Dhaka is among the six South Asian cities with at least 15 million inhabitants and included as 20 mega cities of the world. The other cities are Kolkata, Mumbai, Delhi , Karachi and Lahore . Other notables include Chennai, Bangalore , Hyderabad , Islamabad , Colombo , and Kathmandu , and all these cities are facing air pollution problems.

Dhaka city is expanding in all directions - east to west, north to south. The population, energy and infrastructure demands are increasing but the amenities (including environmental) can not keep pace with the growing demand. A recent study in 2008 concluded that an estimated 15,000 premature deaths, as well as several million cases of pulmonary, respiratory and neurological illness are attributed to poor air quality in Dhaka

Figure 1 presents an overview of the measured monthly average PM_2.5 concentrations at the Sansad Bhavan (Parliament).

The air quality in the Dhaka city has deteriorated over the last decade due to a rapid change in the vehicular fleet, increased congestion, and a large increase in the industrial activity (in and around the city). The annual average concentrations for PM_2.5 are ~100 mg/m³, above any of the standards for clean air and better health.

In the city, transport dominates as an air pollution source. Among the industries, the brick kilns are the major source, especially during the manufacturing season of October to March, depending on the monsoonal rains. In Figure 1, the PM_2.5 concentrations are split between the brick making season and the rest, clearly presenting a distinct change in the pollution trends. The seasonal averages have shifted over the years and during the brick manufacturing season, pollution peaks measured ~230 mg/m³ in January 2008.

A majority of the brick kiln clusters are to the North of the city, presented in Figure 2 and the measured peak values represent the worst case scenario of the maximum wind blowing towards the city (the red line boundary). The clusters account for ~700 brick kilns.

The analysis presented in this study focuses on the review of the current source apportionment studies (Annex 1) quantifying the share of the brick kiln emissions affecting the air pollution in the Dhaka city and modeling of the impacts of the brick kiln clusters. Discussion of the other sectors (transport, domestic, or other industries) is not included in this paper.

Dhaka Air Pollution Source Apportionment

Based on the source apportionment studies, listed in Annex 1, during the dry season (October to March) - vehicular emissions, particularly motor cycles, diesel trucks and buses (most dominant of the sources in both fine and coarse mode); soil and road dusts arising from civil construction, broken roads, and open land wind erosion; biomass burning in the brickfields and city incinerators (to the fine mode) are the major sources of PM pollution. Figure 3 presents a summary of the source apportionment study conducted by the Bangladesh Atomic Energy Center , Dhaka , Bangladesh for fine and coarse mode particulates at two stations – Farm gate and the Dhaka University premises.

The analysis was conducted using ‘ GENT ’ stacked filter samplers, followed by analysis of the filter samples using PIXE and receptor modeling using PMF. The study included development of city specific source profiles, which provide the necessary information on the biomarkers to identify sources, analyze the measured samples, and estimate the percentage contributions of various sources. It is important to note that the source apportionment results cannot be generalized to the whole city, but they do provide a basis for further analysis and an understanding of the mix of sources contributing to the air pollution, especially around the hot spots. The process of source apportionment, applied in this study, is expensive, which prohibits inclusion of many measurement points, unlike a mobile monitoring station which can be used to measure multiple points, but cannot be used to calculate source contributions.

The motor vehicles are a known and visible source of particulate pollution in Dhaka , which require interventions ranging from technical (emission standards) to institutional (inspection and maintenance). The dust, due to the resuspension on the roads, an indirect source of motor vehicle activity, is a major cause of air pollution (in the coarse mode of PM), due to lack the sufficient infrastructure (paved roads) to support the growing fleets and congestion on the roads. The study estimates at least 50 percent of the coarse PM is linked to the resuspension of the road dust, followed by the industrial smelters.

The clusters of brick kilns lying north of Dhaka contribute ~40 percent of the measured fine PM pollution. Growing construction activity (also contributing to the fugitive dust) is leading the demand for brick kilns and burning of biomass and low quality coal is resulting in pollution.

Similar source apportionment studies were conducted in other Bangladeshi cities, e.g. Rajshahi, Chittagong and Khulna . Results from the preliminary work conducted at Rajshahi station are presented Figure 4. Similar to Dhaka , the results suggest the brick kiln industry dominates the fine PM pollution in Rajshahi.

With the seasonal fine PM averages of ~150 mg/m³ (Figure 1) a 40 percent share of ~60 mg/m³ presents a tremendous opportunity for air pollution control from the brick kilns in and around Dhaka .

Dispersion of Emissions from Brick Kilns

The source apportionment studies, due to their limitation of not able to analyze multiple stations, a bottom up analysis of the energy consumption, emissions analysis, and dispersion is necessary to better understand the physical extent of the impact of these brick kiln emissions on Dhaka ’s air quality.

Note that this study was conducted for the brick kiln emissions only and it is NOT a representation of the Dhaka ’s entire air quality. The results are compared to the known source apportionment results for consolidation of the two methodologies and make way for further analysis and introduction of effective interventions for brick kiln emissions control.

ATMoS – Dispersion Model

A number of modeling systems exist with varying capacities to analyze industrial, urban, regional, and global scale air quality. For Dhaka , the air pollution dispersion was conducted using the Atmospheric Transport Modeling System (ATMoS), a meso-scale three-layer forward trajectory lagrangian puff-transport model. The model schematics are presented in Figure 5.

The model was developed for sulfur pollution modeling, as part of the Regional Air Pollution Information System for Asia (RAINS-Asia). The model was, since, extensively applied for sulfur and particulate modeling studies in Asia for regional, national, and urban scale studies

In the ATMOS/UrBAT model, the emissions are modeled as puffs released every 3 hours or every hour respectively, for regional and urban simulation. Each puff is assigned a mass proportional to the source strength and uniformly mixed in the assigned vertical layer and diffused along a Gaussian distribution in the horizontal direction. Individual emission puffs are tracked for 120 hours (5 days) or until the mass falls below a cut-off value of 0.1 percent of the emission source strength or puffs fall out of the simulation domain. Puffs transported beyond the modeling domain are no longer tracked.

The model separates the vertical dimension into two layers during the day and three layers at night. The day layers are the boundary and upper layer. The night layers are the surface, boundary, and upper layers (see Figure 6). The night surface layer extends from ground to 300 meters. The boundary and upper layers are separated by the critical inversion, which determined from the vertical temperature profile. The maximum height for the model is 6000 meters.

The model has flexible temporal and spatial resolution. The horizontal spatial resolution can be varied from 1000 meter for an urban scale study to 1° x 1° (approximately 90 km) for a regional scale study. However, using this model for analysis of episodic analysis is not recommended, due to the applied advection schemes, which tend to simply the interaction between the horizontal and vertical layers. The model is applicable for the regional and urban scale studies, analyzing the seasonal and annual air quality for long term trends and evaluating “what-if” scenarios.

For modeling purposes, the emissions are emitted at the center of the grid point for area sources and at the exact locations for elevated sources (in hundredths of a degree). Area emission sources (small industries, incinerators, boilers and mobile sources) are released into the boundary layer during the day and into the surface layer at night, and the elevated sources (large point sources, LPS’s) are released into the boundary layer during the day and above the surface layer at night. Volcanic sources are released into the upper layer at all times.

The released puffs expand and follow their own trajectory, depending on the prevailing meteorological conditions. However, during day to night and night to day transitions the puff branches into multiple puffs. This action is performed to simulate vertical wind shearing of the puff. Furthermore, the mixing height acts to trap urban surface emission and magnify their influence. The model uses a time-step of 1 hour on which advection and surface processes (wet and dry deposition) are calculated and averaged over the desired time period.

The model is run separately for sulfur, nitrogen, and particulate pollution. For particulates, due to physical and chemical differences in PM₁₀ and PM_2.5, two separate bins are distinguished with varying dry and wet deposition functions. In addition, sulfate and nitrate concentrations are added to the PM₁₀ and PM_2.5 fractions to reflect the contribution of the secondary particles.

Meteorological Data

The ATMoS model uses the meteorological data from the NCEP/NCAR ( National Center for Environmental Prediction) Reanalysis Data. The global dataset is available at 2.5° by 2.5°, from which is the grid cells containing the Dhaka city is extracted. The meteorological input data includes six hourly precipitation, mixing height, and wind vector information. Figure 7 presents the seasonal average windrose functions for the year 2004.

The brick kiln manufacturing season runs from October to March, during which the northerly winds enhance the contribution of their emissions to the Dhaka air quality. The precipitation and mixing heights information is not presented here. All the calculations are conducted using the meteorological data from 2004.

Brick Kiln Emissions

The emissions analysis and data on the physical location of the brick kiln clusters was collected by the team from BUET. The details (latitudes and longitudes) are presented in their final project report. At the brick kilns, measurements included an emission rate of 44 gm/sec of TSP. The emission rates were converted to PM₁₀ (using a ratio of 0.3 to TSP) and to PM_2.5 (using a ratio of 0.3 to PM₁₀). This amounts to a total of 108 ktons of PM₁₀ for 180 days of operating 530 brick kilns presented in Figure 2.

Impact of Brick Kiln Emissions

The Figure 8 (left panel) presents the seasonal (October to March) average contribution of brick kiln emissions, followed by the modeled monthly maxima over the Dhaka metropolitan area. It is important to note the contribution of the brick kilns is incremental to the daily vehicular, domestic, and industrial activities during these months. It is evident in the Figure 1, which demonstrates the dramatic increase in the PM_2.5 levels measured in the city premises.

The seasonal averages estimate a contribution of 15-60 mg/m³ of PM_2.5, which translates to 30-40 percent of the measured PM_2.5 concentrations in Figure 1, which corresponds to the estimated source contribution of brick kilns in Figure 3 (via source apportionment). The monthly maxima (right panel) correspond to the worst case scenario of heavy northerly winds and possible maximum modeled contribution of 30-100 mg/m³ of PM_2.5 over the season.

Figure 8 presents monthly average contributions of brick kilns for the 6 months of operation.

The calculations presented in this paper should not translate to undermining the influence of the other sectors on Dhaka ’s air quality. At the ground level, the transport sector (the direct vehicle exhaust, road dust due to resuspension, and idling emissions) contributes more to the exposure levels than the long range transport of the emissions from outside the city.

In Conclusion

The main objective of this analysis was to demonstrate the schematics of the modeling approach in understanding the contribution of brick kiln emissions in Dhaka , the characteristics of the air pollution dispersion, and the physical extent of the influence of these emissions on public health.

The brick kilns contribute significantly to the Dhaka air pollution problems, especially during the heightened manufacturing season and requires stringent interventions to reduce their incremental impact of the local air quality and health.

For example, the incremental pollution of 40 mg/m³, due to the brick kilns in the north, translates to an increase in ~5,000 premature deaths annually in the Dhaka city.

A final report was prepared by the BUET (followed by ISC3 modeling of the impact of Brick Kiln emissions) and the World Bank team, focusing on the interventions to improve the combustion technologies, the financial implications, and possible reductions on Dhaka air quality. Those details are not presented here and the reader should contact the project team for further information.

The complete study can be accessed at SIM Working Paper Series No. 21-2009