岩石圈磁场数据网格化方法研究

选取最小曲率、克里格、改进Shepard、反距离加权和径向基函数等5种网格化数学模型,对小江断裂地磁总强度加密区岩石圈磁场数据进行数据网格化,采用均方根预测误差和插值数据残差均方根等评价指标对网格化结果进行评价,结果表明,克里格插值与反距离加权插值法的精度最高。进一步比较克里格插值与反距离加权插值法的网格化图形质量,结果显示克里格插值网格化过程中兼顾了数据的平滑性和各实测点与待估点之间的空间位置关系,避免了系统误差,得出克里格插值更适用于岩石圈磁场数据网格化的结论。

Study on Gridding method of lithospheric magnetic field data

The paper compared and discussed five commonly used interpolation methods. The interpolation methods were evaluated using the parameters such as the root mean square prediction error and the interpolation data root mean square for the lithospheric magnetic field data in the fault zone of Xiaojiang. The results showed that the Kriging interpolation method and inverse distance interpolation method give the highest accuracy. Further, the grid graphics of the two interpolation methods were evaluated. The result showed that Kriging interpolation takes into account the smoothness of the data and the spatial position relationship between the measured points and the points to be estimated during the Gridding process, avoiding systematic errors. It can be concluded that Kriging interpolation is more suitable for Gridding of lithospheric magnetic field data.