Relation between soil forming factors and scaling properties of particle size distributions derived from multifractal analysis in topsoils from Galicia (NW Spain)

- PSDs were measured in 187 topsoils from NW Spain under various soil forming factors.
- PSDs behaved as multifractals, with scaling properties varying in different samples.
- PLS showed the entropy dimension D1 as the index better related to forming factors.
- Parent material and climate were the most influential in scaling properties of PSDs.
- Values of D1 could be related to the degree of evolution of the soils.

Soil particle size distribution has been previously shown to behave as a multifractal. In this study, we analyzed the relation between soil forming factors and several multifractal parameters derived from generalized dimension and singularity spectra calculated from particle size distribution of soils from Galicia (NW Spain). A number of 187 samples were collected from the upper horizon (0-30 cm depth) of soils located in different parts of the region, covering a wide range of parent materials, climatic conditions and land use. Particle size distributions of fine earth fractions (< 2 mm) were measured using a laser particle analyzer that yielded a number of data points per sample between 85 and 161. Rényi and singularity spectra were calculated for each sample and parameters Dmax, Dmin, Dmax-Dmin, D0, D1, D2, and D1/D0 of the Rényi spectra, and αmax, αmin, α0, f(αmax), f(αmin), αmax-αmin, αmax-α0, and α0-αmin of the singularity spectra were extracted. The spatial distribution of the multifractal indices and their relation with soil forming factors were predicted using a PLS approach with climate, geology, land use and topography as covariates. The accuracy of the estimates was evaluated by cross-validation upon the same set of samples. The performance of the model was relatively good in the case of D1 (R2 = 0.50, RMSE = 0.02). As a result, the heterogeneity of particle sizes in topsoils from Galicia can be partly explained by a combination of climatic factors, land use and geology. Parent material and some climatic indexes were the most influential variables explaining the scaling properties of particle size distributions.