Spatial distribution of aerosols burden and evaluation of changes in aerosol optical depth using multi-approach observations in tropical region

Understanding of Aerosol optical depth (AOD) parameter is important for air quality assessment. This study aims to evaluate and validate AOD measurements from combine datasets to improve air quality for a period 2005-2020 using Aerosol Robotic Network (AERONET) at Ilorin site (8.320 degrees N, 4.340 degrees E) in Nigeria. AOD outputs from Community Atmosphere Model Version 6 with chemistry (CAM6-chem) at 1 degrees horizontal resolution and Modern-Era Retrospective analysis for Research and Applications (MERRA-2) are investigated in addition to validation of two satellites AOD retrievals: Moderate Resolution Imaging Spectroradiometer (MODIS) and Multi-angle Imaging Spectroradiometer (MISR). Result of spatial distribution of AOD shows high values > 1 in the North and Western Sahara compared to Central Africa. Desert dust shows largest contribution in the North and Western Africa that is up to 2 magnitude larger than other aerosol types. Primary organic matter (POM) and secondary organic aerosols (SOAs) both presents high burdens with later been dominant at around 10 degrees band, and black carbon (BC) largest burden (2.6 x 10(-5) kgm(-2)) is seen in the model from oil and gas exploration site in Nigeria. Inter-comparison of MERRA/MISR/MODIS and AERONET AOD using linear correlation of the seasonal dependence demonstrated high correlation (r = 0.864 - 0.973) subjected to Root Mean Square Error (RMSE = 0.069 - 0.211), suggesting good agreement between the datasets. When compared to seasonal mean maximum AERONET AOD value of 0.978 MERRA is similar to 5%, MISR similar to 28% and MODIS similar to 29% lower with stronger correlations observed in the wet and pre-harmattan seasons. Similarly, MODEL AOD at 550 nm and dust burden were found to be similar to 34% and similar to 67% lower in context to AERONET AOD annual mean value of 0.627. Positive relationships that indicate an upward slope exist between all the computed datasets with moderate value of AERONET/CAM-chem spearman partial correlation, and MERRA/MODIS and MODIS/MISR showing strong and significant relationship with p-value less than 0.05. Low variance is observed with all measurements except in MERRA.