Detecting reindeer carrion in the Arctic tundra of Svalbard: An exploratory study on spectral dataacquisition and utilization for CDI detection

By Petra Bardócz

Abstract
There is a growing body of research focused on understanding how vertebrate carrion affects both local and ecosystem-wide processes. Essential to this is quantifying available carrion, particularly in nutrient-poor ecosystems like the Arctic tundra of Svalbard. This study investigates the use of remote sensing methods to detect Svalbard reindeer (Rangifer tarandus platyrhynchus) carcasses, the archipelago's largest ungulate species. The research utilized spectral data collected via field spectrometry and UAV-mounted multispectral cameras during a field campaign in August 2023. Two supervised classification algorithms (the spectral angle mapper and a random forest model) were employed to discriminate the cadavers in the UAV-derived orthomosaics. Results showed fresh reindeer carcasses can be detected in UAV images due to their high fur reflectance in the visible region, using a random forest model. The study also examined how arctic climatic factors affect spectral data acquisition over reindeer carcasses and identified a set of good practices that can be employed to enhance data quality during future campaigns.