Development of a grid-based emission inventory and an air quality model for Dhaka city

Abstract

Dhaka, the capital city of Bangladesh, is at the risk of experiencing major health impacts resulting from poor air quality. Reliable emission inventory and air quality model are essential prerequisites for assessment of health impacts and analysis of possible options for air quality management. In this study, a GIS based spatially disaggregated emission inventory has been developed for five air pollutants, namely, PM10, PM2.5, CO, NOx and SOx, considering three major sources i.e. motor vehicles, road dust and bricks kilns in and around the Dhaka city. The developed emission inventory has been used as input in an air quality model (ATMoS) to predict ambient concentrations of particulate matter (PM). The model domain considered for developing emission inventory and air quality model is about 1,800 sq. km. covering areas surrounding Dhaka city, including the brick kiln clusters around the city. The model area has been divided into 200 grids of 0.03° × 0.03°, which is approximately 3 km × 3 km. Since the brick kilns operate only during the dry season (November to mid-April), the emission inventory has a district seasonal variation, with brick kilns dominating the dry season emission, while road dust dominating the wet season emission. The north-western and western parts of Dhaka (representing Savar, Gazipur and Dhamrai areas) account for a large portion of brick kiln emissions during the dry season. Brick kilns located to the eastern periphery of the city (Kaliganj and Rupganj) and to the south of the city (Narayanganj, bandar and Sirajdikhan) are also responsible for significant emission. Vehicular and road dust emissions are significant along major roads and intersections. Average total monthly emission of PM10 for dry months (24,606 tons/month) is about ten times higher than that for a wet season (2,584 tons/month). During dry season, emission from brick kiln accounts for about 87 percent of PM10 emission, followed by road dust (11.2 percent) and vehicle (2 percent). During wet period, road dust becomes dominant contributor of PM emission, accounting for over 80 percent of PM10 emissions, and 64 percent of PM2.5 emissions. Diesel driven vehicles (i.e., buses and trucks) are responsible for majority of PM10, PM2.5, SOx, and NOx emissions. Together, buses and trucks account of about 81 percent of vehicular PM10 emissions, 88 percent of vehicular PM2.5 emissions, 94 percent of vehicular SOx emissions, and 83 percent of vehicular NOx emissions. The predicted ambient PM concentrations within and around Dhaka city has been found to vary widely, depending on the presence of emission sources (brick kilns, major roads) and meteorology (primarily wind direction and precipitation). The areas to the north-east of the Dhaka city i.e., Kaliganj, Sreepur are less polluted, primarily because they are not located down-wind of the major brick kiln clusters. During wet season (April to October), predicted PM concentration are relatively low throughout the model domain, but they are much lower for the boundary areas of Dhaka City (e.g., Kapasia, Savar, Keraniganj), compared to the city-centre areas. For some months, e.g., February and March, the predicted values matched well the values recorded at the CAMS, but in general, the predicted values are lower than those recorded at the CAMS. Inclusion of industrial emissions is likely to improve the model predictions. Although brick kilns are the dominant emission source during the dry season, source apportionment exercise suggest that vehicular emission and road dust account for major fractions of ambient PM concentration within city areas. For example, at the CAMS location near Shangshad Bhaban, vehicular emission, road dust and brick kilns account for about 36 percent, 29 percent, and 17 percent, respectively of ambient PM2.5 concentration in March. The developed emission inventory model is flexible such that it can take as input user defined parameters such as emissions factors, activity rates (e.g., AADT for vehicles), fuel use by different vehicle types, etc. Thus the inventory model can be easily updated as new information about the parameters becomes available. The air quality model together with the emission inventory, when fully developed and calibrated, could become a very useful policy analysis tool for air quality management.