In order to improve the existing phosphorus index assessment methods,using the interactive evaluation index(IEI)as an auxiliary variable,the geographically weighted regression(GWR)was adopted as prediction means.A met...In order to improve the existing phosphorus index assessment methods,using the interactive evaluation index(IEI)as an auxiliary variable,the geographically weighted regression(GWR)was adopted as prediction means.A method of regional soil phosphorus risk assessment was constructed by modifying phosphorus index model(MPIM).The GWR-IEI method more accurately predicted available phosphorus(AP)and soil organic matter(SOM),and the prediction precision and goodness of fit were high.Compared with the ordinary least square(OLS)method,the relative improvement of the root mean squared errors(RMSE)with the GWR-IEI method reached 28.95%for available phosphorus predicted,while that of SOM was 21.24%.The phosphorus loss risk of most of the study area(95.29%)was moderate to low.The areas featuring an extremely high phosphorus loss accounted for merely 0.33%of the total research area.Phosphorus loss depends on the effects of many factors.Areas which have strong source or transfer factors are not necessarily high-risk areas for phosphorus loss.Only the co-occurrence of transfer and source factors leads to high risk and greater potential for phosphorus loss.The GWR-IEI-MPIM method accurately reflected the degree of risk for phosphorus at the regional scale,which provides a valuable reference for risk assessment of phosphorus.展开更多
基金This research was funded by the National Key Research and Development Program of China(2016YFD0300801,2020YFC1908601)the National Natural Science Foundation of China(41471186).
文摘In order to improve the existing phosphorus index assessment methods,using the interactive evaluation index(IEI)as an auxiliary variable,the geographically weighted regression(GWR)was adopted as prediction means.A method of regional soil phosphorus risk assessment was constructed by modifying phosphorus index model(MPIM).The GWR-IEI method more accurately predicted available phosphorus(AP)and soil organic matter(SOM),and the prediction precision and goodness of fit were high.Compared with the ordinary least square(OLS)method,the relative improvement of the root mean squared errors(RMSE)with the GWR-IEI method reached 28.95%for available phosphorus predicted,while that of SOM was 21.24%.The phosphorus loss risk of most of the study area(95.29%)was moderate to low.The areas featuring an extremely high phosphorus loss accounted for merely 0.33%of the total research area.Phosphorus loss depends on the effects of many factors.Areas which have strong source or transfer factors are not necessarily high-risk areas for phosphorus loss.Only the co-occurrence of transfer and source factors leads to high risk and greater potential for phosphorus loss.The GWR-IEI-MPIM method accurately reflected the degree of risk for phosphorus at the regional scale,which provides a valuable reference for risk assessment of phosphorus.
