Do boundaries of soil animal and plant communities coincide? A case study of a Mediterranean forest in Russia

Abstract We studied the relationship between plant and soil animal communities by geostatistical analysis in a piedmont forest close to Novorossiysk (Southern Russia). Vegetation on the slope of a hill was an oak-ash-hornbeam forest, while the vegetation on the foot of the hill was a maple-ash-hornbeam forest. Two plots were studied each including both slope and foot habitats. On every plot samples collected formed a grid of 10 × 5 units with a 5 m distance between them. Soil macroinvertebrates were hand-sorted from the samples, and several soil parameters (soil, pebble, and litter mass, soil moisture) were measured. The analysis did not reveal coincidence between the boundaries of plant and soil animal communities on the bend of the hill. Soil animal communities of the plots were dominated by woodlice, diplopods, and insect larvae, reaching an abundance of 680–990 individuals m − 2 throughout the plots. Number of taxonomic groups per sample and overall animal abundance in the bend were the highest in both plots, whilst these parameters on the slope were the lowest. Variograms and maps of spatial distribution indicated that the boundary between soil animal communities was situated further up on the slope than the vegetation boundary. The size of the animal community was smaller than the size of plots sampled, what probably explained the lack of coincidence between the boundaries. There was a significant correlation between distribution of litter mass and parameters of soil animal communities, which was modulated by depth of soil layer and soil moisture. Soil parameters were more important for explaining boundaries between soil animal communities than plant communities in the forest considered.