【摘要】 沙尘气溶胶是大气污染的主要成分，深刻影响着气候环境、人类健康，也是近年来遥感等领域的研究方向之一。本数据集基于MCD19A2、MERRA-2、ERA5-Land数据集，通过数据预处理、数据上采样、时空匹配将气溶胶光学厚度、?ngstr?m指数、气象参数等多种数据提取，将AERONET地面站点观测值作为目标真值，利用机器学习算法LightGBM训练模型，反演2018–2022年“一带一路”沙漠地区的逐日1公里分辨率陆地沙尘气溶胶光学厚度。独立站点评估结果为：全参模型R(RMSE)为0.8789(0.0725)；非全参模型R(RMSE)为0.8559(0.0805)。本沙尘气溶胶光学厚度数据集为“一带一路”地区的沙尘研究提供基础数据。更多还原

【Abstract】 Dust aerosols are a major component of atmospheric pollution, significantly impacting climate,the environment and human health, and have become an important focus of remote sensing research in recent years. This dataset is based on the MCD19A2, MERRA-2, and ERA5-Land datasets. Through data preprocessing, upsampling, and spatiotemporal matching, it integrates multiple variables, including aerosol optical depth, ?ngstr?m Exponent, and meteorological parameters. Using AERONET ground station observations as the ground truth, we employed the LightGBM machine learning algorithm to train the model and retrieve daily land dust aerosol optical depth at 1-kilometer resolution across the desert regions of the Belt and Road Initiative from 2018 to 2022. The evaluation results for the independent site are as follows: the full-parameter model R(RMSE) is 0.8789(0.0725); the non-full-parameter model R(RMSE) is 0.8559(0.0805).This dataset provides fundamental data for dust research in the region along the Belt and Road.更多还原