In the Menišija area (107 km²), we compared four methods for estimating the abundance and density of red deer and roe deer. The faecal pellet group count with prior clearance was conducted on 100 plots of 400 m² each. The plots were cleaned before winter, and accumulated dung was counted in spring after the snow had melted and before vegetation started growing. Density estimates using camera traps were also conducted on the same plots. The age-at-harvest method for density reconstruction was limited to quadrants within the study area, where data on culling of both target species were collected. Thermal camera drone surveys were conducted along 40 linear transects ranging from 1 to 1.2 km in length. For all methods, we assessed density, its confidence intervals and implementation costs, and then described logistical requirements. For red deer, faecal pellet group count estimated 5.22 individuals/km², camera traps 8.74, age-at-harvest 8.76, and drone ranged from 15.4 to 25.4 (various derivations). Confidence interval widths differed between methods and were particularly wide for drone estimates. We calculated the required sampling intensity for each method to achieve relative accuracy equivalent to the most accurate reference method. This correction was used to adjust the cost estimates for all methods. Considering the correction, the total costs for age-at-harvest would be €0, dung counting €5,815, camera traps €28,215, and drone €9,875. For roe deer we found the following densities: faecal pellet group count 1.51, camera traps 1.17, and age-at-harvest 4.69 ind./km². None of these provided a fully accurate estimate. Adjusting for activity and forest use, camera trap estimates for roe deer would likely be comparable to those derived from the age-at-harvest. For red deer, the results from all methods are very similar, except for drones; faecal pellet group count slightly underestimates, but it is a simple and inexpensive method. Camera traps likely provide the most accurate results, but high costs associated with manual review of images and analyses limit their applicability, which is expected to change with automation. Drone provided the highest density estimates, but with wide confidence intervals. To obtain approximate densities, the age-at-harvest method proved most cost-effective, though it provides no measure of precision. This method is also sensitive to rapid population changes, as it does not necessarily reflect current conditions due to time delay. It can also underestimate density as not all mortality is recorded.
|
