Simulation of urban and regional air pollution in Bangladesh

We have developed a regional scale air quality simulation using the Weather Research and Forecasting (WRF) - Community Multiscale Air Quality Model (CMAQ) to assess the suitability of such an advanced modeling system for predicting the air quality of Bangladesh and its surrounding region. The Regional Emission Inventory in Asia (REAS) was used as the emission input in this modeling approach. Both meteorological and chemical model performance were evaluated with observations including satellite data. Comparison between simulated and observed meteorological parameters revealed that the WRF can generate the necessary meteorological inputs for CMAQ. Comparison of observed and simulated concentrations of different air pollutants revealed that CMAQ greatly underestimates the concentrations of key pollutants. Comparison with satellite observations revealed that CMAQ reproduces the spatial distribution of NO2 with some underestimation in Bangladesh and India. The simulated AOD and satellite-retrieved AOD showed good temporal and spatial agreement mutually, with a correlation coefficient of 0.58. Sensitivity simulation using higher horizontal resolution emission data made by re-gridding the REAS inventory with the population distribution improved the CMAQ performance. Nevertheless, CMAQ underestimated the pollutant concentrations in Dhaka. Uncertainties in the emission inventory and in the lack of time variation in emissions input mainly contributed to the model underestimation. Model predictions show that 36-72% PM10 and 15-60% PM2.5 in Dhaka might be contributed from brick kiln emissions in monthly average of January 2004. The chemical composition of PM2.5 showed that the considerable amounts of secondary aerosols in Dhaka and carbonaceous components (particularly organic carbon) are most responsible for the model underestimation. Results suggest that improvements of emission inputs and more detailed sensitivity analysis of CMAQ model are important to assess the reliability and applicability of WRF-CMAQ modeling system in this region.