Animals migrate to access seasonally variable resources or to escape unfavorable local conditions. Partial (flexible) migration strategies offer animals a fitness trade-off in which individuals may choose whether to incur the energetic costs of migration depending on the potential benefits associated with avoiding adverse conditions. Partial migration patterns may shift because of anthropogenic land use and climate change, possibly reducing the benefits of migration and resulting in unpredictable migration cues. We examined causes, consequences, and changes in individual-level partial migration of white-tailed deer Odocoileus virginianus by analyzing archival telemetry data (1986–2014) across four distinct populations in the Great Lakes region, North America. We hypothesized that if the costs of migration outweigh benefits, we would see migratory behavior decline. Migratory behavior declined from 75% in 1987 to 11% in 2014 independent of population identity. Annual mortality was higher for migratory white-tailed deer, whereas increasing minimum temperatures positively impacted survival but did not influence migration. Our results suggest that migratory behavior in white-tailed deer is declining over time and may be associated with declining winter severity. The loss of migratory behavior in large mammals may have ecosystem-level impacts and our study emphasizes the need to better understand and conserve these migratory traditions.

Partial migration is a potential adaptation strategy various species use to flexibly adapt migratory behavior based on environmental conditions, reducing unfavorable fitness trade-offs (Alerstam et al. 2003; Grayson and Wilbur 2009). Individuals in partially migratory populations may switch migration tactics throughout their lifetime, choosing to migrate some years and remaining resident during other years depending on environmental cues (Spitz et al. 2018; Fagan and Gurarie 2020), and these individuals are termed conditional migrators (Sabine et al. 2002; Fieberg et al. 2008). An individual’s decision to migrate is tied to winter severity interacting with habitat for white-tailed deer Odocoileus virginianus (Nelson 1998; Van Deelen et al. 1998; Sabine et al. 2002; Fieberg et al. 2008), pronghorn Antilocapra americana (White et al. 2007), and roe deer Capreolus capreolus (Cagnacci et al. 2011). In fact, partial migration may be more common than obligate migration for many migratory species (Fryxell and Holt 2013; Peters et al. 2017). Migration can increase access to forage quality, allow escape from predation risk, or both, which can increase survival and result in higher populations (Fryxell and Sinclair 1988; Courtemanch et al. 2017). Partial migration has emerged as an important topic for understanding animal behavior, but there are still unanswered questions about drivers and survival trade-offs between migratory and resident individuals.

Migration patterns may change because of anthropogenic impacts, including habitat alteration (e.g., fencing, development) and climate change, which disrupt corridors and increase habitat loss (Bolger et al. 2008; Bischof et al. 2012; Cole et al. 2015), especially in partially migratory species with flexible behavior. Climate change may impact migration timing and breeding in response to increases in spring air temperatures (Walther et al. 2002). Warming temperatures may also lead to increase in ice-free days, which can impact migration pathways of animals such as caribou Rangifer tarandus that migrate across frozen water bodies (Sharma et al. 2009). Migrations vary in scale, but movement of these large ungulates across landscapes can impact ecosystems by creating spatial and temporal heterogeneity in herbivory and predation (McNaughton 1976; Hebblewhite and Merrill 2009; Mysterud et al. 2011). Despite variation, several motivations commonly drive migratory behavior (Harris et al. 2009). These include four main ecological drivers: 1) seasonal availability of forage, 2) snow depth (specifically N. American and Eurasian migrants), 3) use of traditional areas (i.e., for reproduction), and 4) surface water availability (Harris et al. 2009). Migration can increase exposure to predation risk and has energetic demands that result in a trade-off between the costs and benefits of migration (Nelson and Mech 1991). Due to potential fitness trade-offs, partial migration may be an optimal behavioral strategy.

While environmental changes may lead to the loss or reduction of conditional migratory behavior, questions remain about how long it will take for these changes to occur and what the consequences might be (Shaw 2016). White-tailed deer experience a unique form of partial migration, as they do not move in massive groups across the landscape and the distances between ranges varies, ∼5–40km (Halpin et al. 1987; Nelson 1995). Motivations for partial migrations in white-tailed deer are poorly understood, but proximal environmental cues may play an important role (Van Deelen et al. 1998; Sabine et al. 2002) as well as traditions passed from doe to fawn (Cook and Hamilton 1942; Nelson 1995; Brinkman et al. 2005). Increased winter severity increases migration in conditionally migrating white-tailed deer (Sabine et al. 2002), suggesting that individuals shift from migratory to resident as winter severity declines. Equilibrium theory of migration suggests that conditional migrators exist in a partially migratory population because there are fitness benefits associated with migrating that offset the survival and potential reproduction costs of migration (Fryxell and Holt 2013). As environmental conditions change and winters become less severe individual survival benefits of white-tail deer migration may decline, leading to a population-level shift from partially migratory to resident. To predict the fate of migration with changing environmental conditions, we therefore need to understand the proximal and ultimate drivers of migratory behavior (Shaw 2016).

Migratory white-tailed deer in the midwestern USA offer a case study to explore the environmental drivers of partial migration because researchers have been studying their behavior for over 35 years. If migration is becoming increasingly disassociated with survival probability, then we would expect to see a decline in this behavior either because 1) the proximal environmental cues that initiate migratory behavior are reduced, or 2) the fitness benefits of migration decline. To understand the fitness consequences of migration, we were interested in detecting differences in mean mortality as well as the timing of peak hazard between migratory and resident individuals. White-tailed deer use migration as behavioral tactic to survive severe winters; therefore, we hypothesize that survival costs increase for residential individuals only during severe winters.

To investigate the role of decreasing winter severity on the occurrence of short-distance migrations and on seasonal survival patterns, we analyzed archival telemetry data from four studies spanning the years 1986–2014 in three locations across the Great Lakes region. Each study spanned three to four years, with all study locations occurring at similar latitudes and the oldest and most recent studies occurring at the same location, aiding insight into how migration may be changing at a population level over time and the impacts of these changes on survival. Our analyses are designed to isolate temporal trends in migratory behavior and survival while accounting for inconsistency in geographic location. By quantifying and understanding white-tailed migration in the context of a changing climate, we can gain insight into the complex interactions of changing climate and seasonal movement behaviors.

Study site

Our study areas included three locations at similar latitudes in northern Wisconsin and Michigan (USA, Figure 1). The northern Wisconsin study site (hereafter NWI), centers on the Chequamegon National Forest and includes Sawyer and Rusk counties. Forests at this site are composed of mixed hardwoods and lowlands with mixed conifer swamps (Lewis 1990). Species associations include tamarack Larix laricina-black spruce Picea mariana bogs, northern white cedar Thuja occidentalis and alder Alnus glutinosa swamps (Norton 2015). Predator loads have varied overtime in northern Wisconsin with limited wolf abundance in the 1980s and rapid increases in wolf abundance through the 1990s resulting in stable wolf populations, ∼972 individuals in 2022 (Wisconsin Department of Natural Resources 2022a). Due to minimal presence wolves did not impact mortality during our first study (Lewis 1990) and were not found to have strong effects on mortality in the most recent study (Norton et al. 2021). Despite variation in wolf abundance, Wisconsin supports a large white-tailed deer population with annual harvests exceeding 200,000 white-tailed deer since the early 1980s and a stable population trends with slight increases over time (Wisconsin Department of Natural Resources 2022b).

Figure 1.

Locations of the four white-tailed deer Odocoileus virginianus telemetry studies in Wisconsin and Michigan (U.S.A.). Data are from the Great Lakes region of North America: Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA). Capture and mortality dates, telemetry locations, and migration status for adult females were taken from these four studies.

Figure 1.

Locations of the four white-tailed deer Odocoileus virginianus telemetry studies in Wisconsin and Michigan (U.S.A.). Data are from the Great Lakes region of North America: Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA). Capture and mortality dates, telemetry locations, and migration status for adult females were taken from these four studies.

Close modal

The study site in Michigan’s lower peninsula includes Presque Isle, Montmorency, and Alpena counties. Most land is forested and composed of conifer species; northern white cedar, balsam fir Abies balsamea, white spruce Picea glauca, and jack pine Picea banksiana. Deciduous species including quaking aspen Populus tremuloides, maples, Acer spp. basswood Tilia Americana and oaks Quercus spp. are also present throughout the study area. Within the study area, large patches of agricultural land are surrounded by forest.

Lastly, the site in Michigan’s upper peninsula (hereafter UPMI) comes from the Stonington peninsula and the Whitefish River watershed. The study site is a series of wintering yards and thus contained many stands of cedar-dominated swamps. Cedar-dominated swamps or large dense conifer stands are termed ‘deer yards’, as they provide thermal cover and reduce energetic costs for white-tailed deer during severe winters (Ozoga 1968). Lake Superior lies to the north with Lake Michigan to the south. The area’s uplands are dominated by northern hardwoods with areas of pines. The land use is predominantly forest-related (forestry and recreation) with a few small farms in the Stonington peninsula. Historically, population trends of white-tailed deer in Michigan have varied, with the population increasing from the 70s to early 90s then experiencing a gradual decline since 1995. Populations in the southern third of the state are higher and experience greater harvest with annual harvests exceeding 150,000 since the late 1980s (Mason 2016). Similar to northern Wisconsin, wolves started recolonizing UPMI in the late 1980s resulting in current stable wolf populations (MI DNR). White-tailed deer populations in UPMI and northern Wisconsin were similar and experienced comparable winter conditions and predator loads throughout our study years (1986–2014).

White-tailed deer data

We compiled telemetry data on white-tailed deer from four previous studies covering years 1986–1989 (Lewis 1990), 1992-1995 (Van Deelen 1995), 1994-1996 (Sitar 1996), and 2011-2014 (Norton 2015). The earliest study (Lewis 1990) and most recent study (Norton 2015) were both conducted at site NWI. Common sampling methods included capture and radio-marking white-tailed deer with VHF collars during mid-late winter and early spring (December–April). White-tailed deer were located at least twice a week. Seasonal range boundaries were estimated using the 95% adaptive kernel contours which were converted to PC-ARC/INFO polygon coverages to examine the yearly and seasonal overlap of seasonal ranges (Van Deelen et al. 1998). If a white-tailed deer migrated at least once throughout the study and its summer and winter ranges did not overlap it was defined as migratory (Lewis 1990; Van Deelen 1995). “Conditional migrators,” white-tailed deer that migrate some years but not others, were defined as migratory in our study because of the low frequency of individuals that showed behavior which could be defined as conditional migration (<0.05%)(Van Deelen 1995). Male white-tailed deer have more variable migration patterns and seasonal movements can be confounded by expanded movement during rut so only female records were used in our analysis. Additionally, records of juveniles (≤1) were eliminated to avoid confounding instances of dispersal (Lutz et al. 2015). Total sample size consisted of 172 female white-tailed deer after excluding males and juveniles. Our dataset described migration patterns for 172 female white-tailed deer across 11 years (Table S1)cite S2?. The first data set spanned 1986–1989 (Lewis 1990, n = 38 white-tailed deer) in northern Wisconsin, followed by data from Michigan’s upper peninsula from 1992-1995 (Van Deelen 1995, n = 49), Michigan’s northern lower peninsula spanning 1994-1996 (Sitar 1996, n = 47), and lastly, northern Wisconsin covering 2011-2014 (Norton 2015, n = 38). There were 77 migratory females and 95 resident females.

Weather data

We retrieved weather data from Daymet (Thornton et al. 2014) in December 2019. Daymet provides daily weather data from January 1, 1980, until December 2018 (at the time it was retrieved) at a 1-km2 spatial resolution across North America. The raster package (Hijmans 2020) in R (R Core Team, 2020; all data compilation and analysis were conducted in R) was used to extract weather data corresponding to the geographic centroid of each study location during observation years, summarized (mean, min, max) over two-week intervals. Two-week time intervals were chosen to assess how weather conditions over short periods of time may impact white-tail deer movement and survival. Weather variables were also summarized (mean, min, max) for the fall migration season (September 1December 31) to assess population changes in migration. Weather variables included minimum daily air temperature (°C), total precipitation (mm/day), and snow water equivalent (kg/m2).

To determine changes in winter severity over time we used measurements of Accumulated Winter Seasonal Severity Index (AWSSI) at our study locations from the Midwestern Regional Climate Center. We used Total AWSSI and TempSc from the years 1985–2014. Total AWSSI includes TempSc + SnowSc for the season. SnowSc is calculated by summing points for snow depth and snow fall for the season. TempSc is calculated by summing points for minimum and maximum temperatures for the season. The winter season begins with the first measurable snowfall (>=1in) or maximum temperature at or below 32°F and the season ends when one of these three criteria are met: last measurable snowfall (>= 0.1 inch), last day with 1 inch of snow on the ground, last day with a maximum temperature of 32°F or lower, or February 28/29. We also summarized minimum temperatures for each month of the winter season (September–February) to determine if there were any warming trends during our study years 1985–2014).

Migration changes

We were interested in how environmental change during fall migration affected individual migratory responses. We used generalized mixed-effects linear regression to model the impacts of seasonal weather on migration and changes in migration over time at the annual level. Migration status was the binary response variable with a binomial family in the generalized mixed-effects linear model. Weather variables (minimum temperature, precipitation, and snow water equivalent) were averaged over fall migration season (September 1December 30) every year of study and were included as fixed effects along with year. To control for possible confounding effects of study and year, study was included as a random effect. Thus, the results of our model describe trends over time independent from trends unique to each study location. We generated 80 percent prediction intervals using the predictInterval function in the merTools package with 10,000 simulations for each observation (Knowles and Frederick 2020).

As a point estimate of probability of migration, we report the weighted mean. We note that prediction intervals for a mixed effects model (rather than confidence intervals) provide a more conservative and usefully interpretable metric for additional observations within a mixed design accounting for the uncertainty of the main effect estimates and the random effects (Nagashima et al. 2019).

Weather impacts on survival

Impacts of weather variables (minimum temperature, snow water equivalent, and precipitation) on individual survival were assessed using a cox proportional hazard model with temporally varying covariates using function coxph in R package survival (Therneau and Lumley 2015). Time steps consisted of two-week intervals with the first step corresponding to the first capture date from the first study (Lewis 1990) and the final time step corresponding with the last censor date from the final study (Norton 2015). White-tailed deer database structure included a row for each individual summarizing weather variables for each two-week time step that the individual was active in the study, the final time step for the individual indicated whether end of record was a mortality or censoring (if the individual survived the study entirety). Temporally varying covariates (summarized over two weeks) included mean minimum temperature, precipitation, and snow water equivalent. To account for possible study effects, study was included in the model as a static covariate.

We additionally tested for an interaction between minimum temperature and migration on survival as we hypothesized that survival costs are greatest for residential individuals, but not migratory individuals, when winters are severe. However, we were unable to include both migration and study as covariates in the initial individual survival model because some studies have very limited numbers of migrants or residents, and thus this model would not converge. Instead, we assessed impacts of migration and weather with a separate cox proportional hazard model with the interaction between migration and minimum temperature using function coxph in R package survival (Therneau and Lumley 2015). As before, precipitation and snow water equivalent were included in the model as temporally varying covariates.

Survival between migrators and residents

To quantify and compare the seasonal changes in mortality risk for migratory and residents white-tailed deer, we used function fitcyclomort in package cyclomort (Gurarie et al. 2020) in R to analyze impact of migration on survival. The fitcyclomort function fits a multi-modal periodic hazard function to right-censored survival data to estimate seasonal peaks of mortality using a likelihood-based approach (Gurarie et al. 2020). We used Akaike’s Information Criterion (AIC) to select the number of seasonal peaks in each dataset and determined annual mean mortality, timing of peak hazard, and duration of the mortality pulse for residents and migratory individuals with function fitcyclomort. We compared mean mortality and timing of peak hazard between migratory and resident individuals. To determine if hunting had a greater impact on migratory or resident white-tailed deer, we censored hunting mortalities from the data set and compared seasonal patterns of mortality using the cyclomort approach across all four studies with and without hunting mortality.

Migration changes

Seasonal means (September 1–December 31) of weather variables (minimum temperature, precipitation, and snow water equivalent) did not affect probability of white-tailed deer migrating (Table 1, Figure 2). Year was an important predictor of migration and the probability of migration significantly decreased over time (Figure 1a), even after accounting for study effects. White-tailed deer in the most recent study (Norton 2015) exhibited the lowest probability of migration and those in the oldest study (Lewis 1990) had the highest probability of migration. The total winter severity has declined from 1985-2014 and minimum temperatures for the winter season (September-February) have been slightly increasing from 1985-2014 (Table S5). Overall migratory behavior declined from 75% in 1987 to 11% percent in 2014.

Figure 2.

Trends in white-tailed deer Odocoileus virginianus migration over time and across environmental conditions. Partial residual plots depicting relationships between mean seasonal weather variables (September 1st- December 30th), Year, and probability of migrating in female white-tailed deer based on mixed-effects models. Solid line represents the predicted probability of migration based on each variable with 80% prediction intervals (gray shaded regions). The probability of female white-tailed deer migrating decreases through time with year being a significant predictor (p-value < 0.001). Data are from Great Lakes region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Climate data from Daymet; variables had a 1 x 1-km spatial resolution and a seasonal temporal resolution, (Thornton et al. 2016).

Figure 2.

Trends in white-tailed deer Odocoileus virginianus migration over time and across environmental conditions. Partial residual plots depicting relationships between mean seasonal weather variables (September 1st- December 30th), Year, and probability of migrating in female white-tailed deer based on mixed-effects models. Solid line represents the predicted probability of migration based on each variable with 80% prediction intervals (gray shaded regions). The probability of female white-tailed deer migrating decreases through time with year being a significant predictor (p-value < 0.001). Data are from Great Lakes region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Climate data from Daymet; variables had a 1 x 1-km spatial resolution and a seasonal temporal resolution, (Thornton et al. 2016).

Close modal
Table 1.

Estimates, standard errors (Std. Error), and p-values from a generalized linear mixed-effects model with Year and seasonal mean (September 1–December 30) of weather variables (daily minimum temperature in degrees Celsius, snow water equivalent (kg/m2), precipitation (mm/day)) as predictors, study as a random effect, and migration status as the response variable (Observations = 14, R2 = 0.314). Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al., 2016). White-tailed deer Odocoileus virginianus data come from the following studies conducted in the Great Lakes Region of North America: Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).

Estimates, standard errors (Std. Error), and p-values from a generalized linear mixed-effects model with Year and seasonal mean (September 1–December 30) of weather variables (daily minimum temperature in degrees Celsius, snow water equivalent (kg/m2), precipitation (mm/day)) as predictors, study as a random effect, and migration status as the response variable (Observations = 14, R2 = 0.314). Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al., 2016). White-tailed deer Odocoileus virginianus data come from the following studies conducted in the Great Lakes Region of North America: Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).
Estimates, standard errors (Std. Error), and p-values from a generalized linear mixed-effects model with Year and seasonal mean (September 1–December 30) of weather variables (daily minimum temperature in degrees Celsius, snow water equivalent (kg/m2), precipitation (mm/day)) as predictors, study as a random effect, and migration status as the response variable (Observations = 14, R2 = 0.314). Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al., 2016). White-tailed deer Odocoileus virginianus data come from the following studies conducted in the Great Lakes Region of North America: Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).

Weather impacts on survival

Snow water equivalent and precipitation did not affect annual individual white-tailed deer survival. There was a negative relationship between minimum temperature and hazard risk (Table 2, Figure 3b). Lower minimum temperatures corresponded with higher rates of mortality across all studies (p = 0.05, β = 3.7, +/–0.3; Figure 3b). Annual individual white-tailed deer survival did not differ significantly across study (Table 2, Figure 3d). The interaction between migration and minimum temperature did not influence individual white-tailed deer survival (Table S6).

Figure 3.

Relationships between white-tailed deer Odocoileus virginianus survival and environmental conditions. Odds ratios (y-axis) and confidence intervals (gray shaded regions) from the Cox proportional hazards model of annual survival in white-tailed deer with weather variables (daily minimum 2-meter air temperature in degrees Celsius, snow water equivalent (kg/m2), 4 and precipitation (mm/day)) averaged over two-week time intervals as predictors and survival as the response variable. Climate data from Daymet, variables had a 1-km2 spatial resolution and a seasonal temporal resolution, (Thornton et al., 2016). White-tailed deer data from the Great Lakes Region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Higher odds ratios correspond with a higher probability of mortality.

Figure 3.

Relationships between white-tailed deer Odocoileus virginianus survival and environmental conditions. Odds ratios (y-axis) and confidence intervals (gray shaded regions) from the Cox proportional hazards model of annual survival in white-tailed deer with weather variables (daily minimum 2-meter air temperature in degrees Celsius, snow water equivalent (kg/m2), 4 and precipitation (mm/day)) averaged over two-week time intervals as predictors and survival as the response variable. Climate data from Daymet, variables had a 1-km2 spatial resolution and a seasonal temporal resolution, (Thornton et al., 2016). White-tailed deer data from the Great Lakes Region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Higher odds ratios correspond with a higher probability of mortality.

Close modal
Table 2.

Chi square statistics, degrees of freedom (Df), and p-values from cox proportional hazards survival model. Time varying covariates included Study, minimum temperature (daily minimum 2-meter air temperature in degrees Celsius), snow water equivalent (kg/m2), and precipitation (mm/day) averaged annually over the winter season (September 1st-March 31st) with survival as the dependent variable. Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al. 2016). White-tailed deer Odocoileus virginianus data from the Great Lakes Region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).

Chi square statistics, degrees of freedom (Df), and p-values from cox proportional hazards survival model. Time varying covariates included Study, minimum temperature (daily minimum 2-meter air temperature in degrees Celsius), snow water equivalent (kg/m2), and precipitation (mm/day) averaged annually over the winter season (September 1st-March 31st) with survival as the dependent variable. Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al. 2016). White-tailed deer Odocoileus virginianus data from the Great Lakes Region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).
Chi square statistics, degrees of freedom (Df), and p-values from cox proportional hazards survival model. Time varying covariates included Study, minimum temperature (daily minimum 2-meter air temperature in degrees Celsius), snow water equivalent (kg/m2), and precipitation (mm/day) averaged annually over the winter season (September 1st-March 31st) with survival as the dependent variable. Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al. 2016). White-tailed deer Odocoileus virginianus data from the Great Lakes Region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA) and Norton (2015; northern Wisconsin, USA).

Seasons of mortality

The AIC-optimal models for seasons of mortality for both resident white-tailed deer and migratory white-tailed deer suggested two seasons of mortality indicating that mean daily hazard peaked at two points during the year (Table S2). The first peak in mortality for resident white-tailed deer occurred during late winter/early spring (∼March 2) and for migratory white-tailed deer the first peak in morality occurred during late spring (∼April 30, Table S3, Figure 4). For resident white-tailed deer the first peak in mortality was not significant with confidence intervals overlapping zero, for migratory white-tailed deer the first peak was significant, and duration of the mortality peak was about 74 days. The second peak in mortality for resident and migratory white-tailed deer was late fall/early winter, corresponding to nine-day gun hunting season (∼November 25), the second peak in mortality averaged between 4 and 8 days for migratory and resident white-tailed deer. The periodic hazard model fit three seasons for the resident white-tailed deer and two seasons for the migratory white-tailed deer. This may have been, in part, because the sample sizes were smaller for the latter (n = 95 vs. n = 77; Table S2).

Figure 4.

Annual mortality rates for female white-tailed deer Odocoileus virginianus by individual study. Data was sourced from Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Data are from the Great Lakes region of North America. X-axis represents the mean annual mortality for resident white-tailed deer in each study, the y-axis represents the mean annual mortality for migratory white-tailed deer in each study, and error bars represent SEs across years. The solid gray line represents equal mortality between migratory and resident white-tailed deer. The closed circle represents the mortality for all studies.

Figure 4.

Annual mortality rates for female white-tailed deer Odocoileus virginianus by individual study. Data was sourced from Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA). Data are from the Great Lakes region of North America. X-axis represents the mean annual mortality for resident white-tailed deer in each study, the y-axis represents the mean annual mortality for migratory white-tailed deer in each study, and error bars represent SEs across years. The solid gray line represents equal mortality between migratory and resident white-tailed deer. The closed circle represents the mortality for all studies.

Close modal

Survival between migrators and non-migrators

There was considerable variation in overall survival between the different studies (Figure 5, Table S4). Interestingly, the highest mortality rates occurred in Lewis’s (1990) study, whereas Norton’s (2015) study had the lowest mortality rates (Table S4). These two studies were from the same site (NWI), separated by 26 years. In most studies, annual survival was similar between migratory and resident white-tailed deer (Table S4), except for Van Deelen’s (1995) data, which showed significantly higher survival for migratory individuals (0.79 [SE = 0.04] vs. 0.62 [SE = 0.06]). Norton (2015) observed lower survival for migratory white-tailed deer (0.66 ± 0.17 vs. 0.9 ± 0.02) but with a very small sample size for migratory white-tailed deer (n = 4). Although not significant, mean annual mortality as estimated from fitting the periodic hazard function was slightly lower for resident white-tailed deer than migratory white-tailed deer for all animals pooled across all four studies (0.21 [95CIs 0.17, 0.26] vs. 0.31 [0.25, 0.37]; Table S3).

Figure 5.

The timing of mortality throughout the annual cycle. Annual mean hazard for migratory (a) and resident (b) female white-tailed deer Odocoileus virginianus. Histograms display numbers of mortalities, and the solid line depicts the mean hazard throughout the year. The first peak in mortality for migratory white-tailed deer (a) lasts ∼74 days, whereas the first peak for resident white-tailed deer (b) is not a peak in mortality, rather the mortality is spread throughout the late winter into spring. The second peak in mortality for both migratory and resident white-tailed deer is ∼4-8 days and corresponds with hunting season. Shaded area represents confidence intervals for the mean hazard. Data are from Great Lakes region of the North America, Lewis (1990 northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA).

Figure 5.

The timing of mortality throughout the annual cycle. Annual mean hazard for migratory (a) and resident (b) female white-tailed deer Odocoileus virginianus. Histograms display numbers of mortalities, and the solid line depicts the mean hazard throughout the year. The first peak in mortality for migratory white-tailed deer (a) lasts ∼74 days, whereas the first peak for resident white-tailed deer (b) is not a peak in mortality, rather the mortality is spread throughout the late winter into spring. The second peak in mortality for both migratory and resident white-tailed deer is ∼4-8 days and corresponds with hunting season. Shaded area represents confidence intervals for the mean hazard. Data are from Great Lakes region of the North America, Lewis (1990 northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2015; northern Wisconsin, USA).

Close modal

There was no evidence that hunting selected for migratory white-tailed deer over resident white-tailed deer or vice versa. Without hunting, the Lewis (1990) data had slightly higher mortality for migratory white-tailed deer (0.54 [SE = 0.22] vs. 0.38 [SE = 0.22]) and with hunting annual mortality was equal between migratory and resident white-tailed deer (∼0.64 [SE = 0.1]; Figure S3a). Mortality for resident white-tailed deer increased slightly in the Van Deelen (1995) data set with hunting (0.33 [SE = 0.12] vs. 0.38 [SE = 0.06]), whereas migratory mortality stayed about the same (Figure S3b). Mortality for both migratory and resident white-tailed deer increased with hunting in the Sitar (1996) data set, although the increase was negligible based on overlapping confidence intervals (Figure S3c). Finally, the Norton data set did not have enough mortalities in either the hunting or the hunting-excluded data set to determine if hunting was having a greater effect on migratory or resident white-tailed deer.

Our analysis of archival telemetry data showed that migration may impact individual survival in white-tailed deer and that migratory behavior has changed in the Great Lakes region of North America over the past 28 years, even after accounting for study site (Figure 2a). Our results indicated that migratory white-tailed deer had slightly higher mortality rates than resident white-tailed deer (31% vs. 21%), suggesting a potential fitness cost associated with migration. While the difference we observed in mortality between resident and migratory individuals was not significant at α < 0.5, the lack of significance between residents and migrants may be because of a very small sample size of migratory white-tailed deer in the most recent study (Norton 2015) or because of the higher mortality observed in resident white-tailed deer in the Van Deelen study (1995). The Van Deelen study occurred on the north shore of Lake Michigan which can experience higher snow loads because of lake effects (Hoving 2015). Resident white-tailed deer in this area may experience higher mortality because of the increased snow depths and lack of shelter from deer yards. Thus, the higher mortality of resident individuals in the Van Deelen study may be contributing to a more similar mortality rate between migrants and residents for all studies combined. Migratory white-tailed deer also had a later peak of mortality during spring (∼April 30th; Figure 5a), whereas mortality for resident white-tailed deer is spread throughout the spring (Figure 5b). A spike in mortality during spring is an expected endpoint to winter starvation, and may result from additional energetic costs associated with spring migration (Alerstam et al. 2003).

Our analysis suggested there were fewer females migrating later in the time series (Figure 2a), possibly because of a breakdown of migratory tradition. Additionally, our results suggest that migration was not influenced by changes in precipitation (Figure 2c), snow water equivalent (Figure 2d), or minimum temperature (Figure 2b), although, our data did not capture year to year variation in migration patterns because migratory status was treated as a time-invariant state variable, if an individual migrated at least once during the study it was labeled as migratory even if it did not migrate every year. Lower minimum temperatures increased the hazard risk, supporting the hypothesis that white-tailed deer migrate to deer yards, large conifer stands that provide thermal cover and reduce energetic costs during severe winters (Ozoga 1968), to avoid severe weather conditions (Nelson 1995; Sabine et al. 2002). With winters becoming milder (Table S5), benefits of migration may decrease for migratory white-tailed deer, whereas the energetic fitness costs remain, resulting in population-level loss of migratory behaviors. However, we caution that we did not observe a significant interaction between migration and low temperatures affecting survival (Table S6).

An equilibrium theory of migration suggests that conditional migrators exist in a partially migratory population because fitness benefits associated with migrating offset mortality costs of migration (Fryxell and Holt 2013). In other partial migratory populations, such as elk Cervus elaphus in the western United States, researchers have found that survival between migrating and resident elk was not significantly different suggesting that costs of migration do not outweigh advantages in terms of survival, (Hebblewhite and Merrill 2011; Henderson et al. 2017). Our results indicated that migration is costly, as migratory individuals have higher individual mortality. Our most recent study (Norton 2015) had one mortality among migratory white-tailed deer, but there were few (n = 4) migratory white-tailed deer in the study; however, migratory white-tailed deer still exhibited higher mortality than resident white-tailed deer across our studies. If migratory white-tailed deer do not have a positive trade-off such as higher fecundity or increased winter survival, then they would incur increasing energetic costs and exposure through migration without any benefits and we would expect migratory behavior to decline (Fryxell and Holt 2013). Our results support this equilibrium theory of migration as migratory white-tailed deer have higher individual mortality and we see a decline in this behavior.

In the northern parts of their range, white-tailed deer typically migrate to ‘deer yards,’ which provide shelter from severe temperatures and deep snow (Nelson 1995; Van Deelen et al. 1998). White-tailed deer develop trail systems within these deer yards which lower energetic costs and provides easier escape from predation (Nelson and Mech 1981). Environmental cues such as decreasing temperatures and snowfall likely trigger migration to winter range areas that are used again and again (Nelson 1995; Sabine et al. 2002; Felix et al. 2007; Grovenburg et al. 2009). Our results did not show a relationship between proximal weather cues and migration (Figure 2); however, our analysis could not include conditional migrators (white-tailed deer that migrate some years but not others) but because of data collection methods considers all migratory white-tailed deer as obligate migrators, if a white-tailed deer migrated at least one year it was defined as migratory. Our results indicated lower temperatures negatively impacted individual survival, suggesting migration as a strategy to mitigate severe winters. Consequently, climate change may result in certain individuals choosing not to migrate; with increasing temperatures and milder winters reducing the benefit of migration although the fitness costs remain.

White-tailed deer migration is a behavioral adaptation to seasonally variable conditions triggered by proximal environmental cues but is also re-enforced through learning (Telfer and Kelsall 1984; Nelson 1995). White-tailed deer also have strong homing capabilities (Nelson 1995; MacDonald 1997) suggesting that migration may still occur in individuals without a learned tradition. In the Florida everglades, white-tailed deer were observed abandoning their home ranges because of a catastrophic flood. When conditions improved the white-tailed deer returned to their original ranges, (MacDonald 1997) demonstrating that white-tailed deer with no seasonal migration tradition have strong homing capabilities even when displaced over large distances. Given strong site fidelity, severe winter conditions could trigger abandonment in otherwise resident white-tailed deer. Homing behavior with serial seasonal abandonments likely is the basis for development of migratory traditions in white-tailed deer in the northern parts of their range. Partial migration is an inherently flexible behavior which may explain why it declined so dramatically in white-tailed deer over time.

Partial migration has emerged as an important conservation issue because of uncertainty about the fitness consequences and environmental drivers of migration for many species. In white-tailed deer, individual decisions to migrate have long-term impacts since migration tradition is passed down from doe to fawn. Fewer females deciding to migrate can, in time, shift a population from migratory to resident when fewer fawns are taught to migrate. Here, we see a decline in migration over 28 years, suggesting fewer fawns are learning the migratory routes of their dams. With winters becoming milder (Table S5), the benefits of migration may be diminishing for migratory white-tailed deer, resulting in most populations becoming residential. Alternatively, if migratory white-tailed deer have lower survival rates, migratory individuals could be declining because of differential demographic rates. Other variables that were not examined in this study, such as development, competition, and physiologic condition have also been found to effect migratory behavior (Berg et al. 2019). Understanding the drivers of migration can help determine how to best conserve migratory species, migratory behaviors, and their migration pathways (Berg et al. 2019). Increasing research on migration shows that individual behavior needs more attention because individuals can alter their behavior year-to-year to increase survival or ensure access to valuable resources (Nandintsetseg et al. 2019). Declines in migratory ungulates may have ecosystem level impacts and emphasizes the need to better understand these migratory traditions and how best to conserve them (Bolger et al. 2008). The changes observed in migration may not be isolated to white-tailed deer, as anthropogenic effects continue to impact ungulate migration across the globe, creating a continued need for research on partial migration.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article

Table S1. Summary data set of migratory (Mig.) and resident (Res.) white-tailed deer Odocoileus virginianus in the Great Lakes region (North America), northern Wisconsin (NWI), Michigan’s upper peninsula (UPMI), and Michigan’s lower peninsula (LPMI). Telemetry data from Lewis (1990, Van Deelen (1995), Sitar (1996), and Norton (2015). Only adult females were included. Migratory white-tailed deer were defined by non-overlapping home ranges.

Available: https://doi.org/10.3996/JFWM-23-015.S1

Table S2. Akaike Information Criterion (AIC), log likelihood estimates (logLik), and degrees of freedom (d.f.) for seasonal models of mortality based on likelihood ratio tests (Gurarie et al., 2019) for resident female white-tailed deer Odocoileus virginianus in the upper Great Lakes region of North America. Data from Lewis (1990; NWI), Van Deelen (1995; UPMI), Sitar (1996; LPMI), and Norton (2015; NWI). First column represents number of seasons of mortality, because of lack of data for migratory white-tailed deer only a max of 2 seasons could be computed.

Available: https://doi.org/10.3996/JFWM-23-015.S2

Table S3. Annual mean mortality rates for migratory and resident white-tailed deer Odocoileus virginianus with 90 percent confidence intervals. Peak 1 represents day of the year of the first peak in mortality (∼March 2 for resident white-tailed deer, ∼April 30 for migratory white-tailed deer). Peak 1 for resident is not significant as mortalities are spread out, peak 1 for migratory white-tailed deer has a duration of ∼74 days. Peak 2 represented the second annual peak in mortality for migratory and resident white-tailed deer (both ∼November 26th, duration for both ∼4-8 days). Data from white-tailed deer in the Great Lakes region of North America, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2014; northern Wisconsin, USA).

Available: https://doi.org/10.3996/JFWM-23-015.S3

Table S4. Annual mean mortality rates for migratory and resident white-tailed deer Odocoileus virginianus with standard errors divided by study, Lewis (1990; northern Wisconsin, USA), Van Deelen (1995; Michigan’s upper peninsula, USA), Sitar (1996; Michigan’s lower peninsula, USA), and Norton (2014; northern Wisconsin, USA). The first row represents the annual mean mortality rate for all the studies combined. Data from female white-tail deer in the Great Lakes region of North America.

Available: https://doi.org/10.3996/JFWM-23-015.S4

Table S5. The slope for the total Accumulated Winter Season Severity Slope (AWSSI) and Total Temperature AWSSI Slope from 1985-2014 for Michigan and Wisconsin (USA). Higher AWSSI scores correspond to more severe winters and lower minimum temperatures. The negative slope for total AWSSI indicates an overall decline in winter severity from 1985-2014, the negative slope for Total Temperature AWSSI indicates a decrease in severe winter temperatures. The minimum temperature slopes include minimum temperature data from our study locations from 1985-2014. For each winter month there is a positive slope showing an increase in minimum temperatures from 1985-2014.

Available: https://doi.org/10.3996/JFWM-23-015.S5

Table S6. Estimates, standard errors (Std. Error), and p-values from the Cox proportional hazards model of annual survival in white-tailed deer Odocoileus virginianus with migration status, interaction between migration and minimum temperature, and weather variables (daily minimum 2-meter air temperature in degrees Celsius, snow water equivalent (kg/m2) and precipitation (mm/day)) averaged over two-week time intervals as predictors and survival as the response variable. Climate variables from Daymet had a 1-km2 spatial resolution and a daily temporal resolution, (Thornton et al., 2016). White-tailed deer data come from the following studies conducted in the Great Lakes Region of North America: Lewis (1990; northern Wisconsin), Van Deelen (1995; Michigan’s upper peninsula), Sitar (1996; Michigan’s lower peninsula) and Norton (2015; northern Wisconsin).

Available: https://doi.org/10.3996/JFWM-23-015.S6

Figure S1. Annual mortality rates of migratory (y-axis) and resident white-tailed deer (x-axis) Odocoileus virginianus for Lewis 1986-1989 (a, northern Wisconsin, USA), Van Deelen 1992-1995 (b, Michigan’s upper peninsula, USA), and Sitar 1995-1997 (c, Michigan’s lower peninsula, USA). The Norton study (2011-2014, northern Wisconsin, USA) had negligible mortality with hunting excluded so it is not represented. The red dot represents the mortality rate without including hunting mortalities, solid lines represent confidence intervals; the teal dot represents the annual mortality rate with hunting included. The solid gray line represents the slope where migratory white-tailed deer and resident white-tailed deer mortality is equal. Overall hunting does not show selection for migratory or resident white-tailed deer and does not have a large impact on overall annual mortality.

Available: https://doi.org/10.3996/JFWM-23-015.S7

Figure S2. Annual ratio of migratory versus resident white-tailed deer Odocoileus virginianus by study. The biggest difference in migratory vs. residents is seen in Norton’s study (2011-2014), which had a max of four migratory and over 30 resident white-tailed deer throughout the study duration (Norton 2015) (Wisconsin and Michigan, USA). Lewis (1986-1989) (Lewis 1990) and Van Deelen (1992-1995) (Van Deelen 1995) had more migratory white-tailed deer each year of the study, whereas Sitar (1995-1997) (Sitar 1996) had more resident white-tailed deer. While there is a dramatic decline between the first three studies and the most recent study, each study has a negative slope showing a general decline in migratory behavior.

Available: https://doi.org/10.3996/JFWM-23-015.S8

Data S1. File contains IDs for all white-tailed deer Odocoileus virginianus included in the study with capture date, censor/mortality date, study location, and whether the white-tailed deer was migratory or not. Data file was used to determine seasons of mortality, annual survival rates between migratory and resident white-tailed deer, and if hunting selected for migratory or resident white-tailed deer.

Available: https://doi.org/10.3996/JFWM-23-015.S9 (7 KB)

Data S2. This file contains the climate data summarized in two-week time steps, indicated by the TStart and TStop columns. The duration of all four studies is broken up into two-week time steps with first week of the first study being termed two-week time step one. The TStart column indicates the two-week time step that correlates with capture date of the individual white-tailed deer Odocoileus virginianus. For example, NLP_10 was captured on 2/6/94 which corresponds to the two-week time step 2954 from all studies. Each row contains the individual white-tailed deer ID, the mean climate variables (snow water equivalent, precipitation, and minimum temperature) for that two-week period, study location, year, and mortality status for each individual for that time step. Each individual has multiple rows summarizing these data for each two-week time step it was tracked for throughout the study. TStop for the last row of each individual corresponds to the censor or mortality date for that individual. Data file was used to determine the effect of 2-week climate variables and study on white-tailed deer survival.

Available: https://doi.org/10.3996/JFWM-23-015.S10 (639 KB)

Data S3. File contains data for yearly migration analysis. Data includes years from each study, the climate variables (minimum temperature, snow water equivalent, precipitation) averaged over the fall migration season (September-December), number of migratory and resident white-tailed deer Odocoileus virginianus for that year, and study location. Data file was used to assess any population changes in migration over time.

Available: https://doi.org/10.3996/JFWM-23-015.S11 (1 KB)

We thank Timothy Lewis, Kristie Sitar, and Andrew Norton for collecting data. We also thank Wisconsin and Michigan’s Department of Natural Resources, U.P. Whitetail’s Association, and Hiawatha National Forest for their involvement in data collection. We would like to express our sincere gratitude to the reviewers for their feedback and insightful comments on our manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Alerstam
T,
Hedenstrom
A,
Akesson
S.
2003
.
Long-distance migration: evolution and determinants
.
Oikos
.
103
:
247
260
.
Berg
JE,
Hebblewhite
M,
St Clair
CC,
Merrill
EH.
2019
.
Prevalence and mechanisms of partial migration in ungulates
.
Frontiers in Ecology and Evolution
.
7
:
325
.
Bischof
R,
Loe
LE,
Meisingset
EL,
Zimmermann
B,
Van Moorter
B,
Mysterud
A,
Gaillard
AEJ-M,
McPeek
EMA.
2012
.
A migratory northern ungulate in the pursuit of spring: jumping or surfing the green wave
?
American Naturalist
.
180
:
407
424
.
Bolger
DT,
Newmark
WD,
Morrison
TA,
Doak
DF.
2008
.
The need for integrative approaches to understand and conserve migratory ungulates
.
Ecology Letters
.
11
:
63
77
.
Brinkman
TJ,
Deperno
CS,
Jenks
JA,
Haroldson
BS,
Osborn
RG.
2005
.
Movement of female white-tailed deer: effects of climate and intensive row-crop agriculture
.
Journal of Wildlife Management
.
69
:
1099
1111
.
Cagnacci
F,
Focardi
S,
Heurich
M,
Stache
A,
Hewison
AJM,
Morellet
N,
Kjellander
P,
Linnell
JDC,
Mysterud
A,
Neteler
M,
et al.
2011
.
Partial migration in roe deer: migratory and resident tactics are end points of a behavioural gradient determined by ecological factors
.
Oikos
.
120
:
1790
1802
.
Cole
EK,
Foley
AM,
Warren
JM,
Smith
BL,
Dewey
SR,
Brimeyer
DG,
Fairbanks
WS,
Sawyer
H,
Cross
PC.
2015
.
Changing migratory patterns in the Jackson elk herd
.
Journal of Wildlife Management
.
79
:
877
886
.
Cook
DB,
Hamilton
WJ.
1942
.
winter habits of white-tailed deer in central New York
.
Journal of Wildlife Management
.
6
:
287
291
.
Courtemanch
AB,
Kauffman
MJ,
Kilpatrick
S,
Dewey
SR.
2017
.
Alternative foraging strategies enable a mountain ungulate to persist after migration loss
.
Ecosphere
.
8
:
e01855
.
Fagan
WF,
Gurarie
E.
2020
.
Spatial ecology: herbivores and green waves—to surf or hang loose
?
Current Biology
.
30
:
R991
R993
.
Felix
AB,
Walsh
DP,
Hughey
BD,
Campa
H,
Winterstein
SR.
2007
.
Applying landscape-scale habitat-potential models to understand deer spatial structure and movement patterns
.
Journal of Wildlife Management
.
71
:
804
810
.
Fieberg
J,
Kuehn
DW,
DelGiudice
GD.
2008
.
Understanding variation in autumn migration of northern white-tailed deer by long-term study
.
Journal of Mammalogy
.
89
:
1529
1539
.
Fryxell
JM,
Holt
RD.
2013
.
Environmental change and the evolution of migration
.
Ecology
.
94
:
1274
1279
.
Fryxell
JM,
Sinclair
ARE.
1988
.
Causes and consequences of migration by large herbivores
.
Trends in Ecology & Evolution
.
3
:
237
241
.
Grovenburg
TW,
Jenks
JA,
Klaver
RW,
Swanson
CC,
Jacques
CN,
Todey
D.
2009
.
Seasonal movements and home ranges of white-tailed deer in north-central South Dakota
.
Canadian Journal of Zoology
.
87
:
876
885
.
Gurarie
E,
Thompson
PR,
Kelly
AP,
Larter
NC,
Fagan
WF,
Joly
K.
2020
.
For everything there is a season: analysing periodic mortality patterns with the cyclomort r package
.
Methods in Ecology and Evolution
.
11
:
129
138
.
Halpin
MZT,
Chepko-Sade
BD,
Halpin
ZT.
1987
.
Mammalian dispersal patterns: the effects of social structure on population genetics
.
University of Chicago Press
.
Harris
G,
Thirgood
S,
Hopcraft
J,
Cromsight
J,
Berger
J.
2009
.
Global decline in aggregated migrations of large terrestrial mammals
.
Endangered Species Research
.
7
:
55
76
.
Hebblewhite
M,
Merrill
EH.
2009
.
Trade-offs between predation risk and forage differ between migrant strategies in a migratory ungulate
.
Ecology
.
90
:
3445
3454
.
Hebblewhite
M,
Merrill
EH.
2011
.
Demographic balancing of migrant and resident elk in a partially migratory population through forage–predation tradeoffs
.
Oikos
.
120
:
1860
1870
.
Henderson
CR,
Mitchell
MS,
Myers
WL,
Lukacs
PM,
Nelson
GP.
2017
.
Attributes of seasonal home range influence choice of migratory strategy in white-tailed deer
.
Journal of Mammalogy
.
99
:
89
96
.
Hijmans
RJ.
2020
.
Introduction to the ‘raster’ package version 2.6-7:27
.
Hoving
CL.
2015
.
Ice, snow, and swamp: managing deer in Michigan’s changing climate
.
Michigan Journal of Sustainability
.
3
.
Knowles
JE,
Frederick
C.
2020
.
merTools: tools for analyzing mixed effect regression models, 2020
. Available: https://cran.r-project.org/web/packages/merTools/index.html (May 2024).
Lewis
TL.
1990
.
The effects of supplemental feeding on white-tailed deer in northwestern Wisconsin
. Doctoral dissertation.
Madison, WI
:
University of Wisconsin-Madison
.
Lutz
CL,
Diefenbach
DR,
Rosenberry
CS.
2015
.
Population density influences dispersal in female white-tailed deer
.
Journal of Mammalogy
.
96
:
494
501
.
MacDonald
K.
1997
.
Site fidelity and its effects on survival of Odocoileus virginianus seminolus during a catastrophic flood in the Everglades
.
Master's thesis
.
Gainesville, FL
:
University of Florida-Gainesville
.
Mason
R.
2016
.
Michigan deer management plan
.
Michigan Department of Natural Resources
. Wildlife Division Report 3626.
McNaughton
SJ.
1976
.
Serengeti migratory wildebeest: facilitation of energy flow by grazing
.
Science
.
191
:
92
94
.
Mysterud
A,
Loe
LE,
Zimmermann
B,
Bischof
R,
Veiberg
V,
Meisingset
E.
2011
.
Partial migration in expanding red deer populations at northern latitudes—a role for density dependence
?
Oikos
.
120
:
1817
1825
.
Nagashima
K,
Noma
H,
Furukawa
TA.
2019
.
Prediction intervals for random-effects meta-analysis: a confidence distribution approach
.
Statistical Methods in Medical Research
.
28
:
1689
1702
.
Nandintsetseg
D,
Bracis
C,
Olson
KA,
Böhning‐Gaese
K,
Calabrese
JM,
Chimeddorj
B,
Fagan
WF,
Fleming
CH,
Heiner
M,
Kaczensky
P,
et al.
2019
.
Challenges in the conservation of wide-ranging nomadic species
.
Journal of Applied Ecology
.
56
:
1916
1926
.
Nelson
ME.
1995
.
Winter range arrival and departure of white-tailed deer in northeastern Minnesota
.
Canadian Journal of Zoology
73
:
1069
1076
.
Nelson
ME.
1998
.
Development of migratory behavior in northern white-tailed deer
.
Canadian Journal of Zoology
.
76
:
426
432
.
Nelson
ME,
Mech
LD.
1981
.
Deer social organization and wolf predation in northeastern Minnesota
.
Wildlife Monographs
.
77
:
3
53
.
Nelson
ME,
Mech
LD.
1991
.
Wolf predation risk associated with white-tailed deer movements
.
Canadian Journal of Zoology
.
69
:
2696
2699
.
Norton
AS.
2015
.
Integration of harvest and time-to-event data used to estimate demographic parameters for white-tailed deer
.
Doctoral dissertation
.
Madison, WI
:
University of Wisconsin-Madison
.
Norton
AS,
Storm
DJ,
Van Deelen
TR.
2021
.
White-tailed deer, weather and predation: a new understanding of winter severity for predicting deer mortality
.
Journal of Wildlife Management
.
85
:
1232
1242
.
Ozoga
JJ.
1968
.
Variations in microclimate in a conifer swamp deeryard in northern Michigan
.
Journal of Wildlife Management
.
32
:
574
585
.
Peters
W,
Hebblewhite
M,
Mysterud
A,
Spitz
D,
Focardi
S,
Urbano
F,
Morellet
N,
Heurich
M,
Kjellander
P,
Linnell
JDC,
et al.
2017
.
Migration in geographic and ecological space by a large herbivore
.
Ecological Monographs
.
87
:
297
320
.
Sabine
DL,
Morrison
SF,
Whitlaw
HA,
Ballard
WB,
Forbes
GJ,
Bowman
J.
2002
.
Migration Behavior of white-tailed deer under varying winter climate regimes in New Brunswick
.
Journal of Wildlife Management
.
66
:
718
.
Sharma
S,
Couturier
S,
Côté
SD.
2009
.
Impacts of climate change on the seasonal distribution of migratory caribou
.
Global Change Biology
.
15
:
2549
2562
.
Shaw
AK.
2016
.
Drivers of animal migration and implications in changing environments
.
Evolutionary Ecology
.
30
:
991
1007
.
Sitar
KL.
1996
.
Seasonal movements, habitat use patterns, and population dynamics of white-tailed deer (Odocoileus virginianus) in an agricultural region of northern lower Michigan
. Master's thesis.
East Lansing, MI
:
Michigan State University
.
Spitz
DB,
Hebblewhite
M,
Stephenson
TR,
German
DW.
2018
.
How plastic is migratory behavior? Quantifying elevational movement in a partially migratory alpine ungulate, the Sierra Nevada bighorn sheep (Ovis canadensis sierrae)
.
Canadian Journal of Zoology
.
96
:
1385
1394
.
Telfer
ES,
Kelsall
JP.
1984
.
Adaptation of some large North American mammals for survival in snow
.
Ecology
.
65
:
1828
1834
.
Therneau
TM,
Lumley
T.
2015
.
Package ‘survival.’
R Top Doc
.
128
(
10
):
28
33
.
Thornton
PE,
Thornton
MM,
Mayer
BW,
Wilhelmi
N,
Wei
Y,
Devarakonda
R,
Cook
RB.
2014
.
Daymet: daily surface weather data on a 1-km grid for North America, version 2. Oak Ridge National Lab
. (ORNL), Oak Ridge,
TN (United States).
Available: https://www.osti.gov/biblio/1148868 (May 2024).
Van Deelen
TR.
1995
.
Seasonal migrations and mortality of white-tailed deer in Michigan’s upper peninsula
.
Doctoral dissertation
.
East Lansing, MI
:
Michigan State University
.
Van Deelen
TR,
Campa
H,
Hamady
M,
Haufler
JB
.
1998
.
Migration and seasonal range dynamics of deer using adjacent deeryards in northern Michigan
.
Journal of Wildlife Management
.
62
:
205
213
.
Walther
G-R,
Post
E,
Convey
P,
Menzel
A,
Parmesan
C,
Beebee
TJC,
Fromentin
J-M,
Hoegh-Guldberg
O,
Bairlein
F.
2002
.
Ecological responses to recent climate change
.
Nature
.
416
:
389
395
.
White
PJ,
Davis
TL,
Barnowe-Meyer
KK,
Crabtree
RL,
Garrott
RA.
2007
.
Partial migration and philopatry of Yellowstone pronghorn
.
Biological Conservation
.
135
:
502
-
510
.
Wisconsin Department of Natural Resources
.
2022a
.
Wisconsin wolf management plan
.
Wisconsin Department of Natural Resources, Bureau of Wildlife Management
.
Madison, Wisconsin, USA
.
Wisconsin Department of Natural Resources
.
2022b
.
Deer statistics
. Available: https://apps.dnr.wi.gov/deermetrics/DeerStats.aspx?R=HarvestHistory (May 2024).

Author notes

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Supplemental Material