Winter habitat suitability and spatial niche overlap between Japanese serow (Capricornis crispus) and sika deer (Cervus nippon) in Shonai and Minamiaizu, Japan

Radite Aranda, Dr. rer. Silv. Muhammad Ali Imron, S.Hut., M.Sc. ; Prof. Dr. Hiroto Enari

2026 | Skripsi | KEHUTANAN

Winter snow restricts movement and forage accessibility for sympatric ungulates, potentially altering shared habitat use. This study compared predicted winter habitat suitability and spatial niche overlap between Japanese serow (Capricornis crispus) and sika deer (Cervus nippon) in Shonai and Minamiaizu, Japan. Presence-only records from snow-tracking surveys were analyzed using MaxEnt and seven environmental predictors at 30 m resolution: elevation, slope, snow depth, distances to roads and rivers, aspect, and land cover. Aspect and land cover were categorical predictors. Continuous predictors were screened using Spearman correlations and variance inflation factors. Models were evaluated using five-fold cross-validation, and Schoener’s D was calculated from normalized mean suitability surfaces.

Mean test AUC ranged from 0.936 to 0.959. Distance to roads contributed most to the serow models in Shonai (53.7%) and Minamiaizu (69.4%) and the deer model in Shonai (48.9%), whereas elevation dominated the deer model in Minamiaizu (39.3%). Marginal predicted suitability declined with increasing distance from roads in all models. Snow depth contributed more to the serow model in Shonai than in Minamiaizu (27.1% versus 3.0%). Spatial overlap was greater in Shonai (D = 0.671) than in Minamiaizu (D = 0.441).

These results reveal regional differences in modeled habitat relationships and greater spatial differentiation in Minamiaizu. They inform hypotheses about winter habitat sharing but do not establish interspecific competition as the cause of these differences 

Winter snow restricts movement and forage accessibility for sympatric ungulates, potentially altering shared habitat use. This study compared predicted winter habitat suitability and spatial niche overlap between Japanese serow (Capricornis crispus) and sika deer (Cervus nippon) in Shonai and Minamiaizu, Japan. Presence-only records from snow-tracking surveys were analyzed using MaxEnt and seven environmental predictors at 30 m resolution: elevation, slope, snow depth, distances to roads and rivers, aspect, and land cover. Aspect and land cover were categorical predictors. Continuous predictors were screened using Spearman correlations and variance inflation factors. Models were evaluated using five-fold cross-validation, and Schoener’s D was calculated from normalized mean suitability surfaces.

Mean test AUC ranged from 0.936 to 0.959. Distance to roads contributed most to the serow models in Shonai (53.7%) and Minamiaizu (69.4%) and the deer model in Shonai (48.9%), whereas elevation dominated the deer model in Minamiaizu (39.3%). Marginal predicted suitability declined with increasing distance from roads in all models. Snow depth contributed more to the serow model in Shonai than in Minamiaizu (27.1% versus 3.0%). Spatial overlap was greater in Shonai (D = 0.671) than in Minamiaizu (D = 0.441).

These results reveal regional differences in modeled habitat relationships and greater spatial differentiation in Minamiaizu. They inform hypotheses about winter habitat sharing but do not establish interspecific competition as the cause of these differences 

Kata Kunci : Japanese serow, sika deer, winter habitat suitability, niche overlap, MaxEnt