Performance of soil particle-size distribution models for describing deposited soils adjacent to constructed dams in the China Loess Plateau |
| |
Authors: | Pei Zhao Ming-an Shao Robert Horton |
| |
Institution: | (1) Institute of Soil and Water Conservation, CAS & MWR, Northwest A&F University, Yangling, 712100 Shannxi, China;(2) Golder Associates Ltd, 1721, 8th Street East, Saskatoon, SK, S7H 0T4, Canada;(3) SoilVision Systems Ltd, 202, 640 Broadway Avenue, Saskatoon, SK, S7N 1A9, Canada |
| |
Abstract: | Soil particle-size distributions (PSD) have been used to estimate soil hydraulic properties. Various parametric PSD models
have been proposed to describe the soil PSD from sparse experimental data. It is important to determine which PSD model best
represents specific soils. Fourteen PSD models were examined in order to determine the best model for representing the deposited
soils adjacent to dams in the China Loess Plateau; these were: Skaggs (S-1, S-2, and S-3), fractal (FR), Jaky (J), Lima and
Silva (LS), Morgan (M), Gompertz (G), logarithm (L), exponential (E), log-exponential (LE), Weibull (W), van Genuchten type
(VG) as well as Fredlund (F) models. Four-hundred and eighty samples were obtained from soils deposited in the Liudaogou catchment.
The coefficient of determination (R
2), the Akaike’s information criterion (AIC), and the modified AIC (mAIC) were used. Based upon R
2 and AIC, the three- and four-parameter models were both good at describing the PSDs of deposited soils, and the LE, FR, and
E models were the poorest. However, the mAIC in conjunction with R
2 and AIC results indicated that the W model was optimum for describing PSD of the deposited soils for emphasizing the effect
of parameter number. This analysis was also helpful for finding out which model is the best one. Our results are applicable
to the China Loess Plateau. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|