Fig. 3: Distribution of variable importance across China. | npj Climate and Atmospheric Science

Fig. 3: Distribution of variable importance across China.

From: Deriving PM2.5 from satellite observations with spatiotemporally weighted tree-based algorithms: enhancing modeling accuracy and interpretability

Fig. 3

For each grid cell, there are a series of results over time. a represents the most important variable, which is the mode of the time series values, 1–8 represent AOD, PS, UW, VW, T, HCHO, NO2, and CAMS-PM2.5; b is the average ranking of AOD, where a smaller value indicates a higher ranking.

Back to article page