Development of an aerial population survey method for elk (Cervus elaphus) in Rocky Mountain National Park, Colorado

Abstract

Executive Summary Since the early 1990s, substantial effort and funding have been expended to conduct research to guide development of a 20-year Elk and Vegetation Management Plan for Rocky Mountain National Park (RMNP) in Colorado. One goal of the plan is to maintain the elk ( Cervus elaphus ) population size at the lower end of the natural range of variation. To implement management actions called for in the plan, accurate and reliable population estimates are needed, as well as a better understanding of the spatial and temporal distribution of elk. The previous aerial survey protocol and population estimation model used by the park had not been updated since the model’s initial calibration more than 15 years ago and the model was developed with an insufficient number (n=44) of observations. Thus we initiated research to reevaluate, update, and improve elk population estimation protocols for RMNP. We considered several alternative survey and analysis methods, and concluded that a hybrid population-estimation model using a simultaneous double-observer technique with sighting covariates was the most appropriate and effective aerial survey methodology for this population and the environmental conditions in RMNP. Instructional protocols for conducting these surveys in the future, along with datasheets, are provided in this report’s appendixes. To develop an improved method for aerial elk surveys, we used elk radio-collar location data from our study and other studies, and applied geographic information system analyses to define the survey area and develop effective, repeatable survey transect lines. We used telemetry data from radio-collared elk to inform our understanding of the temporal and spatial scale of elk movements across the park boundary, elk use of tree cover during potential survey hours, and elk use of different elevations within their range. Determining where elk were during surveys helped to fine-tune a survey design that improved spatial cover-age and decreased costs where possible, while standardizing the method to make it repeatable from year to year. We conducted aerial helicopter surveys to test our methodology in an adaptive, iterative process during three winters: 2007–2008, 2008–2009, and 2009–2010. We gained new information on each survey and used results to refine subsequent surveys. Our results confirm that elk movements were highly dynamic with respect to park boundary crossings; an average of six round trips from the park to Estes Park, Colorado, and back per month were taken by global positioning system-collared bull elk. We observed a strong diurnal temporal pattern of bull elk use of trees; elk were found in dense tree cover from 15:00−22:00 but not as often during morning and early afternoon hours when surveys were conducted. We used information on elk use of different altitudes to refine and establish a more efficient survey area. During the time of our study, a concurrent study in the park deployed 120 very high frequency radio collars on elk cows. We used those collar locations during our flights to increase sample size and to evaluate the level of precision we would gain by using “known fates analysis” (in which elk were known to be in the survey area, out of the survey area, or deceased based on radio-collar locations collected simultaneously during aerial surveys). Using radio-collar locations reduced bias by 1.1–8.8 percent and increased precision (that is, reduced the width of confidence intervals). This report provides final population estimates analyzed with and without radio-collar data to demonstrate what is gained by using marked individuals.