基于Hurst系数的流域降雨时空变异分析方法

A Spatial and Temporal Variation Analysis Method of Watershed Rainfall Based on Hurst Coefficient

基于Hurst系数能够反映水文序列长期相关性的原理,结合R/S分析方法和分数布朗运动理论,提出一种从整体上识别与检验时间序列变异及其变异程度的分析方法。该法将变异程度划分为4个等级:无变异(或弱变异)、中变异、强变异、巨变异;结合GIS空间分析方法,进而建立降雨量的时空变异分布图。应用该法对无定河流域年降雨序列进行时空变异分析,结果表明无定河流域大部分面积上的年降雨序列处于无变异或弱变异状态,但也有部分地区如高镇一带处于中变异状态,而且年降雨量有减少的趋势,该结论对于无定河流域的沙漠化治理、水土保持规划、水资源规划等具有重要的指导意义。

On account of the influence of human activities and climate change, the variation of hydrological time series may occur in a basin. According to the principle of long-range dependence analysis of time series based on Hurst coefficient, as well as the R/S analysis method and the fractional Brownian motion theory, a hydrological variation analysis method was proposed, which could identify and test the variation of time series and its degree on the whole, i.e. no variation （weak variation）, moderate variation, strong variation, giant variation. Combining with the method of GIS spatial analysis, the distribution map of spatial and temporal variation of annual rainfall series can be obtained. It has been applied for the Wudinghe Basin successfully. The results indicate that the annual rainfall series in majority of the Wudinghe Basin area are no variation （weak variation）, but the annual rainfall series in a part of the area are moderate variation. For example, Gao Town has a decreasing trend. This conclusion has an important guiding significance for desert governing, water and soil conservation planning, and water resources planning in the Wudinghe Basin.