Productivity Report
Transcription
Productivity Report
Southeast Regional Wild Turkey Reproductive Decline Study A Retrospective Regional Analysis of Wild Turkey Productivity in the Southeast Investigators: Michael E. Byrne and Michael J. Chamberlain Warnell School of Forestry and Natural Resources University of Georgia 180 E Green St Athens, GA 30602 Bret A. Collier School of Renewable Natural Resources Louisiana State University 227 Highland Rd Baton Rouge, LA 70803 History and background This study was initiated as a response to persistent declines in annual reproductive indices of eastern wild turkeys (Meleagris gallopavo silvestris) based on surveys conducted by member states of the Southeast Wild Turkey Working Group (SEWTWG). There was concern among state biologists that the observed reproductive declines were an indicator of large-scale regional declines in wild turkey populations that was presently occurring, or was likely to occur in the near future. At the 2010 SEWTWG meeting held in North Carolina, focused discussion on reproductive declines led to the identification of research priorities relating to the decline, including population dynamics, habitat, and survey techniques. At the 2011 SEWTWG meeting in Florida, research priorities were formalized and member states decided that prior to initiating new field studies, it would be informative and cost effective to study demographic trends retrospectively using existing datasets maintained by member states. The analysis of historic trends in wild turkey demographic parameters would provide a long-term, large-scale perspective on wild turkey populations in the region. In doing so, survey techniques could be examined critically, and hypotheses could be developed from the existing data to further refine research priorities. At the 2011 SEWTWG meeting, it was agreed that the cost of the project would only be that necessary to cover 2 years’ salary for a post-doctoral researcher stationed at the University of Georgia (UGA). The idea was to hire a post-doctoral researcher and have the cooperating states each contribute equally to the effort, thereby minimizing costs to all cooperators. Each state was given flexibility to decide how funds were contributed; some state agencies provided funds whereas other states provided funds through their respective state chapters of the National Wild Turkey Federation (NWTF). All monies were directed to NWTF headquarters in Edgefield, SC, who ultimately routed the funds to UGA through a contract with the Warnell School of Forestry and Natural Resources. As per NWTF policy, there were no indirect costs associated with these funds. A full report consisting of a critical examination of survey techniques used by states to measure productivity of wild turkeys and potential improvements to such techniques was delivered to the SEWTWG in January 2014. The present report deals specifically with the analysis of historic demographic trends. Our specific goals were two-fold: 1) Summarize and analyze trends in measurable demographic parameters temporally and spatially. 2) Generate hypotheses to account for observed declines in productivity and identify avenues of future research. As we received and summarized the available data, our approach naturally had to evolve to suit what data were available. Ultimately, we decided to take a relatively broad-scale approach and examine trends at the state level in the style of a meta-analysis. This was primarily a consequence of data availability and quality, which as we detail below, was highly disparate among states and years. After examining trends for individual states, we were able to identify regional commonalities and draw biological inferences. We ultimately decided on this approach because we did not want to sacrifice biological interpretability for the statistical complexity that would have been required to standardize and combine data at smaller spatial scales and across state boundaries. The following report does not proceed in the format of a standard scientific paper, with distinct sections devoted to methods, results, and discussion. Given the complexity of the study and the many aspects incorporated in making inferences from the data and developing hypotheses, we instead decided to present our results in the form of a case-study. As such, we detail the process step-by-step, from data collection, through various analyses, and finally through our development of a theory to explain the observed reproductive declines. Our goal with this presentation was to allow the reader to follow our thought process, thereby providing insight into why we chose the various approaches outlined below, and how we ultimately reached conclusions and developed hypotheses. Data collection and availability Data acquisition began in April 2012, coinciding with the initiation of the contract that allowed funding to be routed from NWTF to UGA. At that time, we contacted the turkey coordinators of cooperating states and asked them to provide all available data and historical records regarding productivity indices, harvest records, and turkey restoration and restocking information. We requested coordinators provide as much of the associated metadata and background information as was available for each dataset. Thirteen states were able to provide records of productivity indices (Table 1). Data from Oklahoma includes only that from the southeastern portion of the state occupied by the eastern subspecies. Alabama initiated a state-wide productivity monitoring program in 2010, and we did not use those data for making inferences on long term productivity trends (Table 1). The primary metric used to index reproduction region-wide was the poult per hen ratio (PPH). In a general sense, the PPH ratio is defined as the ratio of the total number of poults to the total number of females observed during the summer brood-rearing period. The methods of data collection and calculation of PPH ratios varied somewhat among states. In most states, sightings of turkeys are recorded opportunistically during the summer months by agency personnel, or a combination of agency personnel and citizen volunteers, as they go about their daily activities. Generally observers are asked to record observations of females without broods, as well as females with broods. West Virginia is different in this regard, as observations of females without young were often not recorded. Thus, the reader should bear in mind that the reported PPH ratios for West Virginia have a slightly different biological meaning than other states. For most states, the observation period included the months of June – August; Tennessee reports PPH ratios based only on observations recorded in August, the sample period for Kentucky and North Carolina is July – August, and West Virginia uses observations of broods made from May - September. All states ask observers to record individual sightings as separate events. When calculating PPH ratios however, the total numbers of poults and females from individual observations are combined to provide a single estimate. Guidelines used for filtering spurious and unlikely observations prior to calculating PPH ratios, if they exist at all, are not standardized across states and in many cases are not standardized across time within a state (for instance, a current turkey biologist may filter observations whereas his predecessors may not have done so). Of the states that monitored productivity, Virginia was the only state in which the PPH ratios we used were not calculated from summer observations. Rather, PPH ratios were derived from the ratio of yearling to adult females in the fall harvest based on reports at mandatory hunter check stations. While Virginia does record summer brood observations, inferences regarding productivity were traditionally based on fall harvest data as the spatial distribution was more consistent, and sample sizes larger, than summer observations. Virginia discontinued using fall harvest as a means of indexing productivity in 2010, but since this represents a continuous data set of 26 years and was traditionally used by the state as the primary measure of productivity, we considered it the most appropriate for our purposes. All states were able to provide harvest records. However, there was significant inconsistency regarding what data were available, how data were obtained, and length of historical records. The one metric common to all states was an annual estimate of total spring harvest; however, the analytical methods used to develop estimates differed among states and often changed within states through time. Estimates of total spring harvest were variously derived from information gathered at check stations, or through hunter surveys conducted via mail, phone, internet, or various combinations thereof. Calculations of harvest metrics were variously conducted in-house by agency personnel, or by outside entities such as universities or consulting firms. Harvest estimates are often used as proxies of population density, although few studies have tested the veracity of this relationship (Lint et al. 1995). A more robust index would theoretically be gained by accounting for hunter effort. Unfortunately, the data necessary to estimate hunter effort was only available in 9 states and normally only for a limited number of years. Nine states allow, or have allowed, a fall turkey harvest in the past, but we concentrated our analyses on spring harvest because 1) all cooperating states have a spring harvest season and 2) spring seasons see much greater hunter interest and participation than fall seasons, especially in recent decades. Alabama, for example, experienced an 89% reduction in fall hunter numbers between the years 1977-2011, and these trends are not restricted to the southeast (Tapley et al. 2010). Presently fall harvest is greatly surpassed by spring harvest in terms of birds killed and numbers of participating hunters in all states in the region. Detailed historical restoration information was available from 6 states that also provided productivity data (GA, LA, MO, MS, NC, and TN). This information primarily consisted of the number of turkeys released into a specific area. In some cases, releases were detailed to the specific release site, and in other cases releases were summarized by county. Given the largescale nature of this study, we tabulated the total number of birds released state-wide in a given year. State agencies work independently in collecting and summarizing turkey population metrics and the resulting inconsistencies makes comparisons among states problematic. As such, we used data from the USGS Breeding Bird Survey (BBS) as an independent and methodologically consistent index to overall population trends for each state. The BBS is an annual road-based survey designed to track large-scale changes in avian abundance over time. Over 4,000 survey routes exist in the United States and Canada; each route is 39.43 km in length, is surveyed once annually, and is often run by the same observer consecutively for a series of years. Each route contains 50 evenly spaced survey sample points. At each sample point, the observer conducts a 3 minute point-count, recording all bird species seen and heard. BBS data are summarized for individual states as well as larger physiogeographic regions of the continent. For any geographic region, a Bayesian-based hierarchical model which accounts for varying regional survey quality and observer effects (Link and Sauer 2002) is used to estimate an annual population index (measured as observations/route). For our purposes, the standard protocol over space and time is an obvious advantage to the incorporation of BBS indices. Additionally, the BBS has been ongoing since 1966, providing a long term data set that includes the major restoration efforts of the wild turkey in the southeast through the present day. The BBS index incorporates observations of all turkeys observed regardless of age or sex, and as such provides an index of total turkey populations as opposed to most state-derived estimates which are based on data pertaining to specific sexes (i.e. spring harvest data). We queried annual abundance indices and associated 95% credibility intervals for each individual state for the years 1966 – 2011 through the interactive online BBS results and analysis portal (http://www.mbrpwrc.usgs.gov/bbs/bbs.html). Analysis Because considerable differences in data availability and quality among states precluded us from being able to pool data across states, we decided to adopt a meta-analysis approach. We accomplished this by first analyzing available data from each state independently, then making biological inferences based on the broad-scale similarities in trends observed across states. All productivity and harvest datasets are statistically imprecise, primarily driven by inherent biases in data collection methodologies and haphazard survey protocols. Additionally, stochastic annual variation in conditions that may influence nesting success can lead to considerable annual variation from long term productivity averages (Vangilder et al. 1987, Healy 1992). Similarly, factors such as weather conditions and the timing of harvest seasons relative to the timing of a given year’s nesting cycle may introduce noise into estimates of spring harvest. The variation Table 1. The historical availability of statewide productivity data (PPH ratios) from 13 states in the southeastern United States; including the total number of years (n), and the availability of raw data. State Years n Raw data Notes AL 2010-2011 2 NA AR 1982-2012 31 NA GA 1978 -2012 35 1978-2011 KY 1984-2011 28 NA LA 1994 -2010 17 1994-2010 MS 1995 -2012 18 1995-2012 MO 1959 -2012 54 1990-2011 PPH ratio only calculated for brood groups with ≤ 2 hens NC 1988 -2012 25 2001-2011 OK 1985-2012 28 2001-2012 Data only includes SE portion of the state where the eastern sub-species occurs SC 1982 -2012 31 NA TN 1983 -2012 30 2003-2012 PPH ratios calculated only from observations during month of August VA 1979-2010 32 NA Productivity calculated from ratio of adults/young in fall harvest WV 1967-2012 46 1967-2012 Based only on observations of females with broods Table 2. Historical availability of state-wide spring harvest from 15 states in the southeast United States; total harvest = estimated number of total males harvested, hunter numbers = estimates of total hunters or eligible hunters based on permit sales. State Total harvest Hunter numbers AL 1972 – 2012 1972 - 2012 AR 1961 - 2012 NA FL 1989 – 2012 1989 - 2010 GA 2005 - 2013 2005 - 2013 KY 1978 – 2012 1997 - 2012 LA 1980 – 2012 1980 – 2010 MS 1981 - 2012 1981 - 2009 MO 1960 - 2012 1960 - 2011 NC 1977 – 2011 1976, and every 7th year since 1983 OK 1990 – 2012 NA SC 1976 - 2011 1977 - 2010 TN 1990 – 2010 NA TX 1995 – 2012 1983 - 2011 VA 2004 – 2013 NA WV 1996 – 2012 NA caused by such stochastic factors may obfuscate the underlying historical trends representing the larger population-level processes in which we are interested. To account for the expected nonlinearity in our data due to these stochastic effects, we used generalized additive models (GAM, Wood 2006) to fit smoothed spline regression functions to time series data sets of productivity and harvest to elucidate the underlying historical trends. We adjusted the number of knots (similar to piecewise regression) used in producing the spline curve to improve the fit of each model and evaluated goodness of fit for each model visually based on diagnostic residual plots. We report the model predicted estimate and 95% confidence interval for each year. GAM’s were fit in the statistical program R (R Core Team 2013) using the mgcv package (Wood 2013). Establishment of long term demographic trends Our first priority was to identify long-term trends in demographic data regarding productivity and estimated population size. Productivity, as measured by PPH indices, has generally declined across all states where data were available (Figure 1). The nature and severity of declines was somewhat variable however; for example Tennessee has experienced a particularly steep historical decline whereas productivity in Mississippi has remained relatively stable over the record-keeping period (Figure 1). We offer that direct comparisons among states regarding the actual PPH ratios are tenuous, as there are many confounding, latent variables to consider. As such, it is difficult to parse out whether differences in scale represent actual differences in true productivity among states, or are artifacts of differing sample protocols. For example, West Virginia has consistently greater estimates than other states, but West Virginia’s sampling protocol also varies considerably from other states since females without poults are not recorded. The important observation is the consistency in the generally declining trend regionwide, and not the absolute PPH estimates. This applies to total harvest data, and to an extent, the BBS data as well. Trends in spring harvest numbers were highly variable across states, and it is difficult to generalize an overall region-wide trend (Figure 2). A number of states, including Kentucky, North Carolina, and Tennessee exhibited increasing trends in spring harvest through time, whereas states such as Arkansas, Missouri, South Carolina, and West Virginia exhibited a trend of stabilizing or decreasing harvest following a peak in the late 1990’s or early 2000’s. The most precipitous and persistent decline was observed in Mississippi (Figure 2). Inferring a direct relationship between harvest and population is tenuous as harvest is a function of availability and rate, and there is obvious age and location specific selection in harvests of wild turkeys. Harvest may be a sufficient index if hunter effort remains constant over time (Lint et al. 1995), otherwise trends in overall harvest may be representative of hunter population dynamics more so than turkey populations. We performed simple linear regression on harvest data from 8 states (AL, FL, GA, KY, LA, MO, MS, and SC) that provided some estimate of hunter numbers (derived variously through license sales or hunter surveys) to test for independence between hunter numbers and spring harvest. We did not include Texas in this analysis because hunter numbers and total harvest are calculated for a number of species based on hunter survey data, including eastern wild turkeys in the portion of the state in which they are extant. However, harvest data of eastern wild turkeys in Texas are also collected each year based on the mandatory check-in of harvested birds. Estimates of eastern wild turkeys harvested each spring in Texas derived from the hunter survey differed considerably from the corresponding number of turkeys reported from mandatory check-ins (which we assume to be more accurate). This, combined with the fact that estimates of hunter numbers calculated from survey data exhibited very large confidence intervals, leads us to believe that these estimates are inaccurate. We suspect the lack of accuracy stems from a low sample size of hunter response in the relatively small area of the state in which eastern wild turkeys are hunted. We found that correlations between hunter numbers and spring harvest were high in all states, and that hunter numbers were a useful predictor of total harvest (Table 3). The implication is that caution should be used in extrapolating information regarding population trends from estimates of total harvest alone, as these estimates may be more representative of hunter participation than wild turkey population density. Theoretically, accounting for hunter effort should provide a more accurate link between harvest data and turkey population density. If turkey populations are experiencing noticeable long-term changes in density to an extent that hunting quality is affected, then there should be corresponding changes in harvest success. To test for this, we modeled trends in spring hunter success (defined as the % of hunters to harvest ≥ 1 turkey), or harvest proportion (defined as total turkeys harvested ÷ turkey hunters) for 7 states that reported such information directly, or in which we were able to calculate these metrics ourselves from data provided. In Florida we chose to model trends in harvest success rather than harvest rates because this metric was based directly on the responses to hunter surveys, and thus did not introduce bias associated with an Table 3. Results of linear regression models (F statistics, degrees of freedom (df), P-values, and adjusted R2 values) of the influence of spring hunters on total spring harvest estimates for 8 states in the southeast. Sample size (n) is the number of years in which both estimates of harvest and hunter numbers were available. State n F df P- value adjusted R2 AL 41 120.0 39 <0.001 0.75 FL 21 112.4 19 <0.001 0.85 GA 9 11.4 7 0.012 0.57 KY 16 32.7 14 <0.001 0.68 LA 27 33.4 25 <0.001 0.55 MS 27 136.9 25 <0.001 0.84 MO 53 936.7 51 <0.001 0.95 SC 35 361.4 33 <0.001 0.91 estimation of hunter numbers. We modeled hunter success for Mississippi as well, as these values were reported annually for the years 1980 – 2009 in records maintained by Mississippi State University (Hunt 2010). We found that harvest trends that accounted for hunter effort were generally more stable than raw harvest trends (Figure 3). For instance, hunter success in Mississippi has consistently hovered around 50%, despite persistent declines in the numbers of total birds harvested. The regional implication of these trends is that broadly speaking, hunters participating in spring hunts have not experienced any significant historical declines in success as we measured it. Trends in data from the BBS were relatively consistent region-wide and suggested relative population increases to some degree over time in all states (Figure 4). The magnitude of increase varied considerably across states; Tennessee for instance exhibited a particularly sharp population increase beginning in 2000. These trends were least pronounced in Louisiana and Mississippi. BBS estimates for any given geographic region are partially contingent on route coverage and birder interest as surveys are run by volunteers (Link and Sauer 2002). This likely explains to some degree the variation we observed among states in terms of the absolute values of annual indices. A lack of volunteer observers to sample routes in rural areas of Louisiana and Mississippi might partially account for the low trend estimates we observed in these states. We did not show BBS data for Oklahoma and Texas because the eastern subspecies only occurs in a small portion of each state. The BBS analysis is performed at the state level, thus it is impossible to separate the index of eastern wild turkeys from Rio Grande wild turkeys present in large portions of these states. Our findings demonstrate that there is evidence of long term declines in productivity (to varying degrees) region-wide; however, there is little evidence to support a congruent large scale decrease in overall population densities. Unfortunately, no robust way to directly measure wild turkey population densities on large scales presently exists. However, we can infer from the data available that at a regional scale, populations are presently in a state of relative increase or stabilization. We draw this inference from examination of harvest rates and hunter success, and trends in BBS indices. Annual BBS indices for any species are modeled as expected observations per route in a given state. Considering BBS routes are surveyed only once annually and that not all routes in a given state necessarily traverse ideal turkey habitat, the fact that turkeys are encountered with sufficient frequency so that relative abundance estimates have consistently increased range-wide strongly suggests turkey population increases over time. Possible exceptions may be southeastern Oklahoma and eastern Texas which represent the western margin of the eastern wild turkey’s range (Stangel et al. 1992). Populations existing at the distributional margin of a species range are often highly dynamic in nature (Williams et al. 2003, Guo 2005), so this is not unexpected. Eastern Texas in particular has had a history of difficulty in introducing and maintaining stable populations. Demographic trends and landscape change We originally postulated that large-scale landscape changes may have acted as potential drivers of the observed demographic changes through time. However, we decided to reduce effort along this line of inquiry after our initial investigations indicated that this was not a fruitful path to follow. For the time periods in which large-scale land cover data were available, we found only slight variations in relevant habitat types, and the observed demographic trends 1) appeared to operate independent of landscape and 2) suggested that some metrics were statistically different that were biologically implausible. Furthermore, given the general contrasting trends in the various demographic parameters within states, attempting to draw inferences on the relationship between land cover and demography would be problematic because inferences could contradict depending on what dependent variable was considered. We illustrate this for MO and TN. The most suitable data for quantifying long term trends in land cover change was the USGS National Land Cover Database (NLCD, Homer et al. 2012). The NLCD characterizes land cover at 30m resolution across the conterminous United States based primarily on Landsat Enhanced Thematic Mapper satellite imagery, and is available for the years 1992, 2001, and 2006. For each year, we calculated the percentage forested and urbanized land respectively for each state based on the percentage of cells classified as each land cover type. To ensure consistent classification methodology between years, we used the 1992/2001 Retrofit Land Cover Change Product, which updated the classification methodology of the 1992 data for accurate comparison with the later datasets. All NLCD data was accessed from the Multi-resolution Land Characteristics Consortium (MRLC) website (http://www.mrlc.gov/). We plotted changes in the percentage of forest cover and urbanization for each state from 1992 – 2006, along with respective trends in PPH ratio, spring harvest, and BBS index. We found that from 1992 – 2006, total forest cover and urbanization changed by < 1% within each state (Figure 5). During the same period, PPH ratios, spring harvest numbers, and BBS indices exhibited considerable changes as detailed herein. Thus, we can infer that at the state-level, land cover as quantified by the NLCD was not responsible for the observed changes in demographic parameters. This is not to say that habitat has no bearing on turkey populations, but these effects are likely localized, and as such are not detectable or appropriate for causative statistical analysis at the large spatial scale in which we are concerned. Characterization of productivity declines Before we could develop hypotheses to explain productivity declines, it was necessary to characterize the nature of these declines, which could be the result of 2 things: 1) an increasing proportion of females being observed without broods, or 2) a decline in the average size of observed broods. This is important because, while not necessarily mutually exclusive, each scenario suggests a different set of possible mechanisms underlying the observed trends. A proportional increase in the number of females observed without broods may suggest that nesting success, or the proportion of females attempting to nest, has been declining. If however the proportion of females observed with broods has remained relatively stable while brood sizes have declined, this may be indicative of declining poult survival or decreased fecundity. Eight states reported the annual percentage of females observed without broods, or provided raw data that allowed us to calculate this value. We observed that, especially in the last 15 years, the percent of females observed without young has generally increased through time in all states (Figure 6). The strength of this trend varied to some extent among states, but in 5 states the ranges in model predicted estimates approached or exceeded 20% (MS = 18.7%, LA = 19%, MO = 25.8%, OK = 26.8%, TN = 28.8%). Since 1995, the only state to show evidence of a declining trend was Missouri, and only since 2010. The implication is that observed declining trends in PPH ratios have been at least in part due to an increasing percentage of females observed without broods, suggesting a general historic decline in per capita nesting success is likely partially responsible for observed declining trends in productivity. Brood sizes were harder to estimate based on the nature of the available data. Simply calculating the annual ratio of total poults to total females based on observations of females with young was superficially appropriate. However, there were several confounding factors with this method that have potential to introduce considerable bias into the results. Among them is the fact that females without broods are known to travel with females tending broods, and the fact that this method assumed that observed young were distributed evenly among all females. Therefore, we reasoned that the most accurate way to measure mean brood size was to rely solely on observations of single females with young. For states that provided raw data that included records of individual observations, we estimated the mean annual brood size and corresponding 95% confidence interval based on observations of single females with ≤ 16 poults. We chose 16 poults because based on our personal field experiences and records reported in the literature, clutch sizes > 16 are exceptionally rare. Thus, we assumed that observations of single females with broods of more than 16 poults likely represented an erroneous or incomplete observation. On the surface, there appeared to be a general negative trend, especially for states with long term data availability (Figure 7). However, the scale of variation was rather small, with the difference between the largest and smallest mean brood sizes < 2 poults for all states except Oklahoma and West Virginia. Likewise, there was considerable overlap in confidence intervals among years. The most persistent downward trend appeared in West Virginia, although sample sizes early in the historical record were generally low, which is reflected by the wide confidence intervals. Thus, the apparent decline in mean brood size from very high values in the early part of the West Virginia record was likely influenced to some degree by sampling effort. There appears to be little evidence that changes in brood sizes have occurred through time in the southeast, suggesting that wild turkey populations are not experiencing any significant large-scale reductions in fecundity or poult survival. Productivity and restoration efforts The fact that turkey populations exist in many parts of the southeast is largely the result of intensive large scale restoration efforts. Restoration was largely accomplished through the live release of turkeys wild captured from extant populations into areas in which turkeys were absent (or existed at very low densities) that contained suitable habitat (Kennamer et al. 1992). Given the large scale introductions and dispersals of birds into new areas, we wanted to investigate whether releases conferred a noticeable signal in productivity trends. If a consistent signal was found, this would provide reason for us to examine the relationship between restoration efforts and productivity in more detail. We tallied the total number of birds released each year for each of the 6 states that were able to provide detailed restocking information. We determined when restoration was 50%, 75%, and 95% complete in each state based on total cumulative releases. For example, restoration was considered 50% complete when 50% of the total amount of birds released during all years of restoration was reached. When comparing restoration efforts to productivity trends, we noted that declines in productivity began prior to restoration efforts reaching 50% completion in states that have historical productivity data extending back beyond that point (Figure 8). Restoration efforts in Georgia, Missouri, North Carolina, and Tennessee were characterized by time periods in which restocking activity was especially intense. For instance, in North Carolina there was a clear peak in releases in the 1990’s. However, in all of these states the major downward trend in productivity began prior to these intensive restoration efforts and continued unaltered for the duration of the restocking years. Restocking efforts in Louisiana and Mississippi did not exhibit the clear spikes in activity observed in other states. Louisiana is the only state in which productivity declines began after restoration was 95% complete. Overall, no clear pattern existed to suggest a link between the intensity of restoration efforts, or a consistent change following the completion of restoration, on productivity trends. Relationship between productivity and population size It appears that large-scale declines in productivity have occurred concomitant with overall increasing and/or stabilizing population trends. To illustrate this, we plotted GAM predicted productivity (PPH) values as a function of relative population size (BBS indices). To account for differing scales among states, we first scaled PPH ratios and BBS indices from 0 – 1 for each state, with 0 being the lowest value of each metric, and 1 being the largest value of each metric for each respective state. We observed a clear negative relationship between population size and productivity in all states (Figure 9). There are 2 potential conditions which may allow turkey populations to remain stable or increase in the face of low levels of reproduction. High survival rates of adults, especially females. Long-lived females theoretically have more opportunities to reproduce during their lifetime. To maintain a population each female would have to successfully reproduce only once in her lifetime. High recruitment rates of young into the adult population. To maintain a stable population, the number of young recruited in the breeding population need only equal the number of adult deaths. Recruitment above this threshold would act as a buffer, and potentially lead to population increases. If high female survival is playing a part in maintaining populations, we would expect to see some evidence of a historical trend towards increasing female survival to counter-act declining productivity. To explore this potential, we reviewed the literature of studies that reported female survival. We limited our review to studies which used robust statistical methods such as KaplanMeier (Pollock et al. 1989), Heisey-Fuller (Heisey and Fuller 1985), or known-fate survival models in program MARK (White and Burnham 1999) to derive annual survival estimates from radio-telemetry data (Table 4). Additionally, we queried researchers currently engaged in studies of female survival to provide us preliminary results (Table 4). To bolster sample sizes, we considered studies that spanned the entire geographic range of the eastern subspecies. To investigate how survival may have changed over time, we binned studies into one of 3 time periods; 1980 – 1989, 1990 – 1999, and ≥ 2000 and recorded the mean annual survival estimate for each study. For multiyear studies that spanned across time bins and reported annual survival estimates for individual years, we classified years into their respective time period and calculated a mean annual survival rate. For example, Miller et al. (1998a) reported annual survival rates during 1984 -1994. In this case, we calculated a mean survival rate for the years 1984 – 1989, and for the years 1990 – 1994. If a study spanned 2 time periods but did not report individual survival rates for each year, we counted the study as belonging to the time period in which most of the study occurred. For example, Moore et al. (2010) reported only a single survival rate estimate from a study that spanned 1998 – 2000. Because most of the study occurred prior to 2000, we grouped this study into the 1990’s. Additionally, in cases in which successive studies of specific study sites built on and incorporated date reported in earlier studies, we only used survival estimates reported in the most recent study. For example, Byrne (2011) incorporated date reported in Wilson et al. (2005), and as such only the results from Byrne (2011) were used. We found evidence that female survival rates have generally increased over time, from an average annual survival rate of 0.51 (range: 0.44 - 0.61) for studies in the 1980’s to 0.68 (range 0.58 - 0.78) for studies in the 2000’s (Figure 10). While the notion of low productivity and simultaneous stability in wild turkey populations may initially seem counter-intuitive, this relationship has been discussed in the turkey literature before. Most notably, Vangilder et al. (1987) used modeling approaches to conclude that even when female success rates (defined as the portion of females alive each spring to successful hatch a brood) were as low as 30-40%, high population densities could be maintained as long as annual female survival rates averaged ~ 0.44. Table 4. Studies reporting annual survival estimates of female eastern wild turkeys used in tracking survival trends over time. All studies derived survival rates based on radio telemetry data. Study years Study location Landscape Reference 1981 – 1989 MO Mixed forest/ag Vangilder and Kurzejeski 1995 1984 – 1985 MO Mixed forest/ag Kurzejeski et al. 1987 1984 – 1994 MS Mixed hardwood/pine Miller et al. 1998a 1987 – 1990 MS Pine plantation Palmer et al. 1993 1988 – 1994 WI Mixed forest/ag Wright et al. 1996 1990 – 1993 NY Mixed forest/ag Roberts et al. 1995 1990 – 1993 MO Mountain hardwood Vangilder 1995 1990 – 1994 VA/WV Mountain hardwood Pack et al. 1999 1993 – 1996 IA Mixed forest/ag Hubbard et al. 1999a 1998 – 2000 SC Coastal pine Moore et al. 2010 2002 – 2006 OH Hardwood Reynolds and Swanson 2010 2002 – 2010 LA Bottomland hardwood Byrne 2011 2003 – 2005 IN Agricultural Humberg et al. 2009 2010 – 2012 DE Pine plantation J. Bowman, pers com 2012 AR Mixed hardwood/pine T. Pittman, pers com Previous studies have detailed poult survival rates through the summer brood-rearing period (Vangilder et al. 1987, Miller et al. 1998b, Godfrey and Norman 1999, Hubbard et al.1999b, Thogmartin and Johnson 1999, Norman et al. 2001). While studies such as these offer possibilities to draw some indirect inferences regarding recruitment, there are no studies that we are aware of that directly quantified eventual recruitment into the breeding population. The dearth of information in the literature on this subject prevents us from being able to comment on historical and current trends in realized recruitment, except to postulate that it has likely declined in conjunction with productivity declines. Clearly, this is a fertile and important research area in which to focus future efforts. Potential mechanism driving productivity trends The negative correlation between population and productivity indices leads us to hypothesize that the large-scale historical declines in productivity observed in the southeast are evidence that reproduction is mediated in a density-dependent manner. The idea of population self-regulation (i.e. density-dependence) is one of the oldest and most well established aspects of population ecology (Turchin 2001), and the phenomena is ubiquitous across a wide range of organisms (Brook and Bradshaw 2006). It is physically impossible for any population to continue to grow indefinitely, thus increasing population densities are expected to trigger negative feedback loops that limit population growth. As such, it is expected that a negative relationship should exist between population size and the rate of population increase. A review by Guthery and Shaw (2013) points out that evidence of density dependence in upland game birds has existed in the literature since the 1940’s. More recently, Porter et al. (1990) and McGhee and Berkson (2007) both found evidence of density-dependent population growth in wild turkey populations based on analyses of various harvest indices. One way in which density-dependence may manifest itself is through decreased recruitment. Density-dependent effects on reproduction have been documented across a variety of avian taxa through both experimental (Dhondt et al. 1992, Both 1998, Pöysä and Pöysä 2002, Sillett et al. 2004, Brouwer et al. 2009) and observational studies (Larson and Forslund 1994, Ferrer and Donazar 1996, Bennetts et al. 2000, Armstrong et al. 2005, Carrete et al. 2006). Negative relationships between population density and reproduction in gallinaceous birds were documented in the literature as early as the 1940’s; specifically Errington (1945) found such a relationship in both northern bobwhite (Colinus virginianus) and ring-necked pheasant (Phasianus colchicus) populations. The existence of a density-dependent effect on wild turkey reproduction has been hypothesized by several authors who have noted lower productivity in populations considered stable compared to recently introduced and expanding populations (Vangilder et al. 1987, Vander Haegan et al. 1988, Miller et al. 1998b). In the present study, we observed this phenomenon across an extensive spatial and temporal scale. The influence of stochastic environmental variables such as rainfall and temperature on short term annual variations in reproduction is well documented in the literature on eastern wild turkeys (Warnke and Rolley 2005). However, the functional relationship between population density and reproduction, which is more likely to drive long term trends, has not been explored. Despite the observational evidence, there is no experimental evidence of a density-dependent relationship between population density and reproduction for wild turkeys, and as a result traditional models of wild turkey population dynamics have assumed no such relationships (Roberts and Porter 1995, Vangilder and Kurzejeski 1995, Rolley et al. 1998, Alpizar-Jara et al. 2001). Here we present a theoretical mechanism in which density-dependent reproduction may operate on wild turkey populations that ties together reproduction, survival, and potential limiting aspects of landscapes. This theory should generate a set of testable hypotheses and highlight particular aspects of wild turkey ecology in which our knowledge is lacking. We hope our thoughts ultimately will help drive future wild turkey research. To understand how density-dependence potentially works in wild turkey populations, it is important to consider life history traits of turkeys. Wild turkeys exhibit life-history traits commonly associated with r-selected species, such as early age of maturity, short life spans, high reproductive potential, and an association with dynamic and/or early successional environments (Stearns 1977). Species with these life-history traits are hypothesized to exhibit their greatest population growth rates at low population densities relative to environmental carrying capacity (K, Fowler 1981). This hypothesis is supported for wild turkeys by the findings of Porter et al. (1990) and McGhee and Berkson (2007), who both found population growth to be greatest at low densities. A correlate of this is that the highest levels of recruitment would thus be expected to be associated with these periods of high growth occurring at low densities, and the historical trends in the southeast largely conform to this pattern. Female mortality rates are often observed to exhibit seasonal variation. In most studies that have detected such seasonal variation, the lowest survival rates are often associated with reproductive seasons (Palmer et al. 1993, Vander Haegen et al. 1988, Wright et al. 1996). Female survival outside of reproductive seasons, assuming harvest (legal or illegal) is not a major cause of mortality, can be quite high (i.e. Pack et al. 1999). This makes intuitive sense, as incubation and brood-rearing activities are behaviors that ostensibly leave females especially vulnerable to predation. Beyond this, it is reasonable to assume that the physiological stress associated with reproduction may extend the increased mortality risk of reproductively successful females beyond the reproductive seasons. Two general mechanisms to explain how density-dependent reproduction arises in populations have been hypothesized. The first, which may be termed the interference hypothesis, considers that in territorial species as population density increases the frequency of agonistic encounters between individuals competing for space results in a hostile social environment. As a result of the increased energy expended towards territorial behavior at the expense of reproduction, reproductive output is uniformly decreased across the landscape with little variation across individuals (Dhondt and Schillemans 1983, Sillett et al. 2004). The second hypothesis is the concept of site-dependent population regulation (also termed the habitat heterogeneity hypothesis) which suggests that as population density increases a progressively larger proportion of the population is forced into using low-quality habitat, resulting in declines in per capita reproductive success (Rodenhouse et al. 1997, McPeek et al. 2001). Under this paradigm, while per capita reproductive output declines, variation among individuals is high as individuals that are able to access high-quality habitat reproduce more successfully than individuals in marginal habitat. Wild turkey males play no part in nesting or brood rearing, and no territorial behavior between nesting females has been documented. Additionally, the observed trends of reduced PPH ratios concomitant with variable proportions of females observed without young suggests a per capita decrease in recruitment along with increasingly variable nesting success among individuals. As such, we believe that density-dependent reproduction in wild turkeys is most likely to be a form of site-dependent population regulation. To understand how site-dependent population regulation may influence the reproductive output of wild turkeys, consider that wild turkeys inhabit heterogeneous landscapes. A natural consequence of this is that the quality of nesting habitat in any given system is heterogeneous as well, ranging from high quality to unsuitable. Thus, only a certain portion of any landscape will be high-quality nesting habitat. When we refer to high-quality nesting habitat we are specifically referring to habitats that are conducive to successful nesting attempts. Given that predation is often identified as the primary cause of nest loss (Vangilder et al. 1987, Vander Haegen et al. 1988, Miller et al. 1998b, Paisley et al. 1998, Byrne and Chamberlain 2013), it may be reasonable in this sense to assume that high-quality areas are those associated with low nest predation risk. At low population densities, a large proportion of the breeding population will be able to access what high quality nesting habitat is available. As a result, per capita recruitment would be expected to be high as per capita nesting success is also expected to be high. As the population continues to grow the proportion of females forced to nest in sub-optimal habitat grows as well. At high population densities, a small proportion of the total female population will be able to nest in these high-quality areas, with the remainder forced to attempt to nest in lower quality areas. As a result per capita recruitment is reduced, as nest-loss is significant for the proportion of the population nesting in sub-optimal habitat. This would be expected to be reflected in annual summer brood counts. After allowing for some degree of annual variation resulting from density-independent environmental factors (temperature, precipitation, etc.), the absolute numbers of females successfully producing poults may be relatively consistent across years. However, despite increasing population density the absolute number of successful females remains relatively static, and the proportion of the population that experienced failed nesting attempts rises. The expected result is what we see in the historical trends; reduced PPH ratios concomitant with an increasing percentage of the female population observed without broods. The decline in per capita nesting success as a result of increased population density should precipitate a rise in female annual survival rates. This is because females that experience early nest loss, or alternatively do not attempt to nest at all, are spared the increased predation risk associated with reproductive activities. Collier et al. (2009) documented the negative influence of reproductive effort on survival for Rio Grande wild turkeys (M. gallopavo intermedia). As the proportion of these reproductively unsuccessful females increases, the overall population-level survival rate increases. In populations subject to density-dependent regulation, this inverse relationship between reproduction and survival is a mathematical necessity to maintain stable populations (Guthery and Shaw 2013). In this regard, a stable population is one that has occupied the available habitat and shows no trend towards an increasing or decreasing population size, although year to year variability may be substantial. Essentially, this refers to a population that has reached a saturation point, and in which annual demographic variation is ruled primarily by stochastic density-independent factors (i.e. weather). To maintain such populations, survival and reproduction cannot increase in tandem, as doing so would lead to biologically implausible geometric population increases over time. Conversely, concomitant persistent declines in reproduction and survival would result in population crashes. Thus, to maintain stable populations, survival and reproduction rates must be negatively correlated (Guthery and Shaw 2013). We have formulated this theory as a framework to explain patterns in the observed historical demographic trends in the southeast based on our knowledge of wild turkey biology and our personal research experience, as well as general concepts of population ecology. From this we are able to develop a number of specific hypotheses based on assumptions we used that could be explicitly tested via properly designed studies. The results of such studies would provide evidence to support or challenge our hypotheses, and ultimately serve to better elucidate the mechanisms linking density and reproduction. The key hypotheses are: 1) For any given landscape, population-level reproductive performance is negatively correlated with population density. 2) For a given landscape, the mean annual survival rate of females is positively correlated with population density. 3) For a given landscape, females that do not attempt to reproduce or experience early nest failure have greater survival rates than females that are reproductively active. 4) If quality nesting areas are settled first as site-dependent regulation would suggest, then for a given landscape with all else being equal, there should exist a negative correlation between nest initiation date and nest success when population densities are high. This is to say that early nesting females will first settle and occupy limited available quality habitat, and as a result late nesting females will be forced into marginal habitats. The strength of this relationship should be weak when population densities are low. Well-designed studies of an experimental nature are required to specifically test the veracity of these hypotheses; however, there are still inferences we can draw from published observational studies. The results of research conducted on Sherburne WMA (hereafter Sherburne), located in the Atchafalaya Basin in southern LA offers an interesting case-study. Intensive radio-telemetry studies of wild turkeys on this area occurred from 2002 – 2010 (Grisham et al. 2008, Byrne 2011, Byrne and Chamberlain 2013). Studies specifically concerned with nesting ecology and female survival occurred during the years 2002-2005, and 2007-2010 (Byrne 2011, Byrne and Chamberlain 2013). At the time of these studies, before the catastrophic flooding that occurred in May 2011, turkey densities on Sherburne were considered among the highest in the state (J. Stafford, pers. comm.). Female success rates (percentage of females alive each spring to successfully hatch ≥ 1 egg) were among the lowest reported in the literature at 24%, as were nesting rates (percentage of females alive each spring to successfully reach incubation) at 60% (Byrne and Chamberlain 2013). However, nest success rates (percentage of known nests in which ≥ 1 egg was successfully hatched) at 39% were relatively consistent with the range-wide average (Byrne and Chamberlain 2013). These results broadly conform to several predictions of site-dependent population regulation. For example, based on female success rates, we see that per capita reproductive success is low, as hypothesized for a high density population (hypothesis 1). Quality nesting habitat appeared to be of limited availability in this bottomland hardwood system, where dense canopy coverage and persistent seasonal floods resulted in a relatively sparse understory with little available ground-level cover. In a high density population scenario, we expect only a small proportion of the females to be able to access what high quality nesting cover was available, and this appears to be reflected in the low nesting rates. The implication is that a large number of females were forced to attempt nesting in sub-optimal habitat which resulted in a high percentage of nest loss prior to incubation and subsequent observer detection. However, nest success rates suggest that the females that were able to access suitable nesting habitat were as successful in hatching a brood as females in other landscapes. As expected by hypothesis 2, with low per capita reproductive success and a high population density mean annual female survival rates were relatively high at 0.58±0.06 (Byrne 2011). Additionally, Byrne (2011) found that annual survival rates of reproductively active females (defined as those that reached nest incubation) were lower than non-reproductive females as expected by hypothesis 3. Most known female mortalities (67.7%) occurred during the nesting or brood-rearing seasons. Similarly, Grisham et al. (2008) found male annual survival rates on Sherburne to be high as well (0.64±0.06), although this is likely at least partly due to conservative spring harvest regulations. These studies provide some support for the idea of a negative correlation between reproduction and survival, and suggest that despite low reproductive rates, high female survival rates allowed the population to maintain itself at relatively high densities. As such we believe that what we observed at Sherburne can be interpreted as a microcosm of the historical trends observed on larger scales in the present study study. The few studies that have examined the relationship between incubation start date and nesting success have found that earlier nesting attempts are generally more successful than later attempts (Thogmartin and Johnson 1999, Norman et al. 2001). Recent work in Georgia has shown Julian date of nest initiation to be the strongest determinant of nest survival (Little et al. in review). This suggests support for hypothesis 4, which predicts that earlier settling females are able to access the highest quality habitat, which translates into higher rates of nest success for those individuals. While there are numerous examples in the literature that seemingly provide some level of support for our hypotheses, and our proposed mechanism of site-dependent population regulation in general, we stress the need for future research aimed at specifically testing these hypotheses. Ideally these studies would allow for some level of experimental manipulation of wild turkey population densities, while controlling for extraneous variables as much as possible. Management Implications Understanding the mechanisms by which density-dependence operates in wild turkeys would provide tangible benefits to both habitat and harvest management applications. Perhaps the most consequential implication of density-dependence is that in populations that have reached the point of stability and self-regulation some management actions may be in some respects self-defeating. This is due to the fact that actions aimed at increasing reproduction will necessarily entail a decrease in survival, and vice versa; as discussed above this is a necessary relationship in maintaining stable populations. In the case of habitat management for instance, schemes aimed at increasing nesting and brood-rearing habitat under the assumption that increasing productivity will necessarily translate to increasing population sizes may be naïve. In populations that have already occupied the available habitat and stabilized, such a strategy may increase reproduction but we would expect those gains to be counter-acted by associated decreases in survival. The end result is that even with increased productivity the population remains stable (in terms of numbers at least, age ratios may very well change). Population growth could potentially be stimulated by providing new habitat for turkey populations to expand into. Essentially, creating new habitat (as opposed to managing or improving habitat) equates to lowering population densities in relation to habitat availability, and we would expect populations to increase as new habitat is colonized. However, population growth will only increase until the new stabilization point is reached, at which point self-regulatory processes again take control. The realization of the importance of accounting for density-dependence in regards to harvest management strategies has been increasingly recognized in recent years (Warnke and Rolley 2005, McGhee et al. 2008). By accounting for density-dependence in population models, managers could theoretically use a yield curve paradigm to estimate expected yield for a given population given its size relative to K. Managers could use this information to determine the relative population density that maximizes yield, and could enact harvest management practices to maintain populations at said densities (McGhee et al. 2008). Population density relative to K should factor prominently into management goals as practices that are effective at low population densities may not provide similar benefits at high densities, and vice versa. While it may not be feasible to measure population density directly, an understanding of the mechanistic workings of density-dependence would allow sufficient inferences to be drawn on the status of a given population by tracking the relationships between demographic parameters such as reproduction, survival, and harvest rates as a function of hunter effort. Furthermore, an understanding of the processes that regulate turkey populations, and how these processes may vary with respect to population density, would allow for further refinement of population models. This would ostensibly allow for the generation of more precise predictive models, and promote more informed decisions regarding harvest management. Additional research questions and needs In the course of analyzing and interpreting data for this study, we were able to identify several additional areas in which information is limited and where we believe future work would be beneficial. 1: Measures of population density The lack of a universally accepted and accurate tool to measure wild turkey population size is a problem that was first identified several decades ago (Mosby 1967). As an absolute count of turkey numbers is not practical, a variety of methods have been used to attempt to derive an index of population density. These include the total numbers of females and poults in summer brood counts (Wunz and Shope 1980), gobbling counts (Porter and Ludwig 1980), observations of summer bait-site use (Hayden 1985, Weinstein et al. 1995), winter flock surveys (i.e. Weinrich et al. 1985), capture-mark-recapture modeling based on band returns (Lint et al. 1995) and camera trap surveys (Olsen et al. 2010). While some of these indices have shown some degree of correlation with harvest, the veracity of these methods has not been tested in respect to actual population size. Given the importance that population density may have on determining demographic changes through time, research aimed at assessing the accuracy of common indices to measure density, or developing new methods to better estimate population size are paramount. 2: Immigration, emigration, and metapopulation dynamics Within-patch demographic parameters regarding reproduction and survival are commonly studied aspects of wild turkey populations. However, population dynamics are also influenced by between-patch processes of immigration and emigration, and to this end little work has been done focused on wild turkeys. The vast majority of turkey demographic studies ignore the potential roles of dispersal and connectivity between populations, although these parameters may have important influences on population stability and fluctuations. This likely due in part to the fact that a majority of studies are focused on single, relatively small, study areas where these processes are unlikely to be detectable. Understanding connectivity across large ecological scales would aid in identifying source and sink populations, as well as important corridors that permit movements between populations. Likewise, researchers could then begin to quantify the permeability of landscapes to large scale movements, and to quantify how isolated specific populations are. This last point may be particularly relevant for populations in east Texas and southeast Oklahoma. Populations on the margins of a species distribution are generally more dynamic than populations towards the center. There is evidence to suggest that densitydependent population regulation affects marginal and interior populations similarly, but marginal populations are more susceptible to density-independent population fluctuations (Williams et al. 2003). This places an isolated, marginal population at greater risk of extinction, as a number of consecutive years of poor reproduction or survival may drive the population below the point from which it can recover. Theoretically, this risk could be mediated to some degree if connectivity existed to allow immigration from the larger population to “rescue” these marginal populations. It is a plausible hypothesis that the lack of connectivity to the larger population, rather than local habitat quality, may be the major factor underlying the historical difficulties in establishing populations in eastern Texas Figure 1: Historic trends in productivity, as measured by poult per hen ratios, of 12 states in the southeastern United States. Lines represent generalized additive model estimates (solid line) and 95% confidence intervals (dashed lines). Figure 2: Historic trends in spring harvest of male wild turkeys in the southeastern United States. Lines represent generalized additive model estimates (solid line) and 95% confidence intervals (dashed lines). No estimate provided for Georgia because data was not sufficient to fit a model. Figure 2 continued: Historic trends in spring harvest of male wild turkeys in the southeastern United States. Lines represent generalized additive model estimates (solid line) and 95% confidence intervals (dashed lines). Figure 3: Historic trends in hunter success or harvest rates of male wild turkeys during spring hunting seasons in the southeastern United States. Lines represent generalized additive model estimates (solid line) and 95% confidence intervals (dashed lines). Figure 4. Historic state-specific trends in USGS breeding bird survey indices for eastern wild turkeys in 13 states in the southeastern United States. Dashed lines represent 95% credible intervals. Figure 4 continued: Historic state-specific trends in USGS breeding bird survey indices for eastern wild turkeys in 13 states in the southeastern United States. Dashed lines represent 95% credible intervals. Figure 5. Percentage of urban and forest land cover for Missouri and Tennessee during the years 1992, 2001, and 2006 based on National Land Cover Data database, and concomitant trends in demographic variables; PPH ratio, spring harvest, and BBS index (red lines). Figure 6. Historic trends in proportion of females observed without broods during summer brood counts in 8 states in the southeastern United States. Lines represent generalized additive model estimates (solid line) and 95% confidence intervals (dashed lines). Figure 7. Mean brood size (± 95% C.I.) based on observations of single females with ≤ 16 poults for 7 states in the southeastern United States. Figure 8: Productivity trends as determined through poult per hen ratios (solid lines) and annual restoration efforts as measured through numbers of wild turkeys released within a given state (grey bars) for 6 states in the southeastern United States. Vertical dashed lines represent 50%, 75%, and 95% cumulative releases respectively. Figure 9. The relationship between annual indices of population density (measured through Breeding Bird Survey data) and productivity (measured through poult per hen ratios) for 11 states in the southeastern United States. Values for both indices are scaled from 0 (lowest values) to 1 (highest value) for each state. Figure 10: Mean (± SE) survival estimates of female wild turkeys from studies conducted on radio-marked individuals during the 1980s, 1990s, and after 2000. References to specific studies can be found in Table 4 Literature Cited Alpizar-Jara, R., E.N. Brooks, K.H. Pollock, D.E. Steffen, J.C. Pack, and G.W. Norman. 2001. An eastern wild turkey population dynamics model for Virginia and West Virginia. Journal of Wildlife Management 65:415-424. Armstrong, D.P., R.S. Davidson, J.K. Perrott, J. Roygard, and L. Buchanan. 2005. Densitydependent population growth in a reintroduced population of North Island saddlebacks. Journal of Animal Ecology 74:160-170. Bennetts, R.E., M. Fasola, H. Hafner, and Y. Kayser. 2000. Influence of environmental and density-dependent factors on reproduction of little egrets. The Auk 117:634-639. Both, C. 1998. Experimental evidence for density dependence of reproduction in great tits. Journal of Animal Ecology 67:667-674. Brooks, B.W., and C.J.A. Bradshaw. 2006. Strength of evidence for density dependence in abundance time series of 1198 species. Ecology 87:1445-1451. Brouwer, L, J.M. Tinbergen, C. Both, R. Bristol, D.S. Richardson, and J. Komdeur. 2009. Experimental evidence for density-dependent reproduction in a cooperatively breeding passerine. Ecology 90:729-741. Byrne, M.E. 2011. Influence of landscape characteristics on the nesting ecology of female wild turkeys and behavior of raccoons. Ph.D Dissertation, Louisiana State University, Baton Rouge, Louisiana. Byrne, M.E., and M.J. Chamberlain. 2013. Nesting ecology of wild turkeys in a bottomland hardwood forest. American Midland Naturalist 170:95-110. Carrete, M., J.A. Donazar, and A. Margalida. 2006. Density-dependent productivity depression in Pyrenean bearded vultures: implications for conservation. Ecological Applications 16:1674-1682. Collier, B.A., K.B. Melton, J.B. Hardin, N.J. Silvy, and M.J. Peterson. 2009. Impact of reproductive effort on survival of Rio Grande wild turkey (Meleagris gallopavo intermedia) hens in Texas. Wildlife Biology 15:370-379. Dhondt, A.A. and J. Schillemans. 1983. Reproductive success of the great tit in relation to its territorial status. Animal Behaviour 31:902-912. Dhondt, A.A., B. Kempenaers, and F. Adriaensen. 1992. Density-dependent clutch size caused by habitat heterogeneity. Journal of Animal Ecology 61:643-648. Errington, P.L. 1945. Some contributions of a fifteen-year local study of the northern bobwhite to a knowledge of population phenomena. Ecological Monographs 15:1-34. Ferrer, M., and J.A. Donazar. 1996. Density-dependent fecundity by habitat heterogeneity in an increasing population of Spanish imperial eagles. Ecology 77:69-74. Fowler, C.W. 1981. Density dependence as related to life history strategy. Ecology 62:602-610. Godfrey, C.L., and G.W. Norman. 1999. Effect of habitat and movement on wild turkey poult survival. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 53:330-339. Guo, Q., M. Taper, M. Schoenberger, and J. Brandle. 2005. Spatial-temporal population dynamics across species range: from centre to margin. Oikos 108:47-57. Guthery, F.S., and J.H. Shaw. 2013. Density dependence: applications in wildlife management. Journal of Wildlife Management 77:33-38. Grisham, B.A., M.J. Chamberlain, and F.G. Kimmel. 2008. Spatial ecology and survival of male wild turkeys in a bottomland hardwood forest. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 62:70-76. Hayden, A.H. 1985. Summer baiting as an indicator of wild turkey population trends and harvest. Proceedings of the National Wild Turkey Symposium 5:245-252. Healy, W.M. 1992. Population influences: environment. Pages 129-143 in J.G. Dickson, editor. The wild turkey: biology and management. Stackpole Books, Harrisburg, Pennsylvania. Heisey, D.M., and T.K. Fuller. 1985. Evaluation of survival and cause-specific mortality rates using telemetry data. Journal of Wildlife Management 49:668-674. Homer, C.H., J.A. Fry, and C.A. Barnes. 2012. The National Land Cover Database. U.S. Geological Survey Fact Sheet 2012-3020. Hubbard, M.W., D.L. Garner, and E.E. Klaas. 1999a. Factors influencing wild turkey hen survival in southcentral Iowa. Journal of Wildlife Management 63:731-738. Hubbard, M.W., D.L. Garner, and E.E. Klaas. 1999b. Wild turkey poult survival in southcentral Iowa. Journal of Wildlife Management 63:199-203. Hunt, K.M. 2010. Trends in resident and non-resident hunter effort and harvest for turkey in Mississippi: 1980 to the present. Human Dimensions and Conservation Law Enforcement Laboratory Technical Document #143, Mississippi State University, 18 pp. Humberg, L.A., T.L. Devault, and O.E. Rhodes, Jr. 2009. Survival and cause-specific mortality of wild turkeys in northern Indiana. American Midland Naturalist 161:313-322. Kennamer, J.E., M. Kennamer, and R. Brenneman. 1992. History. Pages 6-17 in J.G. Dickson, editor. The wild turkey: biology and management. Stackpole Books, Harrisburg, Pennsylvania. Kurzejeski, E.W., L.D. Vangilder, J.B. Lewis. 1987. Survival of wild turkey hens in north Missouri. Journal of Wildlife Management 51:188-193. Larsson, K., and P. Forslund. 1994. Population dynamics of the barnacle goose Branta leucopsis in the Baltic area: density-dependent effects on reproduction. Journal of Animal Ecology 63:954-962. Link, W.A., and J.R. Sauer. 2002. A hierarchical analysis of population change with application to cerulean warblers. Ecology 83:2832-2840. Lint, J.R., B.D. Leopold, and G.A. Hurst. 1995. Comparison of abundance indexes and population estimates for wild turkey gobblers. Wildlife Society Bulletin 23:164-168. McGhee, J.D., J.M. Berkson. 2007. Estimation of a nonlinear density-dependence parameter for wild turkey. Journal of Wildlife Management 71:706-712. McGhee, J.D., J. Berkson, D.E. Steffen, and G.W. Norman. 2008. Density-dependent harvest modeling for the eastern wild turkey. Journal of Wildlife Management 72:196-203. McPeek, M.A., N.L. Rodenhouse, R.T. Holmes, and T.W. Sherry. 2001. A general model of sitedependent population regulation: population-level regulation without individual-level interactions. Oikos 94:417-424. Miller, D.A., L.W. Burger, B.D. Leopold, G.A. Hurst. 1998a. Survival and cause-specific mortality of wild turkey hens in central Mississippi. Journal of Wildlife Management 62:306-313. Miller, D.A., B.D. Leopold, G.A. Hurst. 1998b. Reproductive characteristics of a wild turkey population in central Mississippi. Journal of Wildlife Management 62:903-910. Moore, W.F., J.C. Kilgo, W.D. Carlisle, D.C. Guynn, Jr., J.R. Davis. 2010. Nesting success, nest site characteristics, and survival of wild turkey hens in South Carolina. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 64:24-29. Mosby, H.S. 1967. Population dynamics. Pages 113-136 in O.H. Hewitt, editor. The Wild Turkey and its Management. The Wildlife Society, Washington, D.C. Norman, G.W., J.C. Pack, C.I. Taylor, D.E. Steffen, K.E. Pollock. 2001. Reproduction of eastern wild turkeys in Virginia and West Virginia. Journal of Wildlife Management 65:1-9. Olsen, J.T., M.E. Sunquist, S.Z. Nomani, and M.K. Oli. 2010. Evaluation of camera-traps as a population estimation technique for wild turkeys. Proceedings of the National Wild Turkey Symposium 10:101-109. Pack, J.C., G.W. Norman, C.I. Taylor, D.E. Steffen, D.A. Swanson, K.H. Pollock, and R. Alpizar-Jara. 1999. Effects of fall hunting on wild turkey populations in Virginia and West Virginia. Journal of Wildlife Management 63:964-975. Paisley, R.N., R.G. Wright, J.F. Kubisiak, and R.E. Rolley. 1998. Reproductive ecology of eastern wild turkeys in southwestern Wisconsin. Journal of Wildlife Management 62:911-916. Palmer, W.E., G.A. Hurst, J.E. Stys, D.R. Smith, J.D. Burk. 1993. Survival rates of wild turkey hens in loblolly pine plantations in Mississippi. Journal of Wildlife Management 57:783789. Pollock, K.H., S.R. Winterstein, C.M. Bunck, and P.D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Porter, W.F., and J.R. Ludwig. 1980. Use of gobbling counts to monitor the distribution and abundance of wild turkeys. Proceedings of the National Wild Turkey Symposium 4:6168. Porter, W.F., D.J. Gefell, and H.B. Underwood. 1990. Influence of hunter harvest on the population dynamics of wild turkeys in New York. Proceedings of the National Wild Turkey Symposium 6:188-195. Pöysä, H., and S. Pöysä. 2002. Nest-site limitation and density dependence of reproductive output in the common goldeneye Bucephala clangula: implications for the management of cavity-nesting birds. Journal of Applied Ecology 39:502-510. R Core Team. 2013. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/. Reynolds, M.C., and D.A. Swanson. 2010. Survival of female wild turkeys in southeastern Ohio. Proceedings of the National Wild Turkey Symposium 10:149-155. Roberts, S.D., and W.F. Porter. 1995. Importance of demographic parameters to annual changes in wild turkey abundance. Proceedings of the National Wild Turkey Symposium 7:15-20. Roberts, S.D., J.M. Coffey, and W.F. Porter. 1995. Survival and reproduction of female wild turkeys in New York. Journal of Wildlife Management 59:437-447. Rodenhouse, N.L., T.W. Sherry, and R.T. Holmes. 1997. Site-dependent regulation of population size: a new synthesis. Ecology 78:2025-2042. Rolley, R.E., J.F. Kubisiak, R.N. Paisley, and R.G. Wright. 1998. Wild turkey population dynamics in southwestern Wisconsin. Journal of Wildlife Management 62:917-924. Sillett, T.S., N.L. Rodenhouse, and R.T. Holmes. 2004. Experimentally reducing neighbor density affects reproduction and behavior of a migratory songbird. Ecology 85:24672477. Stangel, P.W., P.L. Leberg, and J.I. Smith. 1992. Systematics and population genetics. Pages 1828 in J.G. Dickson, editor. The wild turkey: biology and management. Stackpole Books, Harrisburg, Pennsylvania. Stearns, S.C. 1977. The evolution of life history traits: a critique of the theory and a review of the data. Annual Review of Ecology and Systematics 8:145-171. Tapley, J.L., M.A. Hatfield, R.K. Abernethy, and J.E. Kennamer. 2010. Status and distribution of the wild turkey in 2009. Proceedings of the National Wild Turkey Symposium 10:19-30. Thogmartin, W.E., and J.E. Johnson. 1999. Reproduction in a declining population of wild turkeys in Arkansas. Journal of Wildlife Management 63:1281-1290. Turchin, P. 2001. Does population ecology have general laws? Oikos 94:17-26. Vander Haegen, W.M., W.E. Dodge, and M.W. Sayre. 1988. Factors affecting productivity in a northern wild turkey population. Journal of Wildlife Management 52:127-133. Vangilder, L.D. 1995. Survival and cause-specific mortality of wild turkeys in the Missouri Ozarks. Proceedings of the National Wild Turkey Symposium 7:21-31. Vangilder, L.D., E.W. Kurzejeski, V.L. Kimmel-truitt, and J.B. Lewis. 1987. Reproductive parameters of wild turkey hens in north Missouri. Journal of Wildlife Management 51:535-540. Vangilder, L.D., and E.W. Kurzejeski. 1995. Population ecology of the eastern wild turkey in northern Missouri. Wildlife Monographs 130:3-50. Warnke, D.K., and R.E. Rolley. 2005. Lessons from wild turkey population dynamics studies. Proceedings of the National Wild Turkey Symposium 9:91-99. Weinrich, J., E.E. Langenau, JR., and T. Reis. 1985. Relationship between winter census and spring harvest of wild turkeys in northern lower Michigan. Proceedings of the National Wild Turkey Symposium 5:295-301. Weinstein, M., B.D. Leopold, and G.A. Hurst. 1995. Evaluation of wild turkey population estimation methods. . Proceedings of the Southeastern Association of Fish and Wildlife Agencies 49:476-487. White, G.A., and K.P. Burnham. 2005. Program MARK: survival estimation from populations of marked animals. Bird Study 46: Supplement 120-138. Williams, C.K., A.R. Ives, and R.D. Applegate. 2003. Population dynamics across geographical ranges: time-series analyses of three small game species. Ecology 84:2654-2667. Wilson, W.B., M.J. Chamberlain, and F.G. Kimmel. 2005. Survival and nest success of female wild turkeys in a Louisiana bottomland hardwood forest. Proceedings of the Southeastern Association of Fish and Wildlife Agencies 59:126-134. Wood, S.N. 2006. Generalized additive models: an introduction with R. Chapman and Hall/CRC. Wood, S.N. 2013. Mgcv: mixed GAM computation vehicle with GCV/AIC/REML smoothness estimation. http://CRAN.R-project.org/package=mgcv. Wright, R.G., R.N. Paisley, and J.F. Kubisiak. 1996. Survival of wild turkey hens in southwest Wisconsin. Journal of Wildlife Management 60:313-320. Wunz, G.A., and W.K. Shope. 1980. Turkey brood survey in Pennsylvania as it relates to harvest. Proceedings of the National Wild Turkey Symposium 4:69-75.