Survival and Cause-Specific Mortality of Mature Male White-Tailed Deer

ABSTRACT Understanding sources of male deer mortality is a prerequisite to a successful management program, especially in Texas, USA, where white-tailed deer (Odocoileus virginianus) are the most economically important game species. South Texas, USA, is one of the few areas where males reach older age classes (> 4.5 yr), in part because of intense population management. Therefore, we obtained survival rates and causes of mortality of 48 mature male deer in south Texas, USA, over 2 years. We calculated Kaplan—Meier survival estimates during 2 study years modified for a staggered-entry design and annual survival rates for one cohort of deer from 1998 to 2004 using recapture and radiotelemetry data. We documented 21 mortalities (16 harvest and 5 nonhunting mortalities). Average annual survival of the known-aged 1998 cohort was 82% with 52% of surviving to 6.5 years of age. Survival in study year 2 (0.497 ± 0.069) was less than in study year 1 (0.781 ± 0.073; P = 0.0047), largely because males had finally reached harvestable age (> 6.5 yr old). All but one non-harvest mortality occurred during the rut or postrut periods. It appears that a large percentage of males can reach mature age classes under intense population management, making them available for harvest when at peak antler size. This allows for increased economic returns on intensively managed white-tailed deer populations.     

Incorporating harvest rates into the sex-age-kill model for white-tailed deer

Although monitoring population trends is an essential component of game species management, wildlife managers rarely have complete counts of abundance. Often, they rely on population models to monitor population trends. As imperfect representations of real-world populations, models must be rigorously evaluated to be applied appropriately. Previous research has evaluated population models for white-tailed deer (Odocoileus virginianus); however, the precision and reliability of these models when tested against empirical measures of variability and bias largely is untested. We were able to statistically evaluate the Pennsylvania sex-age-kill (PASAK) population model using realistic error measured using data from 1,131 radiocollared white-tailed deer in Pennsylvania from 2002 to 2008. We used these data and harvest data (number killed, age-sex structure, etc.) to estimate precision of abundance estimates, identify the most efficient harvest data collection with respect to precision of parameter estimates, and evaluate PASAK model robustness to violation of assumptions. Median coefficient of variation (CV) estimates by Wildlife Management Unit, 13.2% in the most recent year, were slightly above benchmarks recommended for managing game species populations. Doubling reporting rates by hunters or doubling the number of deer checked by personnel in the field reduced median CVs to recommended levels. The PASAK model was robust to errors in estimates for adult male harvest rates but was sensitive to errors in subadult male harvest rates, especially in populations with lower harvest rates. In particular, an error in subadult (1.5-yr-old) male harvest rates resulted in the opposite error in subadult male, adult female, and juvenile population estimates. Also, evidence of a greater harvest probability for subadult female deer when compared with adult (≥2.5-yr-old) female deer resulted in a 9.5% underestimate of the population using the PASAK model. Because obtaining appropriate sample sizes, by management unit, to estimate harvest rate parameters each year may be too expensive, assumptions of constant annual harvest rates may be necessary. However, if changes in harvest regulations or hunter behavior influence subadult male harvest rates, the PASAK model could provide an unreliable index to population changes. © 2012 The Wildlife Society.     

Survival,cause-specific mortality,and population growth of white-tailed deer in western Virginia

Understanding the role of recruitment in population dynamics of white-tailed deer (Odocoileus virginianus) is important for management. In the central Appalachian Mountains, deer are part of a largely forested ecosystem that supports 3 carnivore species thought to be capable of influencing white-tailed deer recruitment: black bears (Urus americanus), coyotes (Canis latrans), and bobcats (Lynx rufus). Yet little is known about predation, how other environmental factors influence recruitment, or the importance of neonate survival to white-tailed deer population performance in the region. Our objectives were to identify causes of mortality for neonates, analyze effects of landscape attributes on survival of neonates, estimate survival rates for neonates and adult female white-tailed deer, and to model population growth trends based on current vital rates and hypothetical harvest and neonate survival scenarios. During 2019–2020, we captured 57 neonate deer in Bath County, Virginia, USA, by monitoring 38 pregnant females equipped with global positioning system collars and vaginal implant transmitters and by conducting transect searches for recently born neonates. We observed 37 neonate mortalities and identified cause of death using field and genetic evidence. Mortalities included 28 predation events and 9 deaths from other causes (e.g., abandonment, malnutrition, disease). Black bears accounted for 48.6% of neonate mortalities, and 64.2% of predation events (n = 18), followed by bobcats (n = 5) and coyotes (n = 3). Annual survival for adult female deer was 0.871 and neonate survival to 12 weeks old was 0.310. Elevation was a significant predictor of neonate survival; mortality risk increased 20% for every 100-m increase in elevation. Models of annual population growth using observed vital rates predicted an increasing population (λ = 1.10). A 10% increase in female harvest would still result in a potential population increase of 2% (λ = 1.02), but a 20% increase in harvest rate would result in a potential 7% decline (λ = 0.93). Neonate survival was higher near fertile valley bottoms and lower along forested ridges characterized by shallow, infertile soils and limited edge or early successional forests. While predation, largely influenced by black bears, was the leading cause of neonate mortality and contributed to low neonate survival, we observed little evidence of population decline, and suggest there is opportunity for a modest increase in harvest of female deer.     

White-Tailed Deer Population Dynamics Following Louisiana Black Bear Recovery

Changing predator communities have been implicated in reduced survival of white-tailed deer (Odocoileus virginianus) fawns. Few studies, however, have used field-based age-specific estimates for survival and fecundity to assess the relative importance of low fawn survival on population growth and harvest potential. We studied white-tailed deer population dynamics on Tensas River National Wildlife Refuge (TRNWR) in Louisiana, USA, where the predator community included bobcats (Lynx rufus), coyotes (Canis latrans), and a restored population of Louisiana black bear (Ursus americanus luteolus). During 2013–2015, we radio-collared and monitored 70 adult (≥2.5 yrs) and 21 yearling (1.5-yr-old) female deer. Annual survival averaged 0.815 (95% CI = 0.734–0.904) for adults and 0.857 (95% CI = 0.720–1.00) for yearlings. We combined these estimates with concurrently collected fawn survival estimates (0.27; 95% CI = 0.185–0.398) to model population trajectories and elasticities. We used estimates of nonhunting survival (annual survival estimated excluding harvest mortality) to project population growth (λ) relative to 4 levels of harvest (0, 10%, 20%, 30%). Finally, we investigated effects of reduced fawn survival on population growth under current management and with elimination of female harvest. Despite substantial fawn predation, the deer population on TRNWR was increasing (λ = 1.06) and could sustain additional female harvest; however, the population was expected to decline at 20% (λ = 0.98) and 30% (λ = 0.94) female harvest. With no female harvest, the population was projected to increase with observed (λ = 1.15) and reduced fawn survival (λ = 1.02), but the population could not sustain current female harvest (10%) if fawn survival declined (λ = 0.90). For all scenarios, adult female survival was the most elastic parameter. Given the importance of adult female survival, the relative predictability in response of adult survival to harvest management, and the difficulty in altering fawn survival, reducing female harvest is likely the most efficient approach to compensate for low fawn survival. On highly productive sites such as ours, reduction, but not necessarily elimination, of harvest can mitigate effects of low fawn survival on population growth. © 2020 The Wildlife Society.     

Effects of Maternal Status on the Movement and Mortality of Sterilized Female White-Tailed Deer

Abstract: Surgical sterilization by tubal ligation has been proposed as a technique for controlling white-tailed deer (Odocoileus virginianus) populations in urban or suburban areas where other forms of population control are impractical, but little is known about demographic rates in populations under management with surgical sterilization. We analyzed seasonal movement and mortality data collected during a 4-year study of surgical sterilization in suburban Chicago, Illinois, USA. We calculated 323 home range size estimates for 62 individual females within season and year. Non-gravid females without young exhibited home range sizes 52% larger than gravid females and females with fawns. Mortality rate was positively correlated with home range size. We suggest that the increased mortality rate observed in surgically sterilized females may be due to greater movement by non-maternal females. Population managers will need to account for potential effects of maternal status on movement and mortality when considering the use of sterilization for management of suburban populations of white-tailed deer.     

Evaluating Differences in Harvest Data Used in the Sex-Age-Kill Deer Population Model

Abstract: A steady increase in archery hunting participation and frequent changes in hunter regulations led to an evaluation of harvest data used in a common white-tailed deer (Odocoileus virginianus) population model. Our goal was to determine if model parameters and population estimates traditionally estimated solely by firearm harvest data were biased with respect to altered sex and age ratios brought about by increases in archery hunting and harvest success. The sex-age-kill (SAK) model, commonly used by state agencies, was developed in the mid-1900s when deer numbers were low and firearm harvest was predominant. Management actions were concentrated on increasing deer numbers, and model assumptions relied heavily on a stable age distribution and a minimal antlerless deer harvest. We evaluated the reliance of SAK in a modern hunting scenario using a 10-year dataset obtained from Michigan, USA, that encompassed a variety of climatic regions, hunting seasons, and regulation scenarios. We found that firearm and archery harvest sex and age ratios differed among 5 geographic groups and study years for males, females, and fawns (P<0.001, P = 0.001, and P = 0.037, respectively). Also, the addition of archery harvest data increased population estimates but did not alter overall trends. We recommend that managers reassess harvest-based population estimates in 2 situations: 1) if regulation changes affect antlerless deer harvest, and 2) when trends in hunter success rates cause fluctuations in harvest data.     

Influence of Landscape Composition and Structure on Antler Size of White-Tailed Deer

Abstract: Cause for spatial variation in phenotypic quality of white-tailed deer (Odocoileus virginianus) populations is of great interest to wildlife managers. Relating phenotypic variation of populations to large-scale land-use patterns may provide insight into why populations exhibit spatial variation and elucidate how management can influence population phenotype. We used an information-theoretic approach to relate average antler size of 203 deer populations to composition and structure of the habitat occupied by those populations. We used interspersion, edge, and diversity indices to represent habitat structure and percentage of area in vegetation types to represent habitat composition. Landscape composition was a better predictor of deer population antler size than was landscape structure. Percentages of the management unit in agriculture, pasture, and pine forest were variables commonly found in the region-specific set of best models. Model-averaged estimates of agriculture and pasture parameters were always positive and estimates of pine forest parameters were always negative, which suggests that land-use types that promote growth of early successional herbaceous plants will positively influence antler size and, most likely, body growth and reproduction of white-tailed deer populations. Conversely, our findings suggest landscapes dominated by pine forests did not provide optimal amounts of quality forages for white-tailed deer. Pine forest effects should be mitigated using a combination of increased harvest to lower deer density and silvicultural practices like thinning, prescribed burning, and selective herbicide applications that stimulate growth of high-quality forages beneath the forest canopy to improve deer phenotypic quality.     

Integrated Population Modeling of Black Bears in Minnesota: Implications for Monitoring and Management

Background

Wildlife populations are difficult to monitor directly because of costs and logistical challenges associated with collecting informative abundance data from live animals. By contrast, data on harvested individuals (e.g., age and sex) are often readily available. Increasingly, integrated population models are used for natural resource management because they synthesize various relevant data into a single analysis.

Methodology/Principal Findings

We investigated the performance of integrated population models applied to black bears (Ursus americanus) in Minnesota, USA. Models were constructed using sex-specific age-at-harvest matrices (1980–2008), data on hunting effort and natural food supplies (which affects hunting success), and statewide mark–recapture estimates of abundance (1991, 1997, 2002). We compared this approach to Downing reconstruction, a commonly used population monitoring method that utilizes only age-at-harvest data. We first conducted a large-scale simulation study, in which our integrated models provided more accurate estimates of population trends than did Downing reconstruction. Estimates of trends were robust to various forms of model misspecification, including incorrectly specified cub and yearling survival parameters, age-related reporting biases in harvest data, and unmodeled temporal variability in survival and harvest rates. When applied to actual data on Minnesota black bears, the model predicted that harvest rates were negatively correlated with food availability and positively correlated with hunting effort, consistent with independent telemetry data. With no direct data on fertility, the model also correctly predicted 2-point cycles in cub production. Model-derived estimates of abundance for the most recent years provided a reasonable match to an empirical population estimate obtained after modeling efforts were completed.

Conclusions/Significance

Integrated population modeling provided a reasonable framework for synthesizing age-at-harvest data, periodic large-scale abundance estimates, and measured covariates thought to affect harvest rates of black bears in Minnesota. Collection and analysis of these data appear to form the basis of a robust and viable population monitoring program.     

Predator densities and white-tailed deer fawn survival

Predation is the dominant source of mortality for white-tailed deer (Odocoileus virginianus) <6 months old throughout North America. Yet, few white-tailed deer fawn survival studies have occurred in areas with 4 predator species or have considered concurrent densities of deer and predator species. We monitored survival and cause-specific mortality from birth to 6 months for 100 neonatal fawns during 2013–2015 in the Upper Peninsula of Michigan, USA, while simultaneously estimating population densities of deer, American black bear (Ursus americanus), coyote (Canis latrans), bobcat (Lynx rufus), and gray wolf (Canis lupus). We estimated fawn predation risk in response to sex, birth mass, and date of birth. Six-month fawn survival pooled among years was 36%, and fawn mortality risk was not related to birth mass, date of birth, or sex. Estimated mean annual deer and predator densities were 334 fawns/100 km², 25.9 black bear/100 km², 23.8 coyotes/100 km², 3.8 bobcat/100 km², and 2.8 wolves/100 km². Despite lower estimated per-individual kill rates, coyotes and black bears were the leading sources of fawn mortality because they had greater densities relative to bobcats and wolves. Our results indicate that the presence of more predator species in a system is not entirely additive in its effect on fawn survival. © The Wildlife Society, 2019     

Mortality of Rocky Mountain elk in Michigan due to meningeal worm

Mortality from cerebrospinal parelaphostrongylosis caused by the meningeal worm (Parelaphostrongylus tenuis) has been hypothesized to limit elk (Cervus elaphus nelsoni) populations in areas where elk are conspecific with white-tailed deer (Odocoileus virginianus). Elk were reintroduced into Michigan (USA) in the early 1900s and subsequently greatly increased population size and distribution despite sympatric high-density (>or=12/km2) white-tailed deer populations. We monitored 100 radio-collared elk of all age and sex classes from 1981-94, during which time we documented 76 mortalities. Meningeal worm was a minor mortality factor for elk in Michigan and accounted for only 3% of mortalities, fewer than legal harvest (58%), illegal kills (22%), other diseases (7%), and malnutrition (4%). Across years, annual cause-specific mortality rates due to cerebrospinal parelaphostrongylosis were 0.033 (SE=0.006), 0.029 (SE=0.005), 0.000 (SE=0.000), and 0.000 (SE=0.000) for calves, 1-yr-old, 2-yr-old, and >or=3-yr-old, respectively. The overall population-level mortality rate due to cerebrospinal parelaphostrongylosis was 0.009 (SE=0.001). Thus, meningeal worm had little impact on elk in Michigan during our study despite greater than normal precipitation (favoring gastropods) and record (>or=14 km2) deer densities. Further, elk in Michigan have shown sustained population rates-of-increase of >or=18%/yr and among the highest levels of juvenile production and survival recorded for elk in North America, indicating that elk can persist in areas with meningeal worm at high levels of population productivity. It is likely that local ecologic characteristics among elk, white-tailed deer, and gastropods, and degree of exposure, age of elk, individual and population experience with meningeal worm, overall population vigor, and moisture determine the effects of meningeal worm on elk populations.     

Space Use and Survival of White-Tailed Deer in an Exurban Landscape

Abstract: Exurban development is nonmetropolitan, residential development characterized by a human population density and average property size intermediate between suburban and rural areas. Although growth in exurban areas is outpacing that of urban, suburban, or rural landscapes, studies of deer (Odocoileus spp.) ecology in exurban areas are nonexistent. During 2003–2005, we studied space use (i.e., seasonal home-range and core-area size and habitat use relative to human dwellings) and survival of 43 female white-tailed deer (O. virginianus) in an exurban setting near Carbondale, Illinois. Deer had larger home ranges than most suburban deer populations and generally smaller home ranges than rural deer populations. When we analytically controlled for habitat use, deer exhibited a subtle avoidance of human dwellings, especially during the fawning season. The annual survival rate was among the highest reported in the literature at 0.872 (SE = 0.048). Only 5 deer (cause-specific mortality rate = 0.091) were harvested by hunters, indicating major obstacles for wildlife managers when attempting to manage deer in exurban areas using traditional hunter harvest.     

Interannual Variability in Survival Rates for Adult Female White-Tailed Deer

Adult female survival is an important component to population models and management programs for white-tailed deer (Odocoileus virginianus), but short-term survival studies (1–3 yrs) may not accurately reflect the variation in interannual survival, which could alter management decisions. We monitored annual survival and cause-specific mortality rates of adult female white-tailed deer (n = 158) for 6 years (2010–2012, 2016–2018) in southern Delaware, USA. Annual survival rate differed among years. Survival rates (±SE) and mortality causes were similar in 3 years (2011 = 0.72 ± 0.08, 2017 = 0.68 ± 0.08, 2018 = 0.74 ± 0.09) and comparable to previous research from mixed forest-agricultural landscapes. A relatively low survival rate in 2010 (0.48 ± 0.11) was influenced by hunter harvest and potentially compounded by abnormally severe winter conditions in the prior year. A peracute outbreak of hemorrhagic disease occurred during summer 2012, resulting in an annual survival rate of 0.38 ± 0.11, and to our knowledge is the first reported case of a hemorrhagic disease outbreak in a monitored wild population with known fates. In 2016, we did not observe any harvest mortality, resulting in high annual survival (0.96 ± 0.04). Our results demonstrate the degree of variability in annual survival and cause-specific mortality rates within a population. We caution against the use of short-term survival studies to inform management decisions, particularly when incorporating survival data into population models or when setting harvest objectives. © 2020 The Wildlife Society.     

Spotlight surveys for white-tailed deer: Monitoring panacea or exercise in futility?

Many monitoring programs for white-tailed deer (Odocoileus virginianus) on both private and public lands across the United States have long relied on the use of road-based spotlight surveys for monitoring population size and trends. Research has suggested spotlight surveys are ineffective and that road-based surveys for deer are biased because of highly variable detection rates. To evaluate variability in detection rates relative to the assumption that repeated surveys along roads will provide reliable trend data for use in calculating deer density estimates, we collected 5 years of thermal-imager and spotlight survey data using a multiple-observer, closed-capture approach. Using a Huggin's closed capture model, data bootstrapping, and variance components analyses, our results suggest that density estimates for white-tailed deer generated from data collected during road-based spotlight surveys are likely not reflective of the standing deer population. Detection probabilities during individual spotlight surveys ranged from 0.00 to 0.80 (median = 0.45) across all surveys, and differed by observer, survey, management unit, and survey transect replicate. Mean spotlight detection probability (0.41) and process standard deviation (0.12) estimates indicated considerable variability across surveys, observers, transects, and years, which precludes the generation of a correction factor or use of spotlight data to evaluate long-term trends at any scale. Although recommended by many state, federal, and non-governmental agencies, our results suggest that the benefit of spotlight survey data for monitoring deer populations is limited and likely represents a waste of resources with no appreciable management information gained. © 2012 The Wildlife Society.     

Survival of white-tailed deer fawns on Marine Corps Base Quantico

Some jurisdictions in the eastern United States have reduced harvest of white-tailed deer (Odocoileus virginianus) because of perceived declines in recruitment and population size over the last decade. Although the restoration of American black bears (Ursus americanus) and the colonization of coyotes (Canis latrans) have increased fawn predation in some areas, limited information exists on how temporally dynamic resources and weather influence fawn survival. Therefore, we evaluated fawn survival probability, cause specific mortality, and if factors such as oak (Quercus spp.) mast abundance, winter severity, precipitation, and landscape composition influenced mortality risk on Marine Corps Base Quantico in northern Virginia, USA, from 2008 to 2019. We tracked 248 fawns outfitted with very high frequency radio-collars and predation was the leading cause of mortality (n = 42; 45%). We estimated survival to 133 days and survival pooling all years (2008–2019) was 0.50 (95% CI = 0.42–0.60). Increased annual red oak (Quercus spp.) mast abundance from the previous fall reduced mortality hazard for fawns. The longevity of our study revealed a link between fawn survival and a specific maternal resource (red oak mast) only available during gestation. Our results highlight the importance of oak mast in eastern deciduous forests and, more broadly, overwinter maternal condition on white-tailed deer recruitment.     

Spatio-temporal variation in male white-tailed deer harvest rates in Pennsylvania: Implications for estimating abundance

The performance of 2 popular methods that use age-at-harvest data to estimate abundance of white-tailed deer is contingent on assumptions about variation in estimates of subadult (1.5 yr old) and adult (≥2.5 yr old) male harvest rates. Auxiliary data (e.g., estimates of survival or harvest rates from radiocollared animals) can be used to relax some assumptions, but unless these population parameters exhibit limited temporal or spatial variation, these auxiliary data may not improve accuracy. Unfortunately maintaining sufficient sample sizes of radiocollared deer for parameter estimation in every wildlife management unit (WMU) is not feasible for most state agencies. We monitored the fates of 397 subadult and 225 adult male white-tailed deer across 4 WMUs from 2002 to 2008 using radio telemetry. We investigated spatial and temporal variation in harvest rates and investigated covariates related to the patterns observed. We found that most variation in harvest rates was explained spatially and that adult harvest rates (0.36–0.69) were more variable among study areas than subadult harvest rates (0.26–0.42). We found that hunter effort during the archery and firearms season best explained variation in harvest rates of adult males among WMUs, whereas hunter effort during only the firearms season best explained harvest rates for subadult males. From a population estimation perspective, it is advantageous that most variation was spatial and explained by a readily obtained covariate (hunter effort). However, harvest rates may vary if hunting regulations or hunter behavior change, requiring additional field studies to obtain accurate estimates of harvest rates. © 2011 The Wildlife Society.     

Reduction and Maintenance of a White-Tailed Deer Herd in Central Massachusetts

Abstract: Controlled public hunts have been used in a variety of settings to reduce overabundant white-tailed deer (Odocoileus virginianus) herds. We present the results of a large-scale (160 km²) controlled hunt at Quabbin Reservation (QR) in central Massachusetts, USA. The QR was divided into 5 hunt zones. Hunting was initiated in each zone from 1991 to 1994 and continued through 2004. The management goal was to achieve posthunt deer densities of4 deer/km². Initial estimated deer densities in each zone ranged from 11.4 deer/km² to 27.6 deer/km². The management goals were reached in each zone after 2-4 years of hunting. Posthunt populations were maintained at or below the goal even though total hunter effort was reduced. Hunters were not required to harvest antlerless deer, but antlerless deer comprised 55-83% of the harvest each year. We simulated the effects of 5 years without hunting on deer populations. The simulated deer population exceeded management goals after 2 years. Our results demonstrate that controlled public deer hunts can effectively reduce deer populations and maintain them at desired levels over large areas with minimal hunter restrictions. Managers should prepare stakeholders during the hunt planning process for the need to continue overall harvest rates of >30% during the maintenance phase of a deer management program.     

Predation by coyotes on white-tailed deer neonates in South Carolina

Coyotes (Canis latrans) are novel predators throughout the southeastern United States and their depredation of white-tailed deer (Odocoileus virginianus) neonates may explain observed declines in some deer populations in the region, but direct evidence for such a relationship is lacking. Our objective was to quantify neonate survival rates and causes of mortality at the United States Department of Energy's Savannah River Site (SRS), South Carolina to directly evaluate degree of predation in this deer population. From 2006 to 2009, we radio-monitored 91 neonates captured with the aid of vaginal implant transmitters in pregnant adult females and opportunistic searches. Overall Kaplan–Meier survival rate to 16 weeks of age was 0.230 (95% CI = 0.155–0.328), and it varied little among years. Our best-fitting model estimated survival at 0.220 (95% CI = 0.144–0.320). This model included a quadratic time trend variable (lowest survival rate during the first week of life and increasing to near 1.000 around week 10), and Julian date of birth (survival probability declining as date of birth increased). Predation by coyotes was the most frequent cause of death among the 70 monitored neonates that died, definitively accounting for 37% of all mortalities and potentially accounting for as much as 80% when also including probable coyote predation. Predation by bobcats (Felis rufus) accounted for 7% (definitive) to 9% (including probable bobcat predation) of mortalities. The level of coyote-induced mortality we observed is consistent with the low recruitment rates exhibited in the SRS deer population since establishment of coyotes at the site. If representative of recruitment rates across South Carolina, current harvest levels appear unsustainable. This understanding is consistent with the recent declining trend in the statewide deer population. The effects of coyote predation on recruitment should be considered when setting harvest goals, regardless of whether local deer population size is currently above or below desired levels, because coyotes can substantially reduce fawn recruitment. Published 2012. This article is a U.S. Government work and is in the public domain in the USA.     

An Evaluation of Sex-Age-Kill (SAK) Model Performance

ABSTRACT The sex-age-kill (SAK) model is widely used to estimate abundance of harvested large mammals, including white-tailed deer (Odocoileus virginianus). Despite a long history of use, few formal evaluations of SAK performance exist. We investigated how violations of the stable age distribution and stationary population assumption, changes to male or female harvest, stochastic effects (i.e., random fluctuations in recruitment and survival), and sampling efforts influenced SAK estimation. When the simulated population had a stable age distribution and λ > 1, the SAK model underestimated abundance. Conversely, when λ < 1, the SAK overestimated abundance. When changes to male harvest were introduced, SAK estimates were opposite the true population trend. In contrast, SAK estimates were robust to changes in female harvest rates. Stochastic effects caused SAK estimates to fluctuate about their equilibrium abundance, but the effect dampened as the size of the surveyed population increased. When we considered both stochastic effects and sampling error at a deer management unit scale the resultant abundance estimates were within ±121.9% of the true population level 95% of the time. These combined results demonstrate extreme sensitivity to model violations and scale of analysis. Without changes to model formulation, the SAK model will be biased when λ ≠ 1. Furthermore, any factor that alters the male harvest rate, such as changes to regulations or changes in hunter attitudes, will bias population estimates. Sex-age-kill estimates may be precise at large spatial scales, such as the state level, but less so at the individual management unit level. Alternative models, such as statistical age-at-harvest models, which require similar data types, might allow for more robust, broad-scale demographic assessments.     

Methodology matters when estimating deer abundance: a global systematic review and recommendations for improvements

Deer (Cervidae) are key components of many ecosystems and estimating deer abundance or density is important to understanding these roles. Many field methods have been used to estimate deer abundance and density, but the factors determining where, when, and why a method was used, and its usefulness, have not been investigated. We systematically reviewed journal articles published during 2004–2018 to evaluate spatio-temporal trends in study objectives, methodologies, and deer abundance and density estimates, and determine how they varied with biophysical and anthropogenic attributes. We also reviewed the precision and bias of deer abundance estimation methods. We found 3,870 deer abundance and density estimates. Most estimates (58%) were for white-tailed deer (Odocoileus virginianus), red deer (Cervus elaphus), and roe deer (Capreolus capreolus). The 6 key methods used to estimate abundance and density were pedestrian sign (track or fecal) counts, pedestrian direct counts, vehicular direct counts, aerial direct counts, motion-sensitive cameras, and harvest data. There were regional differences in the use of these methods, but a general pattern was a temporal shift from using harvest data, pedestrian direct counts, and aerial direct counts to using pedestrian sign counts and motion-sensitive cameras. Only 32% of estimates were accompanied by a measure of precision. The most precise estimates were from vehicular spotlight counts and from capture–recapture analysis of images from motion-sensitive cameras. For aerial direct counts, capture–recapture methods provided the most precise estimates. Bias was robustly assessed in only 16 studies. Most abundance estimates were negatively biased, but capture–recapture methods were the least biased. The usefulness of deer abundance and density estimates would be substantially improved by 1) reporting key methodological details, 2) robustly assessing bias, 3) reporting the precision of estimates, 4) using methods that increase and estimate detection probability, and 5) staying up to date on new methods. The automation of image analysis using machine learning should increase the accuracy and precision of abundance estimates from direct aerial counts (visible and thermal infrared, including from unmanned aerial vehicles [drones]) and motion-sensitive cameras, and substantially reduce the time and cost burdens of manual image analysis.