Modelling of Habitat Preferences of water pipits (Anthus spinoletta spinoletta) in "Nationalpark Gesäuse" using Remote Sensing and GIS

In spring 2009 a survey was done on a grid patterned cell basis (100x100m), and data on absence and presence of water pipits (Anthus spinoletta spinoletta) in “Nationalpark Gesause” were collected. From this data we performed a habitat preference model, using logistic regression in a stepwise forward procedure. 64 habitat variables were calculated mainly using biotop data from Nationalpark Gesause, a digital elevetion model (resolution 25X25m), Landsat 5 TM satellite data and data from our bird-survey. We used the satellite data to calculate an index for biomass, the NDVI (Normaliced Differential Vegetation Index). The Nationalpark Gesause was choosen as investigation area because of the presence of very good digital data sources that make it ideal to do such surveys, furthermore the scientific team from Nationalpark Gesause provided great help during preparation and performing the studies as well as organising facilities and contacts. We started the modellbuilding process by calculating univariat models using Nagelkerke R² and choosing the most powerful variables to reduce the number of variables. To avoid multicolliniarity, bivariat spearman rang correllations (Rs) were calculated and if Rs was >0,5 only the variable with the more powerful Nagelkerke R² was chosen for further calculations.