Fig. 1: Validation of estimated PM2.5.
From: Reversal of trends in global fine particulate matter air pollution
a Definition of 13 regions (consistent color-coding throughout the paper) and the distribution of monitoring sites (black dots). b Scatter plot (color-coded for the 13 regions) of annual mean PM2.5 from ground-based monitors (X-axis) and the 1 km satellite-drived estimates (Y-axis) in 2017, with overall statistical metrics indicated at the upper-left. c Regional distribution of R2 (size) and normalized root mean square difference (NRMSD, color, Eqs. 1, 2) for the annual mean PM2.5 through 2015-2019. Number of available sites for evaluation are indicated in the brackets following the region names. d Bootstrapped NRMSD of regional population-weighted PM2.5 as a function of sample size in 2017. The map is created using packages cartopy56 and matplotlib57 in Python.
