基于随机森林模拟的辽宁省降水量空间分布研究

Research on Spatial Distribution of Precipitation in Liaoning Province Based on Random Forest Simulation

绘制精细的降水量空间分布图是高效利用气象台站资料的重要方式之一。以辽宁省为例,基于空间离散、数量有限的站点降水资料,采用随机森林方法构建省域降水量空间分布模型。结果表明,随机森林模型能够较好拟合区域降水量空间分布规律,R2高达0.945~0.965,并且与STRM产品资料具有空间一致性,并且可反映降水分布细节信息。

Drawing an elaborate spatial distribution map of precipitation is one of the important ways to efficiently use meteorological station data. Taking Liaoning Province as an example, based on the spatially discrete and limited number of station precipitation data, the random forest method is used to construct the provincial precipitation spatial distribution model.The results show that the random forest model can better fit the spatial distribution of regional precipitation law,R2 is as high as0.945~0.965, and it has spatial consistency with STRM product data, and can reflect the detailed information of precipitation distribution.