均一化中国陆表气候观测格点数据集（地表太阳辐射、地表风速、相对湿度和陆表蒸散）

日照时数重建可以较好的反应地表太阳辐射的长期变化趋势，但只是站点数据。因此要得到高分辨率网格点数据，同时保证其在长期变化方面上的精度，需要融合多种地表太阳辐射相关数据。利用地理加权回归（GWR）方法，融合了MODIS 0.1°分辨率云量和气溶胶反演以及地面日照时数重建地表太阳辐射站点数据。通过增加相邻点数方案的组合判断，有效地提高了地理加权回归降尺度结果的准确性，得到的中国区域多年平均值、长期趋势与观测和卫星遥感反演结果基本一致。采用地理加权回归等方法，生成0.1度网格的地表风速和相对湿度数据；利用改进的彭曼公式计算了陆表蒸散数据。

Reconstructed sunshine duration can well reflect the long-term trend of surface solar radiation, but it is only available as station-based data. Therefore, to obtain high-resolution grid data while ensuring accuracy in long-term variation characteristics, it is necessary to fuse multiple surface solar radiation-related datasets. Using the Geographically Weighted Regression (GWR) method, we fused MODIS 0.1° resolution cloud cover and aerosol retrieval products with station-based surface solar radiation data reconstructed from ground sunshine duration observations. By incorporating combined judgments of different neighboring point number schemes, the accuracy of GWR downscaling results was effectively improved. The multi-year average values and long-term trends over China derived from this approach are generally consistent with those from in-situ measurements and satellite remote sensing retrieval results. Surface wind speed and relative humidity data at 0.1° grid resolution were generated using methods including GWR; land surface evapotranspiration data were calculated via the modified Penman formula.

搜集汇总

背景与挑战

背景概述

该数据集是一个均一化的中国陆表气候观测格点数据集，包含地表太阳辐射、地表风速、相对湿度和陆表蒸散四个关键变量，时间分辨率为月，空间分辨率为10km-100km，数据覆盖时间从1979年或1983年至今。数据集通过融合地面日照时数观测与卫星遥感数据（如MODIS云量和气溶胶反演），采用地理加权回归等方法生成，旨在提供高精度、长期一致的气候数据产品，适用于中国区域的气候变化研究和分析。