Back to Search
Start Over
Deriving PM2.5 from satellite observations with spatiotemporally weighted tree-based algorithms: enhancing modeling accuracy and interpretability.
- Source :
- NPJ Climate & Atmospheric Science; 6/18/2024, Vol. 7 Issue 1, p1-9, 9p
- Publication Year :
- 2024
-
Abstract
- Tree-based machine learning algorithms, such as random forest, have emerged as effective tools for estimating fine particulate matter (PM<subscript>2.5</subscript>) from satellite observations. However, they typically have unchanged model structures and configurations over time and space, and thus may not fully capture the spatiotemporal variations in the relationship between PM<subscript>2.5</subscript> and predictors, resulting in limited accuracy. Here, we propose geographically and temporally weighted tree-based models (GTW-Tree) for remote sensing of surface PM<subscript>2.5</subscript>. Unlike traditional tree-based models, GTW-Tree models vary by time and space to simulate the variability in PM<subscript>2.5</subscript> estimation, and they can output variable importance for every location for the deeper understanding of PM<subscript>2.5</subscript> determinants. Experiments in China demonstrate that GTW-Tree models significantly outperform the conventional tree-based models with predictive error reduced by >21%. The GTW-Tree-derived time-location-specific variable importance reveals spatiotemporally varying impacts of predictors on PM<subscript>2.5</subscript>. Aerosol optical depth (AOD) contributes largely to PM<subscript>2.5</subscript> estimation, particularly in central China. The proposed models are valuable for spatiotemporal modeling and interpretation of PM<subscript>2.5</subscript> and other various fields of environmental remote sensing. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 23973722
- Volume :
- 7
- Issue :
- 1
- Database :
- Complementary Index
- Journal :
- NPJ Climate & Atmospheric Science
- Publication Type :
- Academic Journal
- Accession number :
- 177963560
- Full Text :
- https://doi.org/10.1038/s41612-024-00692-4