Invited Paper UNDERSTANDING THE NEW NORMAL: WILD TURKEYS IN A CHANGING NORTHEASTERN LANDSCAPE Mary Jo Casalena 1 Pennsylvania Game Commission, 2001 Elmerton Avenue, Harrisburg, PA 17110-9797, USA Michael V. Schiavone New York State Department of Environmental Conservation, Bureau of Wildlife, 625 Broadway, 5th Floor, Albany, NY 12233-4754, USA Andrea C. Bowling Department of Fisheries and Wildlife, Ecology, Evolutionary Biology, and Behavior Program, Michigan State University, 480 Wilson Road, 13 Natural Resources Building, East Lansing, MI 48824, USA Ian D. Gregg Pennsylvania Game Commission, 2001 Elmerton Ave., Harrisburg, PA 17110-9797, USA Justin Brown Pennsylvania Game Commission, 2001 Elmerton Ave., Harrisburg, PA 17110-9797, USA Abstract: Over the past 25 years, eastern wild turkey ( Meleagris gallopavo silvestris ; hereafter, turkey) populations have changed dramatically in the northeastern United States (hereafter, Northeast). Restoration efforts from the 1950s through 2000 reintroduced turkeys into unoccupied range and fostered rapid population expansion. However, since approximately 2001, turkey populations in the Northeast have declined. Reasons for this decline are multi-faceted, including a natural population contraction as turkey populations settled to levels more in line with local environmental conditions, changes in habitat conditions on a landscape scale, changes in predator populations, and changes in harvest management. Our objectives were to summarize these trends and interactions of factors that affect turkey populations, describe how these factors have changed and are projected to continue to change in the Northeast, and attempt to understand the new ecological context within which wildlife managers are challenged to effectively manage turkey populations in the Northeast. We first examined turkey management during population restoration and then took a comprehensive approach to understanding the new normal we expect for turkey populations, those who manage them, and the public who enjoy them. We discuss benefits of taking an integrated approach to turkey harvest management that incorporates a diverse array of biological and social data inputs at a landscape level in a changing, complex system. Finally, we discuss how turkey managers can improve the outlook for turkeys in the Northeast using collaboration and coordination through large scale joint initiatives. By appropriately using harvest and habitat management tools, and targeting future research to reduce key uncertainties, resiliency of turkey populations in the face of factors beyond our control can be maintained and enhanced. Associate Editor: Miller 1 E-mail: mcasalena@pa.gov 45 Proceedings of the National Wild Turkey Symposium 11:45–57 Key words: conservation cooperatives, disease, forest management, habitat, hunting pressure, Meleagris gallopavo silvestris, northeastern United States, predation, Structured Decision Making (SDM), weather, wild turkey, wildlife management. The ecological context in which eastern wild turkeys ( Meleagris gallopavo silvestris ; hereafter, turkeys) in the northeastern United States (hereafter, Northeast) exist has shifted dramatically in the past 25 years. During the 1990s, rapid population expansion was observed, facilitated by a combination of factors, including ongoing restoration efforts, suppressed predator populations, more controlled hunting seasons, and a more diverse landscape than exists today (Dickson 1992, Lewis 2001, Tapley et al. 2001). Evidence shows this system has since changed dramatically (Hughes et al. 2007, Nowacki and Abrams 2008, Backs 2009, Porter et al. 2011, New York State Department of Environmental Conservation [NYSDEC] 2014, McShea et al. 2015), which could explain recent turkey population declines. Whether current population levels are what is to be expected into the future is unclear. Therefore, research and management must now focus on determining and understanding the potential ‘‘ new normal ’’ of turkey population trends by gathering information on interactions of turkey habitat, weather, predation, disease, hunting mortality, and survival and incorporating this new under- standing of a changing ecological context into landscape level management programs for turkeys at the population level. To begin the process of understanding management issues for the future of turkey populations in the Northeast, our objectives were to: (1) summarize trends in turkey abundance and productivity in the Northeast during 1990– 2013; (2) identify and describe major factors (habitat, weather, predation, disease, hunting mortality, and surviv- al), and interactions of factors, that have been found to influence turkey populations in the Northeast; (3) summa- rize how these factors have changed in the Northeast in recent decades and are projected to change in the future; and (4) provide suggestions on management approaches to most effectively manage turkeys in the Northeast given current and potential future ecological contexts. METHODS We used spring harvest data (NYSDEC 2014) during 1990–2013, standardized as turkey harvest per km 2 of land area, as an index of turkey abundance in the Northeast. We recognize limitations of using harvest data as an index of population size (see Byrne et al. 2015), but these were the only available data across the region, and we believe harvest trends generally reflected changes in turkey populations in the region. We summarized trends in productivity using state-level data sets on mean number of poults per hen observed during summer surveys (provided by respective state agency personnel). We conducted a literature review to identify major factors influencing turkey populations in the Northeast (Fig. 1) and how interactions of these factors potentially affect turkey populations. We summarized changes in these factors over time by examining Northeast census data for avian predators (e.g., raptors; Pardieck et al. 2014), state-reported furbearer harvests (Association of Fish and Wildlife Agencies [AFWA] 2014), and forest inventory data (Dessecker et al. 2006). Due to annual variation in spring harvest figures and productivity estimates, and variation in hunting, trapping, or survey participation and effort, we used 3-year means for parameters where data were collected on an annual basis. We used habitat data from Dessecker et al. (2006) and King and Schlossberg (2014), who compiled Forest Inventory Analysis (FIA) data for each Bird Conservation Region (BCR) in the Northeast. Collectively, they used FIA data to calculate percent of each state, within BCRs that covered the Northeast, classified as small diameter forest (stands of trees , 12.5 cm dbh) in 1980 and 2005 to estimate changes in early successional habitat conditions at the landscape scale. RESULTS Trends in Turkey Abundance During 1990–1992 through 2011–2013, spring turkey harvest per km 2 increased by almost 140% in the Northeast, but this trend was not uniform over the 23-year period (NYSDEC 2014; Fig. 2). During 1990–1992 through 1999– 2001, spring harvest per km 2 increased almost 160% region-wide, but during 1999–2001 through 2011–2013, it declined about 8% (NYSDEC 2014). Grouping the Northeast between New England states (Maine, New Hampshire, Vermont, Rhode Island, Connecticut and Massachusetts) and Mid-Atlantic states (New York, New Jersey, Pennsylvania, Maryland, Delaware, West Virginia and Virginia) revealed harvest trends also were not geographically uniform (Fig. 2). Both New England and Mid-Atlantic states observed sharp increases in spring harvest per km 2 during the 1990s through the early 2000s. Since then however, harvest has declined by 25% in the Mid-Atlantic states, while harvest has continued to increase in the New England states, albeit at a slower rate. Trends in Turkey Productivity Over the past 2 decades, we also observed a decline in reproductive success, as measured by mean number of poults per hen observed during annual summer productivity surveys in each state (Fig. 3). This decline was more dramatic in Mid-Atlantic states than in New England. This apparent decline in productivity may help explain patterns of decline (Mid-Atlantic) and slower rate of increase (New England) observed in turkey populations in the Northeast. Habitat Habitat quality grades along a continuum, and increasing quality supports survival of individuals, then 46 Policy and Management Figure 1. Map of northeastern states of the United States. Figure 2. Mean spring wild turkey harvest per km 2 of land area in the 13 northeastern United States, 1990–2013. All 13 states combined in Northeast, and grouped as New England states (Maine, New Hampshire, Vermont, Rhode Island, Connecticut, and Massachusetts) and Mid-Atlantic states (New York, New Jersey, Pennsylvania, Maryland, Delaware, West Virginia, and Virginia). Figure 3. Mean poults per hen across 13 states in the northeastern United States, 1996–2013. All 13 states combined in Northeast, and grouped as New England states (Maine, New Hampshire, Vermont, Rhode Island, Connecticut, and Massa- chusetts) and Mid-Atlantic states (New York, New Jersey, Pennsylvania, Maryland, Delaware, West Virginia, and Virginia). Sample sizes varied by 3-year time period: 1996–1998 (5–6 states); 1999–2001 (6 states); 2002–2004 (7 states); 2005–2007 (7–10 states); 2008–2010 (10–12 states); 2011–2013 (12 states). Understanding the New Normal for Northeastern Wild Turkeys Casalena et al. 47 reproduction, and then population persistence (Hall et al. 1997). Habitat management at the landscape level provides population level impacts and studies have postulated that aspects of habitat condition (e.g., mast abundance, interspersion and juxtaposition of cover types) may be used as a proxy for population numbers and expected harvest of turkeys (Glennon and Porter 1999, Norman and Steffen 2003, Diefenbach et al. 2012). Other studies investigated effects of habitat features, measured at multiple scales, on nest predation, dispersal and habitat use (Fleming 2003, Fleming and Porter 2007, Jones et al. 2007, Fuller et al. 2013). These studies indicated that turkey populations may persist in a more homogenous landscape (e.g., a landscape dominated by mature timber), but reach their highest densities in a heterogeneous landscape with a mix of agriculture (e.g., row crops, pasture), mature forest, and early successional vegetation community types such as old fields, shrublands, and young forest (Glennon and Porter 1999, Porter 2007). Such a diverse landscape provides for year-round habitat needs, including nesting, brood-rearing, roosting, and wintering; increases complexity of patches, thus decreasing predator foraging efficiency (Glennon and Porter 1999, Fleming 2003); and provides food sources to help mitigate severe spring and winter conditions (Porter et al. 1983, Vander Haegen et al. 1988). At least 3 important changes in landscape-level habitat quality in the Northeast are evident over the past 25 years. First, there has been a decline in amount of interspersion of different habitat types across the region (Dessecker and McAuley 2001, Dessecker et al. 2006). Second, percentage of forests composed of mast- producing oaks ( Quercus spp.) has been declining, particularly in the Mid-Atlantic portion of the region (Nowacki and Abrams 2008, McShea et al. 2015). Change in forest composition has been driven by lack of forest management, forest fire suppression policies beginning around the 1920s, browsing by over-abundant white-tailed deer ( Odocoileus virginianus ), gypsy moth ( Lymantria dispar ) mortality since the 1980s (Wunz and Pack 1992, McShea and Healy 2002, Nowacki and Abrams 2008, McShea et al. 2015), and other introduced exotic pathogens, such as chestnut blight ( Cryphonectria para- sitica ), beech bark disease ( Nectria spp. and Neonectria spp.) and emerald ash borer ( Agrilus planipennis; Abrams 1992, Porter et al. 2011, Jones et al. 2015). With little to no regeneration, oak-dominated forests are being converted to maple ( Acer spp.) and other non-mast producing hardwood species that are shade-tolerant and fire-sensitive (Nowacki and Abrams 2008, McShea et al. 2015). Simultaneously, microenvironmental conditions have improved for shade- tolerant species (i.e., cool, damp, shaded conditions), which have created less flammable fuel beds and deteriorated ability of shade-intolerant, fire-adapted oaks to persist. As Nowacki and Abrams (2008) stated, cost and effort to restore fire-adapted ecosystems will escalate rapidly if this process continues at the current rate. Third, forest maturation has reduced amount of early successional habitat conditions important for nesting and brood-rearing by turkeys across the Northeast (Dessecker et al. 2006, Jones et al. 2007, King and Schlossberg 2014). Forest Inventory Analysis data for each BCR in the Northeast revealed that, among 12 states for which data were available (data were not available for Delaware), 10 states exhibited a declining percentage of small diameter forest from 1980 to 2005 (Dessecker et al. 2006, King and Schlossberg 2014). Mean declines ranged from almost 1% (Virginia) to almost 16% (New York, Rhode Island). The only states for which there was an increase in small diameter forest were New Jersey (4%) and Maine (15%). Loss of landscape-scale habitat complexity may be a significant contributing factor in the trends observed in turkey harvest and productivity. Weather Due to a colder climate in northern limits of wild turkey range, weather is an important factor affecting populations in the Northeast (Healy 1992). Research has concentrated on winter and spring weather effects on turkey demographics and knowledge of its effects on management decisions. Winter severity has been shown to reduce survival, but has not been found to have long-term effects on population growth (Wunz and Hayden 1975). During the 1970s, severe winters with extended low temperatures, excessive snow depths, and sudden deep snowfalls caused turkey survival to be reduced for newly established populations that perhaps had not yet acquired behavioral adaptations like flocking (Austin and DeGraff 1975, Wunz and Hayden 1975). Negative impacts of winter severity for established populations appear to act through access to food, with available resources providing less nutritional value (Pekins 2007) and reduced food resources negatively impacting reproductive success (Casalena 2015). Con- versely, agricultural food sources (e.g., manure spreads, silage pits) and supplemental feeding in agricultural areas have been shown to mitigate winter severity effects (Vander Haegen et al. 1989, Hamel 2002, Timmins 2003). Weather has been shown to affect nest incubation initiation, with greater March temperatures correlated with earlier incubation and deeper March snow postponing incubation initiation (Norman et al. 2001 a ). Nest success and daily survival also are affected by weather, as they are positively correlated with average to below average rainfall, negatively correlated with number of rain events, and daily nest survival is positively associated with heating degree days (Priest 1995, Roberts et al. 1995, Roberts and Porter 1998 b ). Roberts and Porter (1998 a ) found that poult survival was negatively correlated with lesser temperatures and greater rainfall. These studies also demonstrated links between nest success and poult survival and abundance. Similarly, annual change in fall harvest has been negatively correlated with annual change in May rainfall (Roberts and Porter 2001), such that spring weather data can be a surrogate to reproductive success to predict annual poult survival, assess trends in fall populations (Norman and Steffen 2003), and adjust fall harvests accordingly (Roberts et al. 1995, Roberts and Porter 1996, Norman and Steffen 2003). Average annual rainfall in the Northeast since 1991 has increased 8% relative to 1901–1960, with more winter and spring precipitation (U.S. Global Change Research Pro- gram 2014). Additionally, there has been an increase in amount of precipitation falling in heavy events in the 48 Policy and Management Northeast, which is projected to continue due to a warmer atmosphere and associated changes in large scale weather patterns (U.S. Global Change Research Program 2014). Interactions of Habitat Conditions and Weather Several studies have investigated importance of both weather and habitat conditions on turkey populations in the Northeast. Porter and Gefell (1996) showed importance of multiple land cover types and spring and early summer temperatures on fall harvest across southern New York. Vander Haegen et al. (1989) found that agricultural food sources buffered effects of winter severity in southwest Massachusetts. Roberts et al. (1995) suggested that improvements to winter habitat in areas with frequent, severe winters would provide greatest benefit to turkey populations, whereas improvements to nesting habitat would be more beneficial in areas with infrequent severe winters. Norman and Steffen (2003) found that fall harvest rate index increased with below-average oak mast (acorn) production in Virginia. Effects of interactions between weather and habitat conditions on turkeys at large scales have also been explicitly examined. Bowling (2014) demonstrated an implicit interaction between weather and habitat conditions revealed by negative effects of spring rainfall on fall harvest in better quality habitat conditions (i.e., more diverse with more agricultural land). Conversely, in poorer quality habitat conditions (less diverse with more forested land), fall harvest was unaffected by spring weather. A smaller-scale interaction occurred within a region where increasing proportion of open lands correlated with increased fall harvest, but this positive relationship was suppressed by increased June rainfall (Bowling 2014). Differences among findings from past studies of effects of weather on reproductive parameters (Roberts and Porter 1998 a ,1998 b ; Norman et al. 2001 a ) are not inconsistent with each other, but rather reveal regional differences in effects due to regional differences in weather (e.g., little variation and little total May rainfall in Virginia and West Virginia), habitat quality, and possible differences in predator densities (Norman et al. 2001 b , Hughes et al. 2007). Disease Turkey populations in the Northeast have fewer disease issues than the southeastern United States (David- son and Wentworth 1992). To date, disease outbreaks have not had discernible long-term, large scale impacts on populations (Weinstein et al. 1996, Jones et al. 2015). However, 2 viral infections of concern are avian pox and Lymphoproliferative Disease Virus (LPDV). Avian pox weakens birds and makes them vulnerable to predation (Davidson and Wentworth 1992). Documented, relatively recently (Allison et al. 2014), LPDV occurs widely but may only produce clinical disease when a turkey’s immune system is weakened from presence of other parasites and diseases (Katrina Alger, State University of New York College of Environmental Science and Forestry, unpub- lished data). Many diseases that affect turkeys are influenced by anthropogenic processes, including spillover from domestic poultry (e.g., histomoniasis), toxicoses (e.g., zinc phosphide from rodenticide applications, particularly in orchards), trauma, and intentional movement of captive- bred and wild animals (Leighton 2002, Caudell et al. 2015). Potential for more subtle impacts of disease on productiv- ity, immunity, energy assimilation, resource allocation, and how disease may interact with other population influences such as habitat and weather are poorly understood (J. Brown, Pennsylvania Game Commission, personal com- munication). Predation Predation is a complex process that involves predator– prey interactions and a suite of factors that influence it including habitat, weather, disease, and presence and abundance of other prey species (Miller and Leopold 1992). Turkeys are preyed upon by a suite of predators across all life stages. Omnivorous meso-mammals and snakes are prolific nest predators, depredating nesting hens and eggs, thus making these life stages the most vulnerable for turkeys (Hughes et al. 2007). As ground nesters, eggs of nesting hens are subject to other, opportunistic, predators, such as a variety of birds, small mammals, house cats ( Felis domesticus ; Miller and Leopold 1992), and occasionally black bears ( Ursus americanus ; M. J. Casalena, unpub- lished data). Larger mammalian predators also prey on nesting hens and eggs and kill adult birds, as do avian predators (e.g., hawks and owls; Miller and Leopold 1992). Predation rates are influenced by habitat quality (on local and landscape scales), weather, and disease. Previ- ously, we mentioned that habitat quality affected risk of nest predation (Fleming 2003). Many studies have linked spring rainfall to nest success through the moisture- facilitated nest depredation hypothesis (Roberts et al. 1995; Roberts and Porter 1996, 1998 b ; Lowrey et al. 2001). Also, prolonged periods of deep snow during winter decreases physiological health of turkeys (Pekins 2007) and may increase their vulnerability to predation. Great densities of turkeys may increase transmission of diseases like avian pox, weakening them and thus increasing their vulnerability to predation (Davidson and Wentworth 1992). Predation alone is not known to be a direct threat to turkey populations, but interactions with other factors increase its potential effects (Hughes et al. 2007). Populations of species that prey on turkeys have increased in number and distribution (AFWA 2014, Pardieck et al. 2014). For example, populations of avian predators common to the Northeast that prey on turkeys (broad-winged hawk [ Buteo platypterus ] and red-tailed hawk [ B. jamaicensis ]) have experienced significant increases from 1966 to 2012 (Pardieck et al. 2014). Rabies swept through the Northeast in the early 1990s, suppressing populations of meso-predators (Smith et al. 2002, Dyer et al. 2013). Since then, state-reported hunter and trapper harvest estimates, an index of furbearer abundance, show that number and diversity of mammalian predators have increased over the past decade (AFWA 2014; Fig. 4). However, this trend was not the same throughout the region. Populations of red ( Vulpes vulpes ) and gray ( Urocyon cinereoargenteus ) foxes, raccoon ( Procyon Understanding the New Normal for Northeastern Wild Turkeys Casalena et al. 49 lotor ), striped skunk ( Mephitis mephitis ), Virginia opossum ( Didelphis virginiana ), and weasel ( Mustela spp.), as measured by the fur harvest per km 2 , were greater in the Mid-Atlantic states than in New England and have gradually increased over the past 2 decades (AFWA 2014; Fig. 4). Conversely, populations have been stable or declining in New England. Coyote ( Canis latrans ), bobcat ( Lynx rufus ), and fisher ( Martes pennanti ), while at lesser densities than other mammalian predators, have been rapidly increasing in both numbers and distribution throughout the Northeast (AFWA 2014; Fig. 4). Hunting Much research has addressed the question of how hunting affects turkey populations and results have varied among study areas and by length of investigation. Hunting pressure did not affect populations in southern New York during 1969 1981 (Porter et al. 1990), but this population was also increasing due to restoration efforts. However, longer fall turkey seasons increased fall harvest in Virginia from 1973 to 2002 (Norman and Steffen 2003) and hen harvest rates in Pennsylvania from 2010 to 2014 (Casalena 2015). Also, across western Virginia and West Virginia during 1989 1994, fall harvest affected survival, popula- tions, and subsequent harvest (Pack et al. 1999, Alpizar- Jara et al. 2001). In south–central Pennsylvania during 1999 2001, Casalena et al. (2007) found a population decline was partially due to great mortality from fall harvest. Population models have shown that harvesting 5– 10% of the fall population will allow continued population growth (Vangilder and Kurzejeski 1995, Healy and Powell 1999), but these studies were conducted when turkey populations were still increasing with great reproductive output. Due to potential of overharvest during fall seasons, wildlife managers exhibit caution when setting them (i.e., season length, season timing, bag limits, permit allocations; Healy and Powell 1999). Current spring harvest rates do not appear to affect populations (Eriksen et al. 2011, Diefenbach et al. 2012), and population modeling suggests that a spring harvest rate of 30% for males is sustainable (Vangilder 1992, Healy and Powell 1999). However, if multiple years with great harvest rates on adult males changes the age structure such that juvenile males make up a larger proportion of the spring population, researchers speculate this may negatively affect both population dynamics and hunter satisfaction (Healy and Powell 1999, Diefenbach et al. 2012). Most Northeast state agencies do not express spring harvest goals in terms of desired harvest rates, but set seasons to ensure conservative harvest rates (Healy and Powell 1999). DISCUSSION Several environmental factors in the Northeast have changed over the past 10–15 years in ways that likely negatively affect turkey populations (e.g., decreased landscape heterogeneity, less mast production, unpredict- able weather, unforeseen effects from disease, and increased predator densities), challenging managers to reexamine their expectations for turkey abundance and productivity and to identify where and how limited resources can best be applied for maximum benefit to turkeys. Northeast turkey populations in the future likely will not reach densities in which they existed during the peak of restoration. Causes for declining abundance and productivity are numerous and complex (NYSDEC 2014), as outlined above. Also, the fact that trends in turkey demographics are not uniform across the Northeast further complicates interpretation. We hypothesize that within- region differences in reproductive success and harvest trends may partially be explained by differences in time since restoration such that, in general, just as restoration began and concluded earlier in the Mid-Atlantic states than in the New England states, northern sections of the region may eventually experience declines similar to those in much of the Mid-Atlantic. Furthermore, we suggest changes in landscape-scale habitat quality and predator communities, and long fall hunting seasons (in some Mid- Atlantic states) contributed to trends observed in turkey Figure 4. Mean furbearer harvest density (take/km 2 of land area) in 13 states in the northeastern United States, 1990– 2013. Mammalian Predators (top) includes red fox, gray fox, raccoon, skunk, opossum, weasel, coyote, bobcat, and fisher. All 13 states combined in Northeast, and grouped as New England states (Maine, New Hampshire, Vermont, Rhode Island, Connecticut, and Massachusetts) and Mid-Atlantic states (New York, New Jersey, Pennsylvania, Maryland, Delaware, West Virginia, and Virginia). Data are from the AFWA National Furbearer Harvest Statistics Database (http:// fishwildlife.org/?section = furbearer_management_resources). 50 Policy and Management abundance and productivity, as illustrated by data presented here. Previous research indicated that optimal habitat conditions may buffer negative effects of weather, predation, disease, and hunting, but with a decline in landscape-scale habitat quality, this buffering effect may no longer be functioning as it once did. Losses of early successional vegetation communities and young forests have most likely reduced quality and quantity of nesting and brood-rearing cover (Glennon and Porter 1999, Porter 2007). More mature forests also lack alternative foods that are typically found in the presence of a diversity of timber age classes and, therefore, exhibit great annual variation for fall and winter food availability for turkeys (McShea and Healy 2002, Nowacki and Abrams 2008, Porter et al. 2011, McShea et al. 2015). Forest composition is transitioning from that of mast-producing tree species to other non-mast producing hardwood species less beneficial to turkeys (Nowacki and Abrams 2008, Porter et al. 2011, McShea et al. 2015). Throughout the Northeast and elsewhere, agricultural areas had been reliable alternative sources of food, but these areas have either been lost outright due to development or reduced in habitat quality due to modern farming practices that are less compatible with turkey needs (e.g., less waste grain, more intensive crop production methods; Porter et al. 2011). In some locales, turkeys have actually become more depen- dent on silage storage areas, becoming a nuisance to the agricultural community (Pekins 2007, Porter et al. 2011). Simultaneously with landscape-scale habitat changes, turkeys in the Northeast face a more abundant and diverse predator community than was the case in the 1990s (AFWA 2014). While predation can play a role in limiting local turkey populations, great predation rates may be symptom- atic of a landscape with poor habitat quality causing turkeys and their young to be more vulnerable to predation. Added negative effects of unpredictable severe weather conditions (winter and spring; U.S. Global Change Research Program 2014) and potential disease outbreaks amplify effects of poor habitat quality and predation. The idea that multiple processes and patterns affect turkey demographics is not new. We know that weather and habitat conditions affect nest success and poult and hen survival. Adverse weather affects nest success through predation loss (Priest 1995; Roberts et al. 1995; Porter and Gefell 1996; Roberts and Porter 1998 a, 1998 b ; Norman et al. 2001 a ; Casalena et al. 2007). Weather affects poult survival directly and indirectly through exposure, predation and food availability (Healy and Nenno 1985, Healy 1992, Lowrey et al. 2001). Effects are more severe where weather patterns are more severe, such as in northern climates and high elevations (Healy 1992, Porter 2007). Quality brood- rearing habitat is essential for poult nutritional require- ments (Porter 1992, Harper et al. 2001, Backs and Bledsoe 2011). These pressures from multiple factors synergistical- ly affect recruitment and population growth. While successive years with above-average reproductive success can allow turkey populations to temporarily increase, the apparent overall downward trend in populations over the past 15 years may be indicative of a larger, systemic problem that is not being counteracted by good production years. This downward trend is likely being intensified due to liberal hunting regulations, which reflect a time lag- induced information gap on current population trends, thus not allowing wildlife managers to quickly adjust hunting seasons. Porter (2007: 310) suggested that the drivers of turkey populations appear ‘‘ to operate at multiple geographic scales ’’ , exacerbating the challenge of prescribing manage- ment actions within and among states that attempt to maintain or improve turkey numbers across a large spatial scale. Regional differences in habitat and weather likely affect populations with different levels of importance and likely interact with each other in varying degrees and across different geographic scales as described by Bowling (2014). Accounting for these differences, and possible interactions, allows better understanding of a population’s capacity to sustain varying levels of harvest. These differences are important to consider when assessing fall harvest potential of a landscape. Recognizing these interactions is especially important with declining turkey populations, which are more vulnerable to harvest, especially when harvest rates and hunter densities are great (Healy and Powell 1999, Norman et al. 2001 b , Norman and Steffen 2003, Casalena et al. 2015). Just as understanding interactions among the many factors that influence turkey populations and productivity requires a well-integrated analytical approach (Weinstein et al. 1996), making appropriate decisions regarding turkey harvest management, one of the primary tools for managing turkey populations, requires a well-organized process to integrate different sources of information and different (sometimes competing) objectives. A particular challenge for turkey managers in the Northeast is setting fall either- sex hunting seasons in light of biological and social tradeoffs involved. Fall hunting opportunity is a tradition that is highly desired by stakeholders, but results in mortality that is generally considered to be additive (Little et al. 1990, Vangilder 1992, Healy and Powell 1999, Pack et al. 1999). Thus, fall hunting may constrain future turkey abundance and, in turn, ability of managers to maximize other aspects of stakeholder satisfaction (e.g., spring hunting opportunity and success, and viewing opportuni- ties). The lesser reproductive capacity of current turkey populations may warrant using a lower threshold for acceptable fall harvest rates when setting fall seasons. These tradeoffs have become more pronounced as weather, habitat quality, and other external influences on populations have become less favorable across much of the Northeast, and in the face of our nascent ability to understand complex interactions of these ecological factors, which tends to force us to manage reactively rather than proactively. MANAGEMENT IMPLICATIONS Due to changes in turkey populations in the Northeast and the environmental variables that influence their abundance and distribution since the early 1990s, adjust- ments to management are necessary to ensure sustainable use of this important resource. Broadly speaking, turkey management entails considering how habitat, weather, disease, predation, and hunting all interact to influence turkey populations and productivity across large spatial scales. Therefore, management of harvest and habitat, and future research, should be viewed at a large (population and Understanding the New Normal for Northeastern Wild Turkeys Casalena et al. 51 landscape level) scale and in an integrated fashion, rather than as localized or isolated issues. The following specific examples may be useful in applying a large scale integrated approach to research and management. Harvest Management Historically, there has been strong interest in standard- izing methods across state boundaries to facilitate popula- tion management objectives. During the late 1990s, once turkey population restoration was complete in much of the Northeast, turkey management focus shifted toward sustainable harvest management. Healy and Powell (1999) compiled a list of 23 methods to estimate or index turkey populations used by northeastern states and Ontario for harvest management. Their objective was to facilitate standardization of methods and protocols for data collec- tion within the region to allow comparability among jurisdictions and aid in proper harvest management. During the next 15 years, northeastern states standardized the most emphasized population index methods for tracking popu- lation trends: spring harvest and productivity via summer brood counts (Massachusetts Division of Fisheries and Wildlife 2013, NYSDEC 2014). These states also devel- oped a regional hard mast survey to help explain annual variation in fall turkey harvests and predict hen reproduc- tive capacity the following spring (Cardoza et al. 2007). Standardized methodologies facilitated identifying current declining harvest trend in the Mid-Atlantic states, and the more slowly increasing harvest trend in the New England states (NYSDEC 2014). Part of the original standard for spring harvest management, developed via population modeling when turkey populations were expanding and, presumably, exhibiting their greatest reproductive rates, allowed no greater than a 10% harvest rate of the adult hen population during spring season for continued population growth (Vangilder and Kurzejeski 1995, Healy and Powell 1999). Since the 1990s, spring season regulations have been liberalized across the Northeast (e.g., earlier opening dates, longer seasons, increased permit allocations, all-day hunting in some states), which has increased potential for greater adult hen harvest (illegal or unintentional) and disruption to breeding activity. Meanwhile, population growth rates have either decreased or are now declining (NYSDEC 2014). Additionally, the standard spring male harvest with limited either-sex fall harvest strategy, also developed via population modeling while turkey popula- tions were expanding, suggested harvests of 5–10% of the fall population allowed for continued population growth, while harvests exceeding 10% would usually lead to population decline (Vangilder and Kurzejeski 1995, Healy and Powell 1999). However, given current lesser and decreasing population trends, the 5–10% fall harvest rate may no longer be appropriate. Furthermore, many northeastern states opened a fall turkey season in a management unit when spring harvests equaled or exceeded 1 turkey/2.59 km 2 of turkey habitat (Pack et al. 1995). This standard was subsequently revised for poorer quality habitat conditions to 0.5 turkeys/2.59 km 2 of turkey habitat (West Virginia Department of Natural Resources [WVDNR] 2010). While this metric has been of value to wildlife managers in many northeastern states, a more inclusive, objective process to management is now necessary that incorporates harvest, survival, and productivity data, landscape scale habitat characteristics, and stakeholder values. A new process for setting fall turkey season being developed by managers and researchers in New York and Pennsylvania uses Structured Decision Making (SDM), a transparent, defensible, objective process that incorporates a range of data sources in an organized fashion (Fig. 5; Clemen 1996, Hammond et al. 1999, Robinson and Fuller 2015). This approach allows managers to develop a decision framework that evaluates management strategies that balance competing objectives (e.g., turkey abundance and harvest opportunity), while seeking to provide for the simultaneous goals of sustainable use of turkey populations and optimal opportunities for hunters and other stakehold- ers (Clemen 1996). Use of SDM facilitates an adaptive management process in which model-based predictions about effects of decisions can be made, evaluated by subsequent monitoring, and iteratively improved and fed back into the modeling and decision-making process in future years so that learning occurs over time and research can be focused on reducing key uncertainties. This decision framework is established at an appro- priate spatial scale, such that ecological regions are grouped according to similar habitat conditions and harvest potentials (e.g., New York is grouped into 3 broad regions; Bowling 2014). Additionally, the framework incorporates multiple variables, accounting for population dynamics, habitat, weather, and stakeholder priorities for setting fall turkey seasons obtained through appropriate social science survey methods (Siemer et al. 2013). This process allows state wildlife agencies to evaluate and select among fall turkey harvest regulatory alternatives (Fig. 6). This process also can be used as a standard management tool across jurisdictions with standardized survey and data acquisition methods, which have historically varied across states. These standardizations also allow for assessing how and why populations differ across states and regions. Additionally, a regional harvest potential model was developed in New York that incorporated how habitat quality and weather affected fall harvest