Estimation of Soil Water Characteristics with Arya and Paris Model

• Main Authors: Hsin-Hung Chen, 陳信宏
• Other Authors: Li-Ling Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/39329967536369480651

碩士 === 國立中興大學 === 水土保持學系所 === 94 === The traditional and frequently used research methods of measuring soil water characteristics, usually take much time and have complicated experimental procedures; this may restrict the number of samples for study. In recent years, in the field new monitoring technologies have been developed and new experimental instruments have been used now to measure and record soil water characteristics effectively. There is less range of error than in the past. Nevertheless they are still time consuming as the research procedures are complex. Therefore, some scholars suggested simulating and estimating the soil water characteristics by indirect methods and models. These are generally called “pedotransfer functions (PTFs).” In this study, the Arya and Paris Model (which is a PTF) was used to estimate soil water characteristics from soil particle size distribution, bulk density, and particle density. This model needs detailed particle size distribution data as the basis for estimating soil water characteristics. Such particle size distribution data is not easy to obtain. For this reason, this research has simplified and made use of particle size distribution data available directly from the unsaturated soil hydraulic databases (UNSODA) to estimate soil water characteristics, and to compare the results with measured data from databases, in order to compare and evaluate their reliability. Soils with five kinds of textures from coarse to fine were selected as study materials. These were sand, sandy loam, loam, silt loam, and clay. The chosen criteria for the soil samples from the database had to correspond to soil properties needed in the model. For all samples, in the first and the second treatment, the parameter values in the model are 1.38 and 0.938. However, in the third treatment, the parameter values were set at 1.285, 1.459, 1.375, 1.15, and 1.16 respectively, according to the textures of the five kinds of soils. The essential soil data were put sequentially into the equations of the model, in order to compute the soil water content and the matric potential, then curve fitting technique was applied to establish the soil water characteristic curve equation. Assuming measured matric potential from database as the same conditions, estimated water content was compared with measured water content; then the estimating performance, and capability of model for different soil textures was discussed. The results indicated that the model performed comparatively well for sandy loam, loam, and clay. But for sand and silt loam the results were less good. Consequently, this model is suitable for the soils with medium texture and uniform particle size distribution, but not so suitable for other kinds of soil. This study also demonstrated that it is also inappropriate to use this method when the parameter value is 0.938. With regard to the first and the second treatment, different soil textures produced different soil water estimates.