Detection of spatial avoidance between sousliks and moles by combining field observations, remote sensing and deep learning techniques.

Nowadays, remote sensing is being increasingly applied in ecology and conservation, and even underground animals can successfully be studied if they leave clear signs of their presence in the environment. In this work, by combining a field study, analysis of high-resolution aerial images, and machine learning techniques, we investigated the interspecies relationships of two small burrowing mammals: the spotted souslik Spermophilus suslicus and the European mole Talpa europaea. The study was conducted for 3 years (2018-2020) at a 105-ha grass airfield where both species coexist (Poland). Both field studies and the analysis of aerial imagery showed that, in the period of low population numbers, the souslik avoided coexistence with the European mole, and the presence of the mole was found to reduce the area of the habitat suitable for the souslik. The presence of other burrowing species may be an important element in the habitat selectivity of the souslik, but this has not yet been included in the conservation guidelines for this species. We discuss the contribution of our results to the knowledge of the ecology of burrowing mammals and their interspecies relationships. We also assess the possibility of using remote sensing and deep learning methods in ecology and conservation of small burrowing mammals.