Comparison of population density estimation methods for roe deer

Abstract

Populations of European roe deer (Capreolus capreolus) are steadily growing in Europe and in Hungary. In order to manage this game species efficiently and to reduce the conflicts related to deer (crop damage, car collisions), it is important to follow their density changes as accurately as possible. The principles of adaptive deer management based on bioindicators could greatly help the work of hunters in the Hungarian Great Plain, but this would require data collected using appropriate methods. New methods and equipment in wildlife management, such as the thermal camera, may offer a new opportunity to survey roe deer populations. In this study, we compared the results of counting roe deer from a car along transects by daylight (0–500 m) and by night with spotlight (0–250 m), as well as from observation points with a thermal camera (in both distance intervals). The investigations were conducted in three lowland small game hunting areas of low forest cover in the No. 101. Tiszazugi Wildlife Management Landscape Unit. We also examined the effectiveness of using a thermal camera at an observation distance greater than the reflector could be used, comparing the distance classes between 0 and 250 m and 250 and 500 m. We performed all three estimation methods for the same five days. There was a significant difference among the population densities determined by the four estimation procedures. The spotlight estimation method gave the highest average value (18,7 individuals/100 ha; SD = 5,2), meanwhile the daylight transect estimation provided the lowest one (11,5 individuals / 100 ha; SD = 3,6). Method using thermal camera resulted in intermediate values between the two other methods (0–250 m: 17,7 individuals/100 ha; SD=6,3; n=5; 0–500 m: 11,6 individuals/100 ha; SD=6,4; n=5). However, post-hoc tests could not reveal any significant differences among the data from different methods. In the case of the thermal camera method, in the distance class between 250 and 500 m, the observed individual density was less than half (8,7 individuals/100 ha; SD=13,9) than in the distance class between 0 and 250 m. Therefore, the detected number of the deer individuals by thermal camera decreased significantly with increasing observation distance beyond the effective range of spotlight. The smallest variance was shown by the data from the daytime transect study, but this method results in an underestimation due to the decreased daytime activity of the roe deer. For the hunting units, the night spotlighting along transects is primarily recommended to determine the minimum population roe deer density, as we were able to detect the most roe deer using this method. Human resource, time and cost requirements of this method are also relatively low and results in the slightest underestimation. The efficiency, human and time costs of the thermal imager might reach a similar level to that when using for a range between 0–250 m. But its high price could be a limit for wide application.