Developing a spatially distributed emission inventory for Dhaka city

Abstract

Air quality of Dhaka city is major concern, especially during the dry season. For addressing the air quality problem, development of an emission inventory is absolutely essential. In this study, a GIS based spatially disaggregated emission inventory has been developed for Dhaka city and its surrounding areas, incorporating major transportation and industrial sources. The study area was bounded by 23o32׳38״ and 24o6׳56״ North latitudes and 90o16׳14״and 90o37׳26״East longitudes, covering Dhaka and its surrounding industrial areas. For developing the emission inventory, the study area was divided into 12 x 21 grids (i.e., 252 grids), each grid having a size of 3 km by 3 km. From the transportation sector, vehicular emission was considered, while from industrial sector brick kilns were considered. The pollutants considered included PM10, PM2.5, CO, NOx and SOx. For estimation of vehicular emission, the data required included emission factors for different types of vehicles using different types of fuels and activity level of vehicles in each grid. Based on an assessment of available information, a set of emission factor was selected for use in this study. The activity level of vehicles (VKT or vehicle kilometer travelled) in each grid was determined by estimating the lengths of the major roads in each grid (using GIS based maps) and multiplying that with the annual average daily traffic (AADT) for these roads, as reported in Strategic Transport Plan (2004). Among the modes of transport considered, cars were found to have the highest activity level within the study area (over 2.64 million km/day); it was followed by autorickshaw (about 2.26 million km/day) and bus/minibus (about 1.41 million km/day). Grid-wise vehicular emission inventory was prepared for each of the pollutants selected and presented in the form of GIS maps for easy visualization. A number of grids/areas located in central Dhaka have been found to be responsible for very high emission. The grid/area responsible for the highest emission (accounting for about 7% of total emission within the study area) was located at intersection of Kazi Nazrul Islam Avenue, part of Airport Road, Panthapath and New Eskaton Road. Diesel driven buses and trucks were found to be mostly responsible for PM10 and PM2.5 emissions. Buses and trucks accounted for 47% and 30% of total PM10 emission, and 54% and 25% of total PM2.5 emission, respectively. Buses and trucks accounted for about 86% of total NOx emission; with bus alone responsible for about 66% of total emission. This is not surprising, because diesel driven vehicles usually emits higher NOx because of higher engine temperature, which encourages higher production of thermal NOx. Buses and trucks, which primarily run on high S-containing diesel were found to be responsible for majority of SOx emission, each responsible for about 42% of total emission. Gasoline driven vehicles were found to emit more CO and Hydrocarbons. Results from this study shows that trucks and buses are responsible for only 19% of total emission of CO, while cars and autorickshaws account for 35% and 30% of total emission of CO, respectively. Pollutants, especially particulate matter, emitted from brick kilns are considered a major source of air pollutants for Dhaka city. In this study, emission of particulate matter from large number of brick kilns located in Gazipur, Savar and Keraniganj have been estimated based on a fixed emission rate and duration of production. The 249 brick kilns located in Gazipur have been found to be responsible for emission of 46.9 kton of PM10 and 14.1 kton of PM2.5 during the five and a half month production period. Brick kilns located in Narayangonj, Keranigonj and Rupgonj were also found to be responsible for significant emission of particulate matter. To understand the uncertainties associated in generation of emission inventory, an uncertainty analysis tool “Monte Carlo Simulation technique” has been used which helped in understanding the uncertainty associated with each pollutant. The most uncertainty was associated with NOx, which is followed by SOx, PM2.5, PM10 and CO. The spatial distribution of emissions for confidence level of each pollutant has been measured by Monte Carlo Simulation technique. A sensitivity analysis has been carried out to understand the impact of input variables on emission inventory. The spatially disaggregated emission inventory developed in this study could be very useful for a clear understanding of emission sources within and around Dhaka city. This emission inventory could be conveniently used as input in air quality models. This inventory could also be upgraded relatively easily incorporating new roads and sources in a grid-wise fashion.