基于多源数据的复杂地形高分辨率旬降水量模拟研究

The continuous high-resolution precipitation spatial data is an important input to hydrological and ecological models and also an important basic data for climate change research. In this study, firstly, mountain terrain in China is topographic decomposition from fine to coarse, from microscopic to macroscopic, and from small terrain to large terrain, and the impact of different scales of terrain on precipitation is analyzed. Then, based on the idea of separation and integration, multi-source data of surface meteorological station observation data, GPM IMERG satellite precipitation products, TMPA satellite precipitation products and DEM topographic decomposition data are used to complete the spatial resolution simulation of high-resolution and dekad precipitation in mountainous areas of China. In addition, a regional sub-ten data cluster scheme is designed to determine the model parameters, and the dominant precipitation azimuth and the maximum precipitation increasing azimuth are introduced into dekad scale. We assess the quality of the downscale method by cross-validation, local pattern analysis, cryptographic station validation, and more. It not only provides a new method for simulating dekad precipitation in complex terrain, but also provides more refined input data in these fields such as eco-hydrology and water resources assessment.

连续高分辨率降水空间化数据是水文、生态模型重要的输入数据，同时也是气候变化研究的基础数据。本研究首先基于EMD方法对我国山区地形进行由细到粗、由微观到宏观、由小地形到大地形的分解，分析不同尺度地形对降水的影响；然后利用地面气象站观测资料、GPM IMERG卫星降水产品、TMPA卫星降水产品、DEM地形分解数据等多源数据，采用分离综合的思想，将降水估算分成为趋势项和地形修正项，引入旬尺度的主导降水方位与最大降水递增率方位，设计区域分旬数据集群方案以确定模型参数，实现基于降水产品的中国山区旬降水降尺度方法。通过交叉验证、局地规律分析、加密站验证等评估旬降水降尺度方法的质量。本研究不仅为复杂地形下旬尺度降水模拟提供一种新的研究方法，也可为生态水文模拟、水资源评价等领域提供更加精细的输入数据。

连续高分辨率降水空间化数据是水文、生态模型重要的输入数据，同时也是气候变化研究的基础数据。本项目采用地面气象站观测资料、GPM IMERG卫星降水产品、DEM地形分解数据等多源数据，实现降水量空间分布的精确模拟。开展的主要工作有：优化地形分解算法；比较中国降水卫星产品的质量；分析地形对降水的局地影响；构建中国分旬降水空间化模型；生成一套中国高分辨率降水数据集。研究结果表明地形对精细化降水分布具有重要影响；地形分解技术可以有效提高降水空间化模拟。本项目不仅为复杂地形下旬尺度降水模拟提供一种新的研究方法，也可为生态水文模拟、水资源评价等领域提供更加精细的输入数据。