Variations in random encounter model density estimates of European mouflon island population

Estimating wildlife density is challenging, and various methods have been developed to account for imperfect detection probability. In this study, we applied the Random encounter model (REM) to estimate the population density of the European mouflon Ovis aries on the Mediterranean island of Dugi Otok, Croatia. Our goal was to test if density estimations vary throughout the year in a closed population. We systematically placed 27 camera traps over an area of 35 km2 at the intersections of 1.5-km grid cells and kept them active for 12 months (March 2022 – March 2023). All REM parameters (i.e., average movement speed, angle and radius) were estimated from the camera trap data. To explore potential correlations of mean temperature and humidity with REM parameters, six data loggers were placed systematically. The mouflon was recorded at 21 locations in 1,227 independent events on 6,719 camera-trapping days. The lowest density was recorded in winter while the highest density was observed in autumn. Our findings indicate that mouflon changed activity patterns between July and October, with trap rates increasing during the rut, which led to a fluctuation in density estimates. Correlations of temperature and humidity with the trapping rate were statistically non-significant. The trend in density estimates from two-month data subsets suggests that the largest variation in density estimation occurred between June and November, coinciding with high temperatures and the rut season. This research suggests that density estimation using REM may be influenced by sampling bias driven by seasonal variation in animal behaviour. Consequently, integrating behavioural ecology into survey design is critical to an appropriate study planning and interpretation of results.