We evaluated the relative importance of tree species diversity and functional group for various ecosystem functions in major European forest types in the six regions Finland, Poland, Germany, Romania, Italy and Spain. We sampled soils in 209 permanent plots along a species diversity gradient from monocultures to 5-species mixtures. This dataset contains soil texture information from 0-10 cm layers of the mineral soil in the FunDivEUROPE Exploratory Platform. The texture analysis by the laser diffraction method was carried out on the same samples used for analysis of soil pH, C and N concentrations and stocks reported in Dawud et al. (2017), Funct. Ecol. 31, 1153–1162, doi: 10.1111/1365-2435.12821.

In each of the 209 pots, we subsampled we sampled one mineral soil core in 9 subplots. The samples were divided into 0–10 and 10–20 cm layers and composited by layer into one sample per core plot. We sampled the mineral soil with a 3.6 cm diameter soil corer. The Spanish region was only sampled to 10 cm depth due to stoniness. Only 0-10 cm samples were analysed for soil texture.

Our study was part of the Pan-European project, FunDivEUROPE (http://www.fundiveurope.eu), which investigated the importance of tree species diversity on forest ecosystem functioning (Baeten et al. 2013, doi:10.1016/j.ppees.2013.07.002). Mature forest stands were selected in six regions (Finland/North Karelia, Poland/Białowieźa, Germany/Hainich, Italy/Colline Metallifere, Romania/Râsca and Spain/Alto Tajo Natural park) representing major European forest types. In each region, plot selection procedures were applied with a set of predefined criteria on evenness, admixture, age distribution, tree density and site-related factors such topography, soil type, previous land-use and management. The plots were particularly selected to minimize effects of site-related variability. More detailed site descriptions and selection criteria are reported by Baeten et al. (2013). A diversity gradient based on stands of one to three, four or five tree species was selected within each region, depending on the local species pool (Baeten et al. 2013). Each core plot of 30 by 30 m was divided into nine 10 by 10 meter subplots in which the soil sampling took place.