| dc.description.abstract | Precise and accurate information about population numbers is crucial within wildlife ecology, for example, to effectively manage disease threats. After chronic wasting disease was detected in a reindeer (Rangifer tarandus) population in Norway, concerns that moose and red deer in the area would contract the fatal disease arose. Aiming to lower the probability of such a spillover, the Norwegian Environment Agency recommended that the at-risk moose and red deer populations should be reduced to less than one animal per km^2. However, accurate and precise estimates on absolute densities of deer are difficult, often impossible, to obtain using traditional data collection methods. This is especially true for elusive species living in inaccessible areas. Current estimates of red deer (Cervus elaphus) abundance and density in Norway are unreliable with no known degrees of uncertainties. Thus, these estimates do not suffice when it comes to determining the harvest quotas needed to reach the population density goal set by the authorities. The purpose of this thesis was therefore to develop a method for more reliable population estimates of red deer using drones. In a case study approach, the drones were used to collect data from four different sampling areas in Lærdal, Norway. Detections of red deer in the drone images were then recorded using a double-observer protocol and hierarchal state-space models were fitted to the data using a Bayesian approach to obtain posterior distributions of absolute deer density. Although the produced critical intervals were rather wide, the results revealed that absolute red deer density estimates with quantifiable uncertainties can be produced using this method. Furthermore, the method showed great potential for reliable spatiotemporal comparisons of deer density estimates. | eng |