基于HYDRUS-2D的土壤水力特征参数反演及其时空变异性研究

Soil hydraulic parameters inversion based on HYDRUS-2D model and their spatial and temporal variability

为研究田间尺度灌溉条件下土壤水力特征参数的时空变异性,基于甘肃省武威市石羊河试验站葡萄园内土壤含水率数据,采用HYDRUS-2D模型反演估计不同生育期不同深度的土壤水力特征参数,评估反演参数的土壤含水率模拟效果,分析其时空变异性。结果表明:田间尺度下采用反演参数模拟不同深度土壤含水率的变化的效果优于同种情况下Rosetta模型预测参数的模拟效果;反演参数在果实膨大期和转色期的模拟效果好于其它生育期,反演的拟合优度R2值由Rosetta模型的0.4~0.5提高到了0.7~0.8,均方根误差RMSE和平均绝对误差MAE值由Rosetta模型的0.04~0.05 cm^(3)/cm^(3)和0.03~0.04 cm^(3)/cm^(3)分别降低到了0.03~0.04 cm^(3)/cm^(3)和0.02~0.03 cm^(3)/cm^(3);此外,反演出的土壤水力学特征参数只适用于参与反演过程类似条件下的模拟,应用于有明显差异的情景时,模拟精度显著降低;反演得出的3个参数土壤饱和导水率Ks、土壤水吸力方程参数α、土壤孔隙大小分布系数n均表现出了一定的时空变异性,其中Ks和n的时空变异性小,α在0~10 m-1,时空变异性大。本研究可为准确预测农田土壤含水率的变化,制定合理的灌溉制度,提高水分利用效率提供科学依据。

To investigate the spatial and temporal variability of soil hydraulic characteristics parameters under field-scale irrigation conditions,HYDRUS-2D model was used to estimate soil hydraulic parameters at different depths at different growth stages based on soil water content data in the vineyard of Shiyanghe Experimental Station in Wuwei,Gansu Province.The effect of soil water content simulation by inversion parameters was evaluated and its spatial and temporal variability were analyzed.The results showed that:The inversion parameters simulated the variation of soil water content at different depths at the field scale were better than those predicted by the Rosetta model under the same conditions.The values of RMSE and MAE decreased from 0.04-0.05 cm^(3)/cm^(3) and 0.03-0.04 cm^(3)/cm^(3) in the Rosetta model to 0.03-0.04 cm^(3)/cm^(3) and 0.02-0.03 cm^(3)/cm^(3),respectively;Moreover,the soil hydraulics characteristics of the inversions were only applicable to simulations under similar conditions to those involved in the inversion process,and the accuracy of the simulations was significantly reduced when applied to scenarios with different conditions.The three parameters derived from inversion,which were soil saturation hydraulic conductivity(Ks),soil water suction equation parameter(α),and soil pore size distribution coefficient(n),all showed some degree of spatial and temporal variability.Ks and n showed small spatial and temporal variability,andαranging from 0 to 10 m-1 showed large spatial and temporal variability.In conclusion,this study can provide a scientific basis for accurately predicting changes in soil water content in agricultural fields developing reasonable irrigation systems,and improving water use efficiency.