Based on the spatial interpolation technique served by GIS,this paper evaluated the prediction accuracies of three interpolation methods,i.e. Ordinary Krig(OK) ,Radical Basis Functions(RBF)and Inverse Distance Weighting (IDW) . Not only do these techniques have the capability of producing a prediction surface,but they can also provide some measure of the certainty or accuracy of the predictions. The results showed that there was a high similarity between Ok and RBF methods at the location of sand beam on which the measurement were conducted. However,there was a big discrepancy at the edge of the interpolation surface. The prediction temperature result of IDW method was lower than that of the OK and RBF methods. Companied with the results of cross-validation and site-validation,we could find that the fitting capacity of these methods follow the sequence of OK > RBF > IDW. Therefore,the OK method was used to produce the spatial patterns of soil temperature over the study area. In addition,the accuracy of the predictions was impacted by several factors,including sample distribution,season variability of soil temperature and so on.%对青海湖湖东克土沙地2010年6月份和8月份从6cm到71cm深度的8个层位土壤地温数据进行了插值方法上的对比研究,定量化评估了普通克里格(OK)、径向基函数法(RBF)以及反距离权重法(IDW)的预测精度.结果表明:OK方法与RBF插值结果在有数据监测点的沙梁位置具有高度一致性,但是在沙梁边缘区域表现出了较大的差异.而IDW插值结果比前两者预测结果要低.结合交叉验证以及站点验证结果,OK插值结果精度最高,RBF方法次之,IDW方法精度最低.因此,OK方法是该研究区内最佳的拟合模型.另外,模型预测精度受到样点分布以及土壤温度的季节变率等诸多因素影响.
展开▼
