土壤入渗模型参数的多元线性预测模型精度的对比分析  

Comparison of Multivariate Linear Prediction Model Precision of Soil Infiltration Model Parameters

作　　者：岳海晶[1] 樊贵盛[1] 

机构地区：[1]太原理工大学水利科学与工程学院,太原030024

出　　处：《灌溉排水学报》2016年第10期73-77,96,共6页Journal of Irrigation and Drainage

基　　金：国家自然科学基金项目(40671081);山西省科技攻关项目(20100311124);文峪河水库高效调度运行方案研究项目(2010-2012)

摘　　要：基于黄土高原大田耕作土壤的入渗试验数据,利用多元线性回归法建立了Kostiakov二参数、三参数以及Philip入渗模型参数的线性预测模型,进行了3种模型参数平均误差的比较以及给定时间的累积入渗量误差的比较,提出了便于应用又具有较高精度的土壤水分入渗参数多元线性预报模型。结果表明,Kostiakov二参数模型的平均误差低于三参数入渗模型和Philip入渗模型,能将平均误差控制在15%以下,低于其他2种入渗模型。因此用Kostiakov二参数入渗模型对土壤水分入渗能力进行预测较好。Based on the soil infiltration experiment sample data of the Loess Plateau in field, the linear prediction model of soil infiltration model parameters in Kostiakov of two parameters, three parameters and Philip were es- tablished by the multivariate linear regression method, the average error of three model parameters and the error of cumulative infiltration amount at a given time were compared, then the multivariate linear forecasting model of soil moisture infiltration parameters which were easy to apply and with high precision was put forward. The re- suits indicated that the average error of Kostiakov infiltration model of two parameters was below 15%, which was lower than those of other two models; the error of cumulative infiltration amount at a given time of two parameters model was also within 15%, which were lower than those of other two models. Kostiakov infiltration model of two parameters was better to predict soil moisture infiltration capacity.

关 键 词：Kostiakov入渗模型 Philip入渗模型 线性预测模型 误差分析 土壤理化参数 

分 类 号：S152.7[农业科学—土壤学]