基于小波分析的日降水时空模型构建及模拟

Establishment of spatial-temporal model for simulating daily precipitation based on wavelet

为了获取气象水文要素的时空分布特征,将汉江上游流域55个气象站点1981年日降水量作为随机信号处理,借助于小波分析理论表达其在时间系列上的变化特征。利用Haar小波函数实现随机性较高的日降水信息的时空离散化,得到较为准确的日降水时空数据集。同时,与利用Kriging方法、反距离方法以及样条方法等3种传统内插方法所得计算结果比较。结果表明,基于小波的内插方法能够对日降水量进行精确的时空模拟,其内插精度与传统内插方法相比,有很大程度的提高。

The daily precipitation of 55 meteorological stations over the Hanjiang River basin is regard as stochastic signal, and its change characteristics in temporal scale is expressed by wavelet. The daily precipitation signal, which is highly stochastic, is scattered utilizing the Haar wavelet and the spatial - temporal mapping of daily precipitation is obtained. Meanwhile, the result is compared with that of the 3 traditional methods, the Kriging method, inverse distance weight method and Spline method. The result show that the daily precipitation could be simulated by spatial - temporal interpolation model based on wavelet, and the inter- oolation nreeision is increased comnaring with the 3 traditional methods.