Assessing the activity of deer and their influence on vegetation in a wetland using automatic cameras and low altitude remote sensing (LARS)

Abstract

It was thought by the author of the present study that if both automatic cameras and LARS were used, instead of one or the other, then it would be possible to efficiently gain knowledge that could be applied in the strategic management of deer in wetlands. In the Mizorogaike wetland, deer (Cervus nippon) graze plants throughout its entire area. Aerial images taken in the winter were used in the deer trail survey, as the vegetation would be withered, and the deer trails become easier to identify. The rare species selected to evaluate the effect of deer on vegetation was Menyanthes trifoliata, which was reported to be on the decline in the Mizorogaike wetland. The deer trails were identified visually. M. trifoliata was detected in the images using deep learning. I investigated the relationship between increases and decreases in areas covered by M. trifoliata during the period from 2016 to 2019 and the distance of these areas from deer trails. Three routes were selected where the deer trails indicated that the deer might be crossing the water. Four camera traps were set up on the land side facing the water. In the Mizorogaike wetland, during the period from 2015 to 2019, the area where deer trails could be observed spread to every corner of the wetland. Deer trails became conspicuous at the southwest edge of the Mizorogaike wetland. Although in 2015 the deer trails were limited to the northern half of the floating mat, by 2019, they had spread down into the southern half. Although camera traps were only used to investigate the wetland in 2019, it is thought that deer began to regularly swim across the water during the period from 2015 to 2019 and that this led to trails forming in the south of the floating mat. The area covered by M. trifoliata decreased by 20% from 2016 to 2019. The areas where M. trifoliata cover decreased tended to be nearer to deer trails than the areas where it increased.