Knowledge is lacking on the factors controlling the structure and spatial distribution of earthworm communities within agroecosystems. We first hypothesized that forested riparian buffer strips (FRBS) within agricultural landscapes would be a refuge for earthworms, as we predicted higher soil moisture and organic matter inputs in FRBS than in adjacent agricultural fields (treatments = FRBS vs.
Field). We further hypothesized that earthworms would be most abundant where the chemical quality of above- and belowground plant litter is high, or where soil disturbance is low.
We conducted a field survey to quantify earthworm species abundances in FRBS and adjacent agricultural fields in two bioregions, namely Eastern Canada and Central Europe. At each of 77 sites, we collected and identified earthworms from three plots within FRBS and within adjacent agricultural fields.
In each plot, we identified the tree species, understory vegetation, soil drainage class, agricultural crop as well as five soil physicochemical properties. In each bioregion and treatment, we found proportionately more endogeic than anecic or epigeic earthworm species.
In Eastern Canada, there were proportionately fewer anecic and more epigeic individuals in FRBS than in fields; in Central Europe there were fewer endogeic and more anecic earthworms in FRBS than in fields. We found significant bioregion x treatment interactions on earthworm abundance and soil moisture.
More specifically, in Eastern Canada we found higher earthworm abundance and soil moisture in FRBS, whereas in Central Europe we found higher earthworm abundance in fields and no treatment effect on soil moisture. In both bioregions, we found higher organic matter in FRBS than in fields.
In Eastern Canada, earthworm abundance in deciduous and mixedwood stands were higher than in coniferous stands; in Central Europe, earthworm abundance was highest in deciduous stands only. Within FRBS in Eastern Canada, the abundance of the prominent endogeic species, Apporectodea rosea, was correlated with herbaceous plants, notably ferns and graminoids.
Conditional regression tree analysis revealed positive relationships between earthworms and soil clay content, pH, moisture and organic matter. Our results suggest that local and landscape patterns in earthworm diversity can be predicted by soil and vegetation attributes, however the relative importance of these factors changes across continental scales due to differences in precipitation patterns and soil moisture availability.
Comparing the distributions of earthworms across different scales provides insights into the potential of different species to spread into new habitats with climate change.