Assessing soil particle-size distribution on experimental plots with similar texture under different management systems using multifractal parameters

Soil particle-size distribution (PSD) is a fundamental soil physical attribute with dominant influence on many other soil properties. Laser diffraction combined with multifractal analyses have proven to be useful to obtain precise information from PSDs. The aim of this work was to assess similitude or difference of PSDs sampled on plots of an experimental field and belonging to the same textural class using multifractal parameters. The field experiment consisted of two tillage treatments and two cropping systems. It was conducted following a randomized complete split-block design with four replications on a Humic Dystrudept. Tillage treatments were conventional tillage (CT) and no tillage (NT) while crop rotations were ryegrass–sorghum (RS) and ryegrass–corn (RC). Particle-size distribution analysis by the sieve–pipette and by laser diffraction corroborate that all the samples were assigned to the same textural class. Singularity spectra f(α) and Rényi spectra, Dq, showed that multifractal distribution was a suitable model for PSDs obtained by laser diffraction. However, in the range of moments − 10 < q < 10, the r2 values for the linear fits leading to a Rényi spectrum, Dq, were higher than those for the singularity spectrum, suggesting the former was better defined than the latter. No significant differences in multifractal parameters were found between plots with contrasted crop rotation, RS and RC. In contrast, Hölder exponent of order zero (α0) and several parameters derived from the left branch of both, the f(α) and the Dq spectra, were significantly different between CT and NT treatments. No effects of mixing by cultivation were detected in our work, so that differences in PSDs between no-tilled and conventionally-tilled plots were simply attributed to patchiness and variation on the experimental field. Multifractal analysis of PSDs measured by laser diffraction provides further insight in verifying patterns of between plot soil texture variations (i.e. randomness or trends) in completely randomized block designs