改进的克里金法在海面风场计算中的应用

Application of improved Kriging on calculation of sea wind field

克里金法是对空间分布的离散数据的一种线性、无偏、最优的插值法。变异函数的好坏对克里金法的效果有重要影响。用最小二乘法求变异函数通常不是最佳选择,而根据经验观察又缺乏说服力。为解决这一问题,文章提出一种用改进的最小二乘法来确定变异函数,并用该方法对东海部分海域的风场进行克里金插值。通过交叉验证法检验,该方法误差较改进前明显减少,且排除了人为影响。

Kriging method is an agonic, linear and best estimation for discrete spatial data. The variogram has a great impact on the effect of Kriging method. The least squares method is usually not the best way to obtain variogram, which is not convincing only based on the experience. To deal with this problem, this paper gives an improved least square method to obtain variogram. The improved method is used to conduct the sea wind field in part of the East China Sea. The cross validation indicates that the error is less than before without human influence.