Modelling and spatial prediction of earthworms ecological-categories distribution reveal their habitat and environmental preferences

Earthworms are one of the important soil animals and have been generally described as soil engineers. Knowledge on environmental conditions driving the distribution and population of this soil animal and the habitat which support these conditions especially at the ecological level is required to understand their responses to these environmental conditions at different habitats so as to guide its usage as bio indicator of soil quality and health. In this study we use RandomForest (RF), a machine learning algorithm to model species distribution, density/abundance based (SDM/SAM) and predict the biodiversity distribution (richness and density, ind.m⁻²) of three basic earthworms ecological categories: epigeic, endogeic and anecic (including the epi-anecic subcategory) across soil and climate variables at multiple habitat type/land uses in Germany. Our study shows there are spatial/ geographic variation in the distribution of the species richness and density among the three earthworms’ ecological categories. Also their environmental and habitat preferences are equally different, while epigeic species are predicted to be climate driven mostly in forests, endogeics are predicted to be the most diverse (in richness and density), but are mostly driven by soil textural contents (clay and silt) and found primarily in arable and grassland. Vineyard and crop flood plain are predicted to be suitable and the preferred habitat for anecics/epi-anecics. This study also identify optimum environmental gradient at which the species density is at the peak in each of the earthworm’s ecological category which would not only provide guide on soil biodiversity monitoring but also the soil health status.