多变量时间剩余协克立格模型理论及其在水文数据分析中的应用

A THEORY OF MULTIVARIATE TIME-RESIDUAL COKRIGING MODE AND ITS APPLICATION TO HYDROLOGICAL DATA ANALYSIS

本文将随机水文学和多变量地质统计学相结合 ,提出了多变量时间剩余协克立格模型。该模型将水文变量的周期性成分和趋势性成分视为确定性分量从原始各测站水文时间序列中剔除 ,剔除剩下的剩余值属于时空随机函数 ,从而以这种剩余值作为研究变量去刻划原水文变量的时空相关特征。该模型可用于水文变量的插补延展和短期预报 。

A multivariate time-residual cokriging model was presented in this paper by means of combining Stochastic Hydrology with Multivariate Geostatistics.The periodic and trend component of hydrology variable were treaded as determinacy variables.Spatial and temporal correlation characters of original hydrology variable were described by random function which are defference between hydrology time series at hydrology stations and their periodic and trend components.The model is easy to insert and extend the hydrology variables and to take short period forecastion with high accuracy.