Open AccessA Simple and Improved Model for Describing Soil Hydraulic Properties from Saturation to Oven Dryness
Author(s) -
Liao Kaihua,
Lai Xiaoming,
Zhou Zhiwen,
Zhu Qing,
Han Qing
Publication year - 2018
Publication title -
vadose zone journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.036
H-Index - 81
ISSN - 1539-1663
DOI - 10.2136/vzj2018.04.0082
Subject(s) - hydraulic conductivity , pedotransfer function , soil water , dryness , soil science , mean squared error , environmental science , saturation (graph theory) , water content , hydrology (agriculture) , geotechnical engineering , mathematics , geology , statistics , medicine , surgery , combinatorics
Core Ideas Soil hydraulic properties from saturation to oven dryness are needed to simulate soil water movement. A new soil hydraulic model was compared with the classic van Genuchten–Mualem models. The new model with different fitted parameters was more efficient than the van Genuchten–Mualem models. Soil water retention characteristics and hydraulic conductivity from saturation to oven dryness are needed for simulating soil water movement and solute transport, especially in arid and semiarid regions. The aim of this study was to derive a new soil hydraulic model and compare this model with the classic van Genuchten–Mualem (VGM) models (M1). A total of 12 soils with varying basic properties were selected from the Unsaturated Soil Hydraulic Database (UNSODA) for the evaluation of the new model. Five pairs of soil hydraulic models were considered, including M1 and the new model with different parameters fitted (M2–M5). Correlation analysis was conducted to analyze the relationships between the new model parameters and basic soil properties. In addition, the Morris method was applied to quantify the parameter sensitivity of the new model. Results showed that M2 to M5 gave slightly better performance than M1 in predicting soil water retention characteristics, with a decrease of ∼15% in the root mean squared error (RMSE). However, they produced substantially better performance than M1 in estimating soil hydraulic conductivities, with a decrease of ∼40% in the RMSE. The new model with all of the model parameters being fitted (M2) was more efficient than those with one or two model parameters not fitted. However, an over‐parameterization problem was detected with M2. Physically unrealistic parameters (e.g., the saturated soil water content and effective grain diameter) exist in M2. These parameters were found to have important influences on the results of the soil hydraulic properties prediction based on the Morris‐based sensitivity analysis. Overall, the new model with two parameters not fitted (M5) can best describe the soil hydraulic properties from saturation to oven dryness and had the potential to simulate water and solute transport in the vadose zone.