Estimation and predicting of soil water characteristic curve using the support vector machine method
Wenjing Qin,Guisheng Fan,Hongxing-Li
DOI: https://doi.org/10.1007/s12145-022-00909-1
2022-12-28
Earth Science Informatics
Abstract:In this paper, we measured, in low suction region, the water-retention curves of undisturbed loess soil samples covering different soil textures and obtained the parameters of classical van Genuchten soil–water equation by using the HYPROP instrument. Meanwhile, the basic physical and chemical properties (soil particle size distribution, bulk density, organic matter content) of tested soil samples were measured by using corresponding experimental methods. Based on this database, we established a model for predicting the parameters of vG water-retention curve in the low water potential range through an improved Support Vector Machine (SVM) method by the Particle Swarm Optimization (PSO) algorithms. The results show that the values predicted by using the established SVM model here (with input variables: clay, silt, bulk density, organic matter content, electricity conductivity and pH) are highly consistent with the experimental data according to the statistical analysis. Specifically, the obtained absolute, relative and root-mean-square errors between predicted and experimental values are less than 0.0153, 0.1071, and 0.0633, respectively. The research results provide a new method for accurately obtaining soil moisture characteristic curves and provide basic data for hydrological and water quality model.
geosciences, multidisciplinary,computer science, interdisciplinary applications