A model for characterizing dry soil aggregate size distribution

Dry soil aggregate size distribution (DASD) is an important parameter in evaluating soil management practices or modeling of wind erosion and dust emissions. Several models including Lognormal, Fractal and Weibull distributions have been developed to quantitatively describe the DASD. In this study, a new model combining power-law and exponential distributions was proposed to characterize DASD. The performances of the Fractal, Weibull and modified Lognormal distributions and the new model were investigated using 253 DASD data from published documents across five countries. The new model best described the observed DASD data across different soil texture, land use and sieving cuttings. The patterns of DASD generally governed the goodness of fit of these DASD models. The Fractal and Weibull distributions could not well depict the right-skewing unimodal and multimodal DASD data. The modified Lognormal distribution could not well describe the multimodal DASD data. The larger aggregate size distribution was better dictated by the power-law distribution, whereas the smaller aggregate size distribution was characterized by the power-law and exponential distributions for the new model. The choice of screen opening for sieving procedure could affect the patterns of the apparent DASD, and further influence the accuracy of the DASD models. The d(r), a parameter in the new model, was significantly related to rock fragments, soil aggregate stability and wind erosion rate. More studies are required to investigate the relationship between the parameters of the new model and soil properties linked with wind erosion.