Seasonality overrides differences in butterfly species composition between natural and anthropogenic forest habitats

Anthropogenic disturbances that alter habitat structures can lead to strong changes in species composition. The importance of certain habitat features for structuring species composition, however, may be determined by the availability of resources that can seasonally fluctuate. The Kenyan Arabuko Sokoke coastal forest consists of three pristine forest types and provides various anthropogenic forest-like habitats, such as plantations and transition zones between forest and agricultural land. In parallel, this region experiences pronounced climatic seasonality which causes differences in resource availability. To test for potential differences in butterfly communities between natural and anthropogenic forest habitats, and whether they vary seasonally, we assessed butterfly species along transects in all five habitat types during the rainy and the consecutive dry season. We compiled ecological characteristics for each transect and for each butterfly species. We found highest species richness and abundances in the three natural forest habitats during the rainy season. Our data show distinct species clusters for each habitat type during the rainy season. In contrast, butterfly species composition does not form distinct community groups during the dry season. During this time period, highest species richness and abundances are found in plantations. These temporal differences in species richness, abundance and species composition between the two seasons are assumed to be driven by limited nectar availability inside of the forest during the dry season. This situation might lead to increased mobility and the invasion of anthropogenic habitats, where nectar sources are still available during the dry season. Thus, plantations might represent a surrogate habitat for forest butterflies during their imaginal stage, but do not represent a suitable habitat providing all resources needed for successful larval development.