6533b853fe1ef96bd12ad619

RESEARCH PRODUCT

ESTIMATING SOIL PARTICLE-SIZE DISTRIBUTION FOR SICILIAN SOILS

Vito FerroGiuseppe GiordanoVincenzo Bagarello

subject

sicilian soilsParticle-size distributionMean squared errorSoil textureMechanical Engineeringlcsh:SBioengineeringSoil classificationSoil sciencelcsh:S1-972Industrial and Manufacturing EngineeringSoil gradationlcsh:AgricultureLoamParticle-size distributionSoil waterRange (statistics)Settore AGR/08 - Idraulica Agraria E Sistemazioni Idraulico-ForestaliParticle-size distribution Particle-size distribution models Soil physical properties.lcsh:Agriculture (General)Mathematics

description

The soil particle-size distribution (PSD) is commonly used for soil classification and for estimating soil behavior. An accurate mathematical representation of the PSD is required to estimate soil hydraulic properties and to compare texture measurements from different classification systems. The objective of this study was to evaluate the ability of the Haverkamp and Parlange (HP) and Fredlund et al. (F) PSD models to fit 243 measured PSDs from a wide range of 38 005_Bagarello(547)_33 18-11-2009 11:55 Pagina 38 soil textures in Sicily and to test the effect of the number of measured particle diameters on the fitting of the theoretical PSD. For each soil textural class, the best fitting performance, established using three statistical indices (MXE, ME, RMSE), was obtained for the F model with three fitting parameters. In particular, this model performed better in the fine-textured soils than the coarse-textured ones but a good performance (i.e., RMSE < 0.03) was detected for the majority of the investigated soil textural classes, i.e. clay, silty-clay, silty-clay-loam, silt-loam, clay-loam, loamy-sand, and loam classes. Decreasing the number of measured data pairs from 14 to eight determined a worse fitting of the theoretical distribution to the measured one. It was concluded that the F model with three fitting parameters has a wide applicability for Sicilian soils and that the comparison of different PSD investigations can be affected by the number of measured data pairs.

https://doi.org/10.4081/jae.2009.3.33