Drivers of earthworm incidence and abundance across European forests

• Factors explaining earthworm presence were studied in 6 European forest landscapes.
• The importance of explanatory factors differed among earthworm functional group models.
• Climate was a poor factor in explaining earthworm presence in European forests.
• Litter- and soil-related predictors best explained earthworm presence.

Earthworms have a significant influence on the structure, composition and functioning of forest ecosystems, but in spite of their role as ecosystem engineers, little is known on the factors controlling their distribution across European forests. Optimised sampling techniques, as well as more advanced statistical tools and geographical information systems have facilitated studies at the landscape scale. But these, and even larger-scale studies, are scarce due to data limitations, taxonomic inconsistencies and practical issues in linking existing databases. In this continental-scale field-based study we used boosted regression tree modelling to identify and evaluate the relative importance of environmental factors explaining earthworm incidence (presence/absence) and abundance (density and biomass) in European forests. To parameterise our models earthworms were sampled in six forest landscapes along a latitudinal gradient from the boreal north to the Mediterranean south in spring or autumn of 2012, together with several environmental variables. Earthworms were sampled using a combined method of mustard extraction and hand sorting of litter and a soil monolith, after which they were weighed and identified to functional group (epigeic, endogeic and anecic). We found that litter- and soil-related variables best explained earthworm incidence and biomass in European forests, leaving only a minor role to climate-related variables. Among the litter related variables, understory vegetation played an important role in explaining earthworm incidence and abundance. The relative importance of explanatory variables differed between models for incidence, density and biomass and between earthworm functional groups. Our results suggested that threshold values for soil C:N ratio, forest floor pH and understory plant biomass and plant nutrient concentrations have to be attained before earthworms can occur. Beyond these threshold values, variables like soil C:N ratio, tree litter C:P ratio and forest floor mass further explain earthworm biomass. Mechanisms behind these observations are discussed in the light of future earthworm distribution modelling at continental scale.