Estimating winter habitat use and population abundances of large ungulates at Fort Riley Military Reservation

Abstract

Rocky Mountain elk (Cervus canadensis) populations have been reintroduced across North America including the Fort Riley Military Reservation (FRMR) in the tallgrass prairie region of Kansas, USA. Although habitat use by elk has been assessed in many areas of North America, we have limited understanding of how elk use space in prairie landscapes and under multiple habitat-management regimes. I used detection/non-detection data from 149 camera-trap sites on FRMR to assess winter habitat use by elk in relation to landscape composition and management efforts. Specifically, I used single-season occupancy models to investigate the effects of woody encroachment, prairie landcover, food plots, distance to refugia and water, and burn history on site occupancy by elk in January - February 2023. The most supported-model indicated that burn timing was the best predictor of elk habitat use. Elk were more likely to occur at sites recently burned than sites that were not burned > 24 months. Interestingly, I did not detect an effect of amount of food plots or permanent water sources near sites. The results demonstrate that managed fire regimes may benefit elk during winter in the lower Great Plains where winter temperatures can be moderate. Precise estimation of population abundances is a vital consideration for biologist who develop harvest regulations and manage populations. Traditionally, methods such as aerial census surveys and spotlight surveys are used to estimate abundances of Rocky Mountain elk and white-tailed deer (Odocoileus virginianus). These approaches can be logistically challenging and may not reveal precise population estimates. Thus, there is a need to evaluate the effectiveness of complimentary abundance estimation procedures. I used presence/absence data from 146 camera-trap sites to evaluate the usefulness of space-to-event (STE) models for estimating deer and elk abundance on Fort Riley Military Reservation in February 2023. I compared The STE estimates to complete census surveys (elk) and distance sampling (deer) currently used by FRMR biologists. Abundance estimates derived from STE models for elk (197.8, 95% CI = 163.7 - 238.9) and deer (2,290.0, 95% CI = 2,140.9 - 2,449.5) were slightly greater than those derived from census sampling for elk (102) and distance sampling for deer (1,526.9, 95% CI = 1,078.8 - 2,160.9), though precision was much better with STE estimators. The results demonstrate that STE models can be a useful and cost-effective tool and the primary impact of using these methods is that they allow for concurrent estimation of elk and deer abundances at large spatial scales. The primary contribution of this work lies in the science-based information provided to FRMR biologists. Specifically, it enhances their understanding of habitat management techniques they use through prescribed burns and offers an efficient method to estimate ungulate abundance within the FRMR area.