Сибирский экологический журнал, 5 (2016) 625639 УДК 630*907.13 DOI 10.15372/SEJ20160501 Modelling the Spatial Distribution of Wildlife Animals Using Presence and Absence Data H.-S. KWON1, B.-J. KIM1, G.-S. JANG2 1 Nationаl Institute of Ecology Seocheon, Chungnаm 33657, South Korea 2Department of Life Sciences, Yeungnаm University Gyeongsan, Gyeongbuk 38541, South Korea E-mail: sunside@ynu.ac.kr. <...> Статья поступила 24.12.15 Принята к печати 30.12.15 ABSTRACT This study was conducted to analyze the habitat preference of six major mammals for various environmental factors based on 100 random points within a mountain area in South Korea. <...> In-situ presence and absence data for the mammals were surveyed and collected, and twelve explanatory variables related to topography, water, greenness, and anthropogenic influence were applied to create a habitat distribution model. <...> The best combination of variables was determined using Morans I coefficients and Akaike criteria information, and applied to estimate the habitat preference for each species using GRASP v.3.0. <...> The predictive map showed that wildlife animals in this study were mainly categorized into two groups: Group I (Korean squirrel, Sciurus vulgaris, mole, Talpa micrura and water deer, Hydropotes inermis), showed equal preference for all mountainous areas; Group II (weasel, Mustela sibirica, leopard cat, Felis bengalensis and raccoon dog, Nyctereutes procyonoides) showed different preferences in a mountain. <...> Key words: Species distribution model, General additive model, GRASP, Habitat preference. <...> Interest in species distribution models (SDMs) of plants and animals has been growing dramatically and they have become increasingly important tools to address various issues in ecology, biogeography, evolution and climate change since the early 1990s [Guisan, Thuiller, 2005]. <...> More recently, prediction of species distributions through field-surveying data has long been recognized as a significant component of conservation planning [Austin, 2002; Elith, Burgman, 2003], and a wide variety of statistical and machine-learning methods have © Kwon H.-S., Kim B.-J., Jang G.-S., 2016 625 been introduced to obtain the best predictions, often in conjunction <...>