Method Study On Determining Soil Infiltration Parameters With Pedo-Transfer Functions

The method study on determining soil infiltration parameters with pedo-transfer functions was supported by the Shanxi natural science foundation research. In this paper, based on indoor infiltration tests and field infiltration tests into freezing and thawing soils, the major factors influencing soil infiltration characteristics and the mathematic relation between those and the infiltration capacity were studied systematically. The Models of soil pedo-transfer functions between soil infiltration capacity and soil infiltration model parameters and soil conventional physicochemical parameter were established, with the help of the means of multivariate linear, multivariate nonlinear and BP model. By means of statistic analysis method, those models were tested and verified. The methods that infiltration capacity and infiltration model parameters were determined with the conventional parameters which reflect field soil physicochemical properties.The studied results show that Soil texture, soil structure, soil moisture content and organic content were the major influence factors on unfreezing soil infiltration characteristics, and that because in period of freezing and thawing, ground temperature is the deterministic factor which results in the phase change of liquid state moisture in soils and determines the degree of the phase change of liquid state moisture, therefore, soil texture, soil structure, soil moisture content, soil temperature and organic content were the major influence factors on the infiltration characteristics of frozen soil.In this paper, the multivariate linear regression model, multivariate nonlinear regression model and BP model predicting the infiltration capacity and infiltration model parameters of soils through the routine soil physics parameters were established respectively. Those models were analyzed and compared, and the result is that the multivariate nonlinear regression model and the BP model are fair to the multivariate linear regression model, and the BP model is fair to the multivariate nonlinear regression model.If the input variables of models are soil moisture content, soil dry bulk density, silt content, sand content and organic content in unfreezing period of soils, and soil moisture content, soil dry bulk density, silt content, sand content, soil temperature and organic content in frozen period of of soils , the predicted results are satisfactory.In view of the increasing tension contradiction between the supply and demand for water resources and the water-saving stratagem of agriculture irrigation that the improvement of surface irrigation technique was the principalest water-saving measure , and the sprinkle irrigation and micro irrigation should be developed conditionally, In this paper, the key problem—soil infiltration model parameters, which have important influence on the water-saving of surface irrigation was researched profoundly, soil pedo-transfer functions of infiltration model parameters were determined indirectly through soil conventional physicochemical parameters. Those results are of theoretical value for promoting the development of soil infiltration theory, and of practical significance for improving surface irrigation technique and heightening the using efficiency of irrigation water supply.