Weather conditions associated with autumn migration by mule deer in Wyoming

Introduction

Maintaining ecological integrity necessitates a proactive approach of identifying threats to species, communities, and the ecological processes that sustain them. For mule deer (Odocoileus hemionus), a culturally and economically important species (Copeland et al., 2014), habitat fragmentation can present barriers to migration, alter migration routes, or increase mortality through deer-vehicle collisions (Sawyer, LeBeau & Hart, 2012).

Highway underpasses have restored connectivity for mule deer in fragmented landscapes (Reed, Woodward & Pojar, 1975; Ng et al., 2004; Clevenger & Waltho, 2005; Braden et al., 2008; Gagnon et al., 2011). The unique structure of highway underpasses and associated fencing, when coupled with remote cameras and weather observations, presents an opportunity to gain substantial information on population age and sex structure (Ikeda et al., 2013), abundance (Rowcliffe et al., 2008), and proximate cues associated with migratory movement. Yet, the relationships between migratory movements and weather conditions are understudied, despite the importance of this for informing policy and adaptive management decisions regarding connectivity in the context of a changing climate.

Our objective was to identify weather conditions associated with autumn migration by mule deer in Wyoming. Our central hypothesis is that entering the winter range too early comes at the expense of reproductive output and survival of young mule deer. Our basis for this hypothesis arises from known nutritional limitations and energetic costs of winter to mule deer (Bartmann, White & Carpenter, 1992; Poole & Mowat, 2005; Bishop et al., 2009); premature entry to the winter grounds presumably provides no advantage to adults or offspring. Based on this hypothesis, we predict that mule deer movements through the underpass will be associated with the timing of snowfall events and snow depth that cover forage on summer grounds. Identification of the specific cue(s) used by mule deer during migration may allow managers to anticipate changes in migration based on weather conditions.

Methods

Study site and camera set-up

We collected animal migration data using a trail camera set up to monitor a highway underpass on Hwy. 789 approximately 8 km north of Baggs, Wyoming (Fig. 1). The underpass was installed in 2009 and equipped with a RECONYX Hyperfire (TM) camera (Reconyx, Holmen, Wisconsin, USA) mounted approximately 1.5 m high centered within the underpass. The camera was pointed towards the direction animals were migrating from, so during autumn the camera was pointed east. Camera settings included a distance from camera to subjects of 18.2 m with a 1/5 s trigger speed; three photos recorded when motion was detected, and photo resolution of 1080P High Definition or 3.1 Mega-pixels. Cameras remained active throughout the year; however, during autumn migrations by mule deer the images were downloaded more frequently. Autumn migration dates were November 1 to December 31 of each year. The Wyoming Game and Fish Department approved this research (Permit #791).

The study site was defined by the Baggs Mule Deer Herd Unit, which encompasses 1,092 km² south of I-80 in southern Carbon County (Fig. 1). This area supports a variety of vegetation types, but is generally characterized by rolling topography, prominent ridges, and dry canyons dominated by sagebrush (Artemisia sp.), black greasewood (Sacrobatus vermiculatus), Utah juniper (Juniperus osteosperma), and other mixed-shrub (Purshia tridentata, Prunus virginiana, Amelanchier alnifolia, Chrysothamnus sp., Cercocarpus sp.). Elevations range from 1,920 to 2,530 m.

Results

Analysis of 19,480 images acquired during fall migration 2009 and fall migration 2010 documented 6,628 counts of mule deer using the underpass (Table 1). The population age and sex structure was consistent across the two years with a mean of 61 fawns, 8 yearling bucks, and 14 adult bucks per 100 does (Table 1).

Model-selection results indicated minimum air temperature and snow depth were the best proximate cues associated with autumn migration. In 2009, minimum air temperature of the previous day was the most supported model for does and bucks, with maximum air temperature of the previous day competing with the most supported model for bucks only (Table 2). In 2010, snow depth was the most supported model for does and bucks with no competing models for either sex. Higher counts of deer occurred on days with lower minimum temperatures in 2009 and on days with greater snow depth in 2010 (Table 3). Bucks and does responded similarly to weather conditions within and across years despite migrating in separate groups.

We used coefficients from the top fitted models to predict use of the highway underpass under the range of weather conditions observed in 2009 and 2010 (Fig. 2). Early, extreme minimum air temperatures and snow depths had the highest predicted counts of does and bucks. Based on these model predictions, thresholds for onset of migration were minimum air temperature of 0 to −5 °C or snow depth exceeding 25.4 cm.

Discussion

Wildlife underpasses and overpasses are used throughout western North America to restore connectivity for migratory and large-ranging species. We demonstrated that monitoring wildlife underpasses provides important demographic and movement information, which when coupled with weather observations can be used to understand connectivity and migratory dynamics associated with weather conditions. The association between weather conditions and counts of deer using the underpass was consistent with our expectation. Specifically, we identified minimal daily temperature and snow depth as proximate cues used by mule deer during migration to winter range. These weather cues were consistent across does and bucks, but differed slightly by year. Additionally, extreme early season snow depth exceeding 25.4 cm or minimum air temperature of 0 to −5 °C may be associated with onset of migration.

Our results using a single camera trap were consistent with radio-telemetry studies of mule deer migration. Previous studies identified decreasing daily temperature, increasing snow depth, and senescing vegetation as factors associated with autumn migration from high-elevation summer ranges to low-elevation winter ranges (Garrott et al., 1987; Nicholson, Bowyer & Kie, 1997; Monteith et al., 2011). However, the specific relationships varied by vegetation type, elevation, and climate. In the sagebrush-steppe ecosystem of the Sierra Nevada Mountains in California, the onset of migration coincided with average daily temperature <5 °C and snow depth >0 cm (Monteith et al., 2011). In a pinyon pine-Utah juniper shrubland complex in northwestern Colorado, mule deer migration occurred with average daily temperature >0 °C and no snow depth (Garrott et al., 1987). In the San Bernardino Mountains of southern California, mule deer exhibited partial migration in response to snow cover (Nicholson, Bowyer & Kie, 1997).

Installation of under and overpasses are effective in reducing deer-vehicle collisions (Clevenger, Chruszez & Gunson, 2001). However, under and overpasses may be cost prohibitive and thus alternative methods of migration route protection may need to be developed. In this situation, variability in the number of mule deer entering the winter range poses substantial challenges to managing mule deer-human interactions. One alternative is to use weather information to better time the use of temporary signage on roads crossed by migration routes. Temporary signage could indicate speed reduction or simply increase awareness that migration movements are more likely on days with suitable weather conditions.

In addition to providing information that could be used to more effectively manage migration route-road crossings, acquiring images of deer using underpasses and overpasses may also increase the efficiency of monitoring efforts by state agencies. Long-term population monitoring for setting harvest regulations typically uses aerial surveys of summer grounds conducted just after the hunting season and prior to migration to wintering grounds. A concern with post-hunt aerial surveys is the potential to miss animals or have a biased sample if some classes of animals migrate prior to the survey. In these situations, underpass and overpass data may complement post-hunt aerial surveys by providing additional sampling periods during spring and fall migration. Having multiple sampling periods would enable within-year estimates of survival and recruitment as opposed to a single, annual estimate.

Conclusions

Underpass and overpass data coupled with meteorological observations can be used to understand the relationship between weather and migration. With short-term data, we identified minimum daily temperature and snow depth as proximate cues associated with mule deer migration to winter range in Wyoming. Long-term data will provide information on migration dynamics, including whether the onset, duration, or magnitude of migration co-changes with weather. Long-term data may also reveal the role of weather in partial migration, cessation of migration or changes in route location over time.

Supplemental Information