4.7 Article

Absorbing Aerosol Optical Depth From OMI/TROPOMI Based on the GBRT Algorithm and AERONET Data in Asia

出版社

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2022.3231699

关键词

Absorbing aerosol optical depth (AAOD); Asia; machine learning; Moderate Resolution Imaging Spectro-Radiometer (MODIS); TROPOspheric Monitoring Instrument (TROPOMI)

向作者/读者索取更多资源

To obtain accurate and high-resolution absorbing aerosol optical depth (AAOD) for pollution tracking, a gradient boosted regression trees (GBRT) method based on joint data from Ozone Monitoring Instrument (OMI), Moderate Resolution Imaging Spectro-Radiometer (MODIS), and AErosol RObotic NETwork (AERONET) is applied to TROPOspheric Monitoring Instrument (TROPOMI). The results show a correlation coefficient greater than 0.6 compared to ground-based data, with a difference generally within +/- 0.04. The underestimation issue compared to OMI data has been greatly improved, and the study provides important insights for research on regional and urban anthropogenic pollution.
Quantifying the concentration of absorbing aerosol is essential for pollution tracking and calculation of atmospheric radiative forcing. To quickly obtain absorbing aerosol optical depth (AAOD) with high-resolution and high-accuracy, the gradient boosted regression trees (GBRT) method based on the joint data from Ozone Monitoring Instrument (OMI), Moderate Resolution Imaging Spectro-Radiometer (MODIS), and AErosol RObotic NETwork (AERONET) is used for TROPOspheric Monitoring Instrument (TROPOMI). Compared with the ground-based data, the correlation coefficient of the results is greater than 0.6 and the difference is generally within +/- 0.04. Compared with OMI data, the underestimation has been greatly improved. By further restricting the impact factors, three valid conclusions can be drawn: 1) the model with more spatial difference information achieves better results than the model with more temporal difference information; 2) the training dataset with a high cloud fraction (0.1-0.4) can partly improve the performance of GBRT results; and 3) when aerosol optical depth (AOD) is less than 0.3, the perform of retrieved AAODs is still good by comparing with ground-based measurements. The novel finding is expected to contribute to regional and even urban anthropogenic pollution research.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.7
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据