Determination of Satellite-Derived PM2.5 (A case study of Kampala District).

Abstract

Particulate Matter with diameter less than 2.5μm (PM2.5) is highly inhalable and has been linked to a number of cardiovascular complications. This has in turn intensified the need to monitor these concentrations over time and thus ground monitoring stations have been established. Although ground monitoring stations provide accurate PM2.5 measurements, they are expensive to maintain and also provide information localized to the locations of the measuring stations hence limited for large scale coverage. This study, therefore to explored the use of satellite images (Landsat-8 and Sentinel-2) in estimating PM2.5 concentrations and its spatial distribution in Kampala district. Firstly, Aerosol Optical Depth (AOD) was computed using the Code for High Resolution Satellite mapping of optical ThIckness and aNgstrom Exponent algorithm (CHRISTINE code). Derived AOD was then characterized with reference to meteorological factors, and then correlated with in-situ PM2.5 to determine satellite-derived PM2.5 using Geographically Weighted Regression. Validation of the results was done using the linear correlation between in-situ and estimated PM2.5 values and the Root Mean Square Error (RMSE) values. From the results, analysis of in-situ PM2.5 measurements showed that PM2.5 varies both spatially and temporally with distinct seasonal differences. Also, correlating in-situ PM2.5 and AOD revealed that the relationship is highly variable over time thus needs to be modelled for each satellite overpass time, rather than having a generic model fitting say, a season. The satellite-derived PM2.5 showed good model performance with coefficient of correlation (R2) values ranging from 0.69 to 0.89. Furthermore, sentinel-2 data showed better prediction, signifying that increasing spatial resolution can improve satellite-derived PM2.5 estimation. Validation of these results showed that the estimated values are comparable with in-situ measurements implying that satellite measurements are representative of ground observations. Therefore, the study recommends targeted monitoring of PM2.5 to be carried out at hotspot areas using higher resolution images to determine PM2.5 concentrations for further assessment and timely intervention.