Spatially Predictive Habitat Modeling of a White Stork (Ciconia Ciconia) Population in Former East Prussia in 1939

ABSTRACT

Historic information is often crucial for assessing changes and drivers for wildlife and habitat changes although it is often plagued with statistically poor quality. Here we developed three habitat models on two different scales for 1939 for the white stork (Ciconia ciconia) in the region of former East Prussia. We used a geographical information system and a statistical modeling algorithm that comes from the disciplines of machine-learning and data mining (TreeNet). The occurrence of white stork nesting grounds is mainly defined by the variables - distance to forest - , - distance to/density of settlement - , - distance to pasture - and - distance to coastline - . The models present for the first time a quantitative predictive distribution estimate for East Prussia. They are a sound foundation but could be further improved by more data regarding the structure of the habitat and more exact spatially explicit information on the location of white stork nesting sites.