Pedo-transfer Functions Of Soil Hydraulic Parameters In Black Soil Area Of Songnen Plain

The black soil area of the Songnen Plain is a typical black soil concentration area in the northeastern part of China.It is an important commodity grain production base and agricultural base.This paper takes the southern part of the black soil area of the Songnen Plain as the research area.The relationship between soil water characteristic curve and saturated hydraulic conductivity and soil physical and chemical properties was analyzed,and soil water characteristic curve and saturated hydraulic conductivity Pedo-transfer functions were constructed respectively.The research results can provide data support for the development of water-saving agriculture,alleviating the shortage of agricultural water,realizing agricultural irrigation,regional water resources management and soil water in the black soil region.Because direct determination of soil water characteristic curve and saturated hydraulic conductivity is labor-intensive,difficult and time-consuming,indirect estimation of soil physical and chemical properties by easy measurement is relatively simple.Therefore,based on the parameters of van Genuchten model,saturated hydraulic conductivity and soil physical and chemical properties,109 groups of experimental samples in different soil layers,the Pedo-transfer functions of the parameters of van Genuchten model parameters and saturated hydraulic conductivity were constructed by nonlinear programming method and grey BP neural network method.The main conclusions were as follows:(1)Four soil water characteristic curves models of VG,BC,MG and LND were selected to simulate the soil water characteristics curves of five soil textures in the study area.The predicted and measured values of the empirical models of the four soil water characteristics curves were compared.The results showed that the VG and MG models had higher accuracy in simulating soil water characteristics in the study area.The VG model was selected as the basic model for constructing the soil water characteristic curve of the study area.(2)The particle swarm optimization algorithm was used to fit the parameters of van Genuchten model,and the relationship between van Genuchten model parameters and soil physical and chemical properties was analyzed by grey correlation.The results showed that the parameter θr of van Genuchten model in 0~20cm soil layer had the highest correlation with fractal dimension,dry bulk density and organic matter content,and the correlation between parameter n and fractal dimension,dry bulk density and clay content was the highest.The parameter α had the highest correlation with sand content,dry bulk density and fractal dimension.The parameter θr of van Genuchten model in 20~40cm soil layer had the highest correlation with sand content,dry bulk density and organic matter content,parameter n and fractal.The correlation between dimension,organic matter content and particle content was the highest,and the correlation between parameter α and sand content,fractal dimension and organic matter content was the highest.(3)Using soil water characteristic curve Pedo-transfer functions was constructed by nonlinear programming method and BP neural network and other regional soil transfer functions predict soil water characteristics in the study area.The results showed that the prediction accuracy of other regional Pedo-transfer functions is low,which was not suitable for the prediction of soil water characteristic curve in the study area.However,the predicted values of the two models constructed in this paper were close to the measured values,and the prediction results are better.By comparing and analyzing the two soil water characteristic curve models of nonlinear programming and grey BP neural network,the grey BP neural network model had a slightly smaller prediction error for the van Genuchten parameter than the nonlinear programming model in different soil layers.(4)The grey correlation analysis was used to analyze the relationship between soil physical and chemical properties and measured saturated hydraulic conductivity in different soil layers.The correlation between soil saturated hydraulic conductivity and soil organic matter,dry bulk density and fractal dimension in 0~20 cm and 20~40 cm soil layers was higher,The results showed that the soil saturated hydraulic conductivity was highly correlated with soil organic matter,dry bulk density and fractal dimension in 0~20 cm soil layer and 20~40 cm soil layer.(5)Using nonlinear programming and grey BP neural network to construct Pedo-transfer functions was more accurate than other regions transfer function.by comparing and analyzing the two kinds of saturated hydraulic conductivity models of nonlinear programming and grey BP neural network,the average error of different soil layers in both models is less than 10%,and the correlation coefficient is above 0.85.The mean square of gray BP neural network model The root error is smaller than the nonlinear programming model.The grey BP neural network model is slightly better than the nonlinear programming model.In this paper,the southern region of black soil area in Songnen Plain is taken as the research object.The relationship between soil water characteristic curve,saturated hydraulic conductivity and measured data of soil fractal dimension,organic matter content,dry bulk density,soil texture(clay content,silt content,sand content)in different soil layers was analyzed by grey correlation analysis.By using the method of non-linear programming and grey BP neural network,the suitable Pedo-transfer functions of soil water characteristic curve and saturated hydraulic conductivity were constructed in different soil layers of the black soil area of Songnen Plain,so as to achieve the purpose of predicting the soil hydraulic parameters of different soil layers in the study area.The research results can provide parameters for soil water and solute transport and soil water prediction in the study area,and also provide technical support for acquisition of regional soil hydraulic properties.