmuom mhAm 02-3. 25.0 553995 (%) .1 9110;) qnxqg ooo~ .23— .m .v. 107 K. B. Reis, 2000 PODD: Index values for descriptive categories Page 1 and page 5 Define land types immediately surrounding crop field and within 10.5 km: Woodlands C1 = 1.0 C2 = 0.5 C3 = 0 Agricultural C1 = 1.0 C2 = 0.5 C3 = 0 Openings C1 = 1.0 C2 = 0.5 C3 = 0 Development C1 = 1.0 C2 = 0.5 C3 = 025* C4 = 0 * C3 = 0 on page 5 (within 10.5 km) Page 2, deciduous woodlots adjacent to crop field Dominant tree C1 = 1.0 C2 = 0.7 C3 = 0.5 Basal area C1 = 1.0 C2 = 0.75 C3 = 0.3 C4 = 0 Canopy cover C1 = 1.0 C2 = 0.5 C3 = 0.25 C4 = 0 Size class C 1 = 1.0 C2 = 0.7 C3 = 0.3 C4 = O Understory vegetation C1 = 1.0 C2 = 0.75 C3 = 0.5 C4 = 0.25 Distance fi'om field C1 = 1.0 C2 = 0.9 C2 = 0.8 C2 = 0.7 Page 3, coniferous woodlots adjacent to crop field Dominant tree C1 = 1.0 C2 = 0.8 C3 = 0.4 Basal area C1 = 1.0 C2 = 0.85 C3 = 0.6 C4 = 0.3 Canopy cover C1 = 1.0 C2 = 0.5 C3 = 0.25 C4 = 0 Size class C1 = 1.0 C2 = 0.7 C3 = 0.3 C4 = 0 Understory vegetation C1 = 1.0 C2 = 0.75 C3 = 0.5 C4 = 0.25 Distance from field C1 = 1.0 C2 = 0.9 C2 = 0.8 C2 = 0.7 Number of conifer stands C1 = 1.0 C2 = 0.6 C3 = 0.25 C4 = 0 Page 4, agricultural fields adjacent to crop field Crop type C1 = 1.0 C2 = 0.8 C3 = 0.6 C4 = 0.3 Percent next to crop field C1 = 1.0 C2 = 0.5 C3 = 0 Management practice C1 = 1.0 C2 = 0 Distance from field C1 = 1.0 C2 = 0.9 C2 = 0.8 C2 = 0.7 Page 4, openings adjacent to crop field Opening type C1 = 1.0 C2 = 0.75 C3 = 0.5 C4 = 0.25 Distance from field C1 = 1.0 C2 = 0.9 C2 = 0.8 C2 = 0.7 Page 5, development adjacent to crop field Development type No index values necessary Percent next to field C1 = 1.0 C2 = 0.5 C3 = 0.25 C4 = 0 Distance from field C1 = 0.3 C2 = 0.2 C3 = 0.1 C4 = 0 Damage prediction ranges based on final PODD value 0-12 13-19 20-26 Low Medium High Final PODD Value 108 A32 .8 N a. N 28E: 8.4 6%? N2: 389$ 283$ 28¢.ch 28¢.ch N “82$ 8. a 3.3m 3.3% $.ch £3wa :25de N “stem 83w 5%? a2: EESTB Ans—«SAC Ao:_a>-mv Ans—95¢ Aoégiv vignm 029d 2:87am o=_a>-m 0395* Sam SN VENNN am ~253th sacs: .8388.— 86 30> 82 .8333? “88.3 .6 cassava 20E 86 $83 _ 03¢ 6:98 $8.8 N a N waste: a: 8.: fig“. 32 M “8:3: is? 225.3 3898 58.8 N “stem 8.5%. 93:. 8. a fig... 8.: Augusta Ans—9A6 33973 Aoégiv Ava—ETB 020>un~ cup—Erna 03a>nm o=~w>unm odd—gum can Boga: 2 .3 man x5 3% an N ENNN am 8.6 30> 3.2%.: co 82 Beam .8 a so: 82 bosom .88 23 Mae 5 8322 533 3on mo mac—om £32m E “swag no.8 bow 9:28:55 mofiatg “55a: =~ wan—5:8 358 05 How $02a>¢ Ea 83m>-..b 39:28 .588..on 8595 23:98 05 MO 338% A 035. m .. m 592mm: 109 @898 A889 :80de :80de afimc vav 3.6mm 8.3% .3qu bQ 33 A883 8898 A338 :08de A889 amdw coin mmdm $.33 KAm fl 38% ..9Q «23 Ao=_m>-nc $33-3 Ans—«>46 Aofigiv 32973 Aoagxc 253$ o=_m>-m o=_m>-m 263i ofigi 023$ «£236 Nagaaaotww xmcmfitmm 8.3x 3ng Nab—0mm com QED 80> owvm 205 Mo BEST—om 20E 35 8:50 - 20E 33. “8:80 .82 Ba $2 a 8322 532 1350 .«O View :89 be 5 owag no.8 Hoot wfiocoscfi mus—amine, “SEE =a waning—co 3on 05 .5.“ Amofigd 98 mofigiv mix—«SN 56.0.2on 8598a 03:38 05 mo $38M .E 033. = u m— Nun—2mm: 110 A888 383$ 38.8 8.2 8.8 3.0 38me 383$ 388$ N Eta: 8.3 8.3% .880. 388$ 38.8 283$ Nesta: 3.3” 8.8 3.3% .880 0387.: 338756 33873 $3935 OBSTR o:_a>n.m o:_a>..m 05—97"— B0E§~0>0Q M83080 NESESLMV 8:009: no.5 2% 05.38 03$ 83 .8200 58.8 288.8 283$ 38.8 8.2 8.: 8.88 8.: 280 on 888.8 288.8 283$ 38.8 28E: 8.2.. 2.: 8.: 8.: .880. 283$ 38.8 808.8 38.8 383$ 2898 38.8 Nesta: EN 8.: 8.2 8.3 8:. v3 23. .88... A030>-nv A030>-nv A039,.-nv €33-56 A03u>-nv $387.6 $3855 a039,»: 03a>-n 030>-n 030>-n 03a>-n 03a>-n 030>-n 030>-n 033i 30Eno~0>0Q $53080 ~0u8~30fMV 2589: 30E 2 00:80.5 00.3 3.3% M0: 58.2033 00.80%. no.5 008 20E .6 “829: 8558820 83m 08323 .33 3 3w30az .2033 33:00 3 020m 302 be 0:0 £0.20 9 308 .80 m0naowo§_ 3 nouwomnnna 3 no @093 _038 ODOn 05 En $0357n 38 $3036 30303 22300023.. 03338 033:3 05 no 330m :— - m X—n—Zmnnd. 111 308.8 83.8 380$ $88 2:88 283$ 8.2 8.0 8.8% 03. 8...: 6.88 .280 .00 32:8 $08.8 $08.8 388$ m 0 0 085: 8.0 3.0 25. 9.2 _ .880 330?“: A030>dv A030>Ac $300758 $3856 A030>Ac 33007.6 03007.3 030>-m 030>-m 03007.3 030>-.m 030?...3 030?"— A0000 ~23 m.mmv ~=0E§~0>0Q $530030 ~005302WV M30080 ~008~00T~MV 038000: no.5 8900 .808 0 230 ”5 3.8 25 380 .8050 0000 220 3850 .8: a 50222 0033 _00300 3 020m 0003 D0 0:0 0.20.20 8 3000.80 00200030— 3 03:03:50 3 no 00009 _083 QQOm 05 .28 A0030>d 000 003038 303000 2300030.: 003508 03333 05 Mo 033003 .2 030B >~ .. m van—Zak: 112 APPENDIX C Agenda for Educational Workshop 9:00 A.M.-4:00 P.M. Main facilitator: Ben Peyton (10 April); Shari Dann (17 April) Workshop leaders: Rique Campa and Kathryn Reis (10 and 17 April); Dean Beyer (10 April); Scott Winterstein (17 April) I. Orientation and Introductions Kathryn briefly explains why the workshop was organized: I Responding to a previous study on deer crop damage at Michigan State University. I Partnering with the Michigan State Chapter of The Wildlife Society to pilot the national communication program Wildlife Information Network (WIN). Next, the leaders/facilitator will introduce themselves followed by the participants (providing name tags, too). Hunters should identify: I town/county of residence, I cover type on which they hunt (private or public?), I if they own land specifically for hunting, and I if they hunt only in the U.P. Farmers should identify: I town/county of residence, I size of their farm, I crops they produce, and 113 I whether or not they hunt (and if so, where). (15 minutes—depends on number of participants) II. Objectives of workshop Kathryn discusses the workshop’s objectives (5-10 minutes)... 1. When the farmers and hunters leave this workshop they will: I appreciate what factors control the size of a population and how biologists manage these factors via direct (hunting) and indirect (habitat management) measures. I appreciate the role and limitations of computer models in making deer management decisions. I understand 2 concepts of population dynamics—resiliency of deer and harvestable surplus production among does and bucks. 2. When the farmers, hunters and biologists leave this workshop they will have an improved understanding how the quantity and quality of deer habitat interplay with population dynamics to influence deer crop damage, especially in landscapes that contain a diversity of vegetation types used by deer. 3. Using a structured communication format farmers, hunters, and biologists will interact to discuss their deer management concerns and objectives for a specific DMU and find general agreement on how to address those concerns /objectives. 4. Evaluating the workshop’s ability to serve as a model that wildlife professionals/agencies could use to enhance communication between biologists and all stakeholders for a local/state wildlife management issue. 114 I Purpose of the workshop model is to share information; not develop management strategies that the group would submit directly to the MDNR. III. Educational session Activity: Group discussion about population control factors and modeling (1.45 to 2 hours) Dean/Scott will ask the audience what factors influence the size of a local deer population (births, deaths, immigration and emigration). Hunting is the only way biologists can directly control population growth, but this cannot be viewed in isolation. Before biologists set hunting quotas, they must know what the buck:doe ratio is and at what rate does reproduce. Do does have 1, 2 or 3 fawns each year? Does this reproductive rate remain the same year after year? What influences the reproductive rate? Secondly, biologists must consider potential weather conditions. Severe winter storms can lead to direct mortality of deer, or a mild winter may cause unusually large numbers of deer to survive. Also, the lack of high quality, nutritious food during long winters and/or dry springs could induce starvation among deer or poor reproduction among does (i.e., miscarriages, lower reproductive rates, increase risk of fawn mortality). Finally, biologists must understand the movement patterns of deer. How many deer migrate to seasonal home ranges and what distances do they travel? Are the same number of bucks and does migrating? (20 minutes) Rique/Scott will present the results of Scott and K. Sitar’s research in the northern Lower Peninsula. He will describe the deer that migrated to summer and winter 115 ranges (when they migrated, average distances they traveled), and compare habitats (type and size of home ranges) and daily behaviors between migratory and non-migratory deer. (20 minutes) Before Dean/Scott introduces POP-II, he will ask the participants how some of the above information is considered when managing a population. Next, he will address the following key points about modeling programs (20 minutes): I Type of data used - variables, sources I Tool for long-term planning — liken to house budget plans and farming (how farmers determine planting regime for their farm based on previous year’s production levels and economic return) I Simulations help identify strengths and weaknesses of data set I Limitations of POP-II: Simplified picture of real world — must know assumptions before interpreting data; does not accommodate density dependent factors or stochastic events (at least not automatically) I Assumptions of POP-II: No immigration or emigration 5 minute break Kathryn will present 3 scenarios and display only the graphical output of POP-H simulations (population graph and sex-age pyramid). Dean/Scott will add to Kathryn’s description of scenario I (if necessary). Before Kathryn introduces scenario 3, Rique will discuss what QDM is and how that hunting strategy impacts population dynamics of deer. For each scenario, Kathryn will explain what information was fed into POP-II to carry out the simulation, but this information will be mostly qualitative (e. g., pre-season mortality encompasses deer death due to predation, starvation/disease, summer shooting 116 permits, poaching and deer-vehicle accidents) although numbers will be provided when necessary (e. g., all 3-6 year-old does give birth to 1.75 fawns). All participants will have copies of the graphical output. The scenarios are: 1. (15-20 minutes) Wildlife biologists, farmers and hunters agree to lower the deer’s population size (1 buck : 6 does) in a particular county. Therefore, liberal regulations are set for the regular firearm season because this year’s winter will be mild (1988). However, when the winter arrives it is both severe and long. Over a 9-year period, how will the deer population respond to 1988’s liberal hunting and harsh winter? Results: The population drops significantly but bounces back fairly quickly. By 1992 the population is larger than it was in 1988. I Kathryn and Dean/Scott address the concept of resiliency and the importance of not relying on immediate results of a management program/policy. Births and immigrations of deer over time allow deer populations to overcome adverse conditions. Also note that in some instances, such as when deer have heavily browsed their habitat, populations display a delayed reaction (i.e., drop in size) to natural disturbances, disease or poor management decisions. 2. (1 5-20 minutes) The same population and same harvesting rate begins this scenario, but the winter is mild. During the following years, hunters harvest deer at their normal levels (i.e., focusing mostly on 2-4 year-old bucks and does). I Results: The population keeps on growing. Overall population is skewed towards the females. 117 3. (IS-20 minutes) [F irst, Rique will query the participants for QDM definitions and discuss its strengths and weaknesses as a harvesting tactic.] Wildlife biologists in DMU xx decide to implement QDM as a tool for lowering the number of antlerless deer and raising the number of bucks in older age classes (i.e., 1 buck : 6 does). Over an 8 year period, bucks < 4-years-old are not harvested and only a small percentage of bucks 2 4-years-old are harvested while the doe population is harvested heavily. I Results: During the 8-year period the number of bucks within each age class grows, but the doe population drops across all age classes. The ratio of bucks .' does is less skewed. 15 minute break Activity: Group discussion about habitat quality and its influences on deer crop damage (30—40 minutes) Rique will discuss the components and characteristics of deer habitat and how they may influence population sizes at different spatial scales. His presentation of K. Braun’s research will demonstrate these relationships. Specifically, Rique will focus on how the juxtaposition of deer habitat and agricultural fields, characteristics of crop fields, and the quality of habitat influenced deer damage in dry bean and alfalfa fields in comparison to how deer are distributed over a larger landscape. Also, he will link these factors to deer distributions and explain why farmers may experience intolerable crop losses even when they farm in an area with low deer densities. Based on what we know regarding deer population characteristics and habitat attributes influencing deer density, Rique will present a scenario that demonstrates how habitat patchiness influences the number of deer potentially available for hunting or 118 causing crop damage. Aerial photographs representing a 1.5 square mile area in an agricultural community will serve as a visual aid, and Scott’s example of 40 deer in a square mile area (initially 10 bucks : 30 does) will illustrate how the number of does and bucks varies between 40 acre blocks. Questions for the participants are: I What does the habitat look like? I How are the deer distributed? Main facilitator will summarize the information Dean/Scott, Rique and Kathryn presented during the morning session and explain how it may factor into the afiemoon session. 45 minute lunch break (approx. 12:15—1:00) IV. Structured communication dialogue: Planning exercise for management of deer crop damage The main facilitator will explain the purpose and logistics of this management planning activity. [In Escanaba, farmers will meet in room 423 and hunters in 426. In Tustin, farmers will meet in the Ford A conference room and hunters in Ford B.] *Note: If the number of participants is small, conduct Stages 1 and 3 as split groups only if the farmers and hunters seem polarized and unable to communicate with one another immediately. 31389—13 Homogeneous groups of farmers and hunters will meet for 30 minutes to list all of their concerns regarding deer management (e.g., deer crop damage, hunting). I I am concerned about deer management in DM U _ because... Each group will have an equal number of biologists sitting in on the discussion and only offering information when a farmer/hunter requests it (e. g., What is the deer 119 density for this DMU? How many deer are typically harvested?) Kathryn will facilitate the farmers’ discussion and Rique will facilitate the hunters’ discussion. Their role is to record the group’s concerns on a flipchart and present those concerns during Stage 2. Kathryn and Rique also will note observations about the groups’ dynamics and ensure that the discussion remains focused and the biologists do not speak when not addressed. The main facilitator will circulate among the groups to verify the discussions remain focused and assist the group facilitators if necessary. StaggZ: Farmers and hunters (and biologists) report back to the main room to present their list of concerns. The main facilitator will make sure this 30—45 minute discussion is focused. [Points/questions that are off the subject or beyond the scope of this workshop but could be addressed when debriefing the workshop will be noted on a flipchart/overhead.] These lists of concerns will be recorded on a flip chart or overhead projector. Also, the main facilitator will use questions and feedback to determine when all participants understand why each group presented their set of concerns. 15 minute break (Rique and Kathryn copy the concerns of each others list onto a flipchart) Stage}: Two mixed groups of farmers, hunters and biologists will meet in separate rooms to discuss management strategies that respond to the concerns each group listed during Stage 2. Kathryn and Rique will facilitate the 45-minute discussion. First each group will determine how the concerns could be narrowed down, if at all. Next the groups will prioritize the concerns (each participant can vote twice). At this point, participants must develop a list of management strategies for the following headings and all recommendations must reflect general agreement among group members. 120 I Recommendation strategies for landowners I Recommendation strategies for hunters I Recommendation strategies for biologists Before each group returns to the main room, they must prioritize their recommendations—which ones must be addressed? The main facilitator will circulate between groups to verify the discussions remain focused and assist the workshop leader if necessary. Stagfi: The groups report back to the main room and post their deer management strategies for the listed concerns. Selected recorders announce their group’s strategies but do not discuss them. Through questions and feedback, the main facilitator will gage how well the participants heard and understood the presented strategies. (30 minutes) Stagej: The main facilitator leads the entire group through a 45-minute discussion to find consensus on as many recommendations as possible (it would be unlikely for the participants to agree on all strategies). A workshop leader will record on a flipchart two lists of strategies (accepted and not accepted) based on the groups’ common or different concerns. V. Debriefing the workshop experience; Kathryn and main facilitator (20 minutes) Afier the main facilitator debriefs the structured communication activity, Kathryn will ask all participants to complete the reflective writing survey found in their packets. The survey questions will be posted on the overhead, too. Kathryn will explain that the information provided would only be used to help the workshop organizers learn 121 whether or not they met their 4 objectives stated at the beginning of the workshop and how to improve the program for future use. No one’s identity will be linked to the survey when reporting its results. 122 APPENDIX D Questionnaire for Workshop Participants . Did this workshop affect how think about deer biology, habitats, populations and management issues? If no, explain why. If yes, explain how and to what extent. . Were you given opportunities to voice your opinion? If no, explain why not. If yes, explain what the opportunity was. . If your answer to question 2 was “yes,” explain whether or not you believe your opinion was valued. . I most appreciated... . If I were to attend this workshop again in the firture I would prefer... 123 "Illllllllllllllllllll