Correlations between weather conditions and common vole (Microtus arvalis) densities identified by regression tree analysis

Population dynamics of fluctuating and cyclic rodent populations can be impacted by particular weather parameters. In temperate areas there are interrelations between different weather parameters, which make identification difficult. However, this is necessary because small rodents are relevant for both the food web and crop damage especially in the face of climate change. We used both, boosted regression tree and classification and regression tree methods to identify weather conditions correlating with the active burrow index (ABI) of common voles (Microtus arvalis) from 1974 to 1998 in the high outbreak risk area of Central Germany. Highest ABI occurred in perennial crops in fall with a maximum of more than 2000 active burrows per 1000 m(2). Boosted regression tree analysis showed that between 12 and 20 weather parameters could have a relative influence on vole ABIs ranging from 2% to 19%. Classification and regression tree analysis highlighted that the number of days with snow cover in December and March, rainfall amount in spring and maximum temperature in October seem to be key indicators for ABIs in the following year in spring. Monthly maximum temperatures of February to June and the amount of precipitation in April and July were correlated to ABIs in fall. Quantitative validation showed an agreement of ABI distribution on regional scale >85%. This represents the first study to identify complex weather conditions including single parameter thresholds correlated with common vole abundance in a temperate area. The results have the potential to aid the development of predictive models for small rodent dynamics and they inspire further detailed search for regulative mechanisms of small mammal dynamics.