River pollution and social inequalities in Dhaka, Bangladesh

Dhaka is one of the fastest growing urban agglomerations in the developing world, home to over 17 million people spread over an area of 307 km². The city is surrounded by six interconnected river systems—the Buriganga and Dhaleshwari in the southwest, Turag and Tongi Khal in the north, and Balu and Sitalakhya in the east—which support trade, transport, and stormwater drainage. However, in the past three decades, the water quality in the rivers surrounding Dhaka has been severely degraded due to indiscriminate discharge of untreated sewage, solid waste, and industrial effluent. Dhaka's sewer system covers only 20% of the city, and feeds into one sewage treatment plant operating at one-third of its capacity (DWASA 2019). In addition to the organic and pathogen pollution load from the 1.2 million m³ of untreated sewage, the rivers receive about 60,000m³ of industrial effluent every day from nine major industrial clusters. There are an estimated 500–700 wet processing and dyeing textile factories releasing a range of chemicals including salts, dyes and bleaches, and 155 tanneries discharging heavy metals, including chromium (Sagris and Abbott 2015). The biological oxygen demand (BOD) and chemical oxygen demand (COD) of the textile wastewater are estimated to be 480 mg l⁻¹ and 696 mg l⁻¹ respectively, compared to a BOD of 2000 mg l⁻¹ and COD of 4500mg l⁻¹ for tannery effluent (Sagris and Abbott 2015).

The textile and leather industries are critical to Bangladesh's economic growth, with an average annual gross domestic product (GDP) growth rate of 6.5% in the fiscal years (FY) from 2009 to 2018 (World Bank 2019). As a subset of the textiles industry, ready-made garments accounted for 83.5% of the country's export revenues and 11.2% of the GDP in FY2017-18 and employs about 4 million people. The leather industry, which contributed to 3.5% of annual exports and 0.35% of GDP in FY2017-18, is also of strategic importance as the country is seeking to diversify its manufacturing export base to sustain the growth trajectory (Hong 2018). While environmental regulations, as articulated in the Environmental Conservation Act (1995), require these 'red category' factories to treat their wastewater, not all factories comply and many do not run their effluent treatment plants due to gaps in financing for infrastructure and technology upgrades, or cost savings associated with the treatment chemicals and electricity needed to run the system (Haque 2017, Restiani 2017). In December 2017, following decades of domestic and international pressure to curb pollution, the tanneries in Hazaribagh (along the Buriganga River) were relocated to a new 200-acre industrial park in Savar along the Dhaleshwari River. The park is equipped with a centralised effluent treatment plant (CETP) with a daily capacity of 25,000m³ compared to 40,000 m³ of wastewater being generated. As of August 2021, it is estimated that the tanneries dumped more than 16.4 million m³ of untreated wastewater over the past three years, causing a parliamentary standing committee to order temporary closure of the industrial park (Hasan 2021).

Recognising the severity of pollution, four of the rivers surrounding Dhaka—Buriganaga, Sitalakhya, Balu and Turag—have been declared as ecologically critical areas in 2009 by the Department of Environment (DoE) (The Daily Star 2009), with the Planning Commission labelling them as 'unsuitable for any human use' (General Economics Division 2015, p.423). In 2016, DoE reported that the mean dissolved oxygen (DO) in dry and wet seasons ranged from 0. 17–2.98 mg l⁻¹ for Buriganga, 4.14–4.25 mg l⁻¹ for Sitalakhya, 0.4–4.51 mg l⁻¹ for Turag, and 2.75–6.05 mg l⁻¹ for Dhaleshwari, compared to the environmental standard of >5mg l⁻¹ for fisheries, irrigation and recreational uses (DoE 2017). Heavy metals contamination in surface water and their accumulation in vegetables (Rahman et al 2017) and fisheries (Ahmed et al 2016, Rashid et al 2017) in areas surrounding industrial clusters have been documented. Hasan et al (2019)'s analysis of hair and nail samples reflected significantly higher chromium accumulation in leather factory workers and residents in Hazaribagh compared to non-exposed people in a control village. The pollution also has significant implications for the city's growing water demand, 87% of which is currently sourced from groundwater as the extreme contamination makes surface water treatment economically and technologically unfeasible (Arfanuzzaman and Rahman 2017).

The industrial clusters in Dhaka are often interspersed with low-income residential areas or slums that accommodate workers and their families. The underdeveloped water and sanitation facilities in slums, home to 35% of the urban dwellers in Bangladesh, counteracts the 'urban health advantage' often generated from better employment and health care services compared to rural areas (Vlahov et al 2005). The Bangladesh Urban Informal Settlements Survey 2016 (Yanez-Pagans 2016), which included a representative sample of 588 households across small, medium and large slums in Dhaka, showed that 68% of the households accessed piped water through a shared connection within the slum compound, with the poorest households sharing a waterpoint with 43 other households on average compared to 23 sharers among the richest households (Haque et al 2020). Landlords were the main providers of water infrastructure in small and medium slums on privately-owned land. In large slums on government owned land, about 27% and 21% of the waterpoints were installed by the Dhaka Water Supply and Sewerage Authority (DWASA) and NGOs, respectively. More than half of the households also reported being exposed to flooding in the past year, which disrupted access to a functional water source for 10 days on average.

In terms of sanitation, only 8% of the slum households had access to a flush toilet connected to a septic tank, while 78% used improved pit latrines and the remaining 10% depended on hanging latrines (Haque et al 2020). A toilet facility was shared among 16 households on average. Among those with improved sanitation facilities, about 68% reported never emptying their pits or septic tanks, while 20% mentioned disposing the faecal sludge directly to drains or nearby waterbodies. Amin et al (2020)'s study in two low-income neighborhoods in Dhaka showed extensive V.cholerae, NoV-GII, Giardia, and Shigella/EIEC contamination in effluent from on-site sanitation systems, to which children and adults may be exposed in their immediate residential environment. A larger study of environment faecal contamination across 10 neighbourhoods found significantly higher E. coli concentrations in non-municipal drinking water, bathing water, surface water, and soil samples in low-income neighborhoods compared to high-income neighborhoods, indicating poor drainage systems, improper child faeces disposal, and poor faecal sludge management in low-income neighbourhoods as the potential contamination routes (Amin et al 2019).

Consequently, many of the immediate environmental burdens of pollution are borne by these low-income residents. Longitudinal observations from hospital visits between 1993 and 2012 showed that children under five living in densely populated slums were significantly more prone to diarrhoeal diseases from Vibrio cholerae and severe dehydration compared to those in non-slum settings (Ferdous et al 2014). In Kamrangirchar, the largest slum in Dhaka and home to about 600,000 people who generated income from the tanneries in adjacent Hazaribagh area for wages, occupational and environmental exposure to toxic heavy metals, coupled with low living standards and poor health seeking behaviour, contributed to high disease burden, including contact dermatitis, work-related asthma, peptic disease, low back pain, malaise and injuries (Muralidhar et al 2017, van Puijenbroek et al 2019).

We collected water samples from the Turag River and Tongi Khal as part of two sub-studies: (1) monthly monitoring of 21 physiochemical parameters at 13 sampling points along the river from February 2018 to January 2019 (figure 1); (2) analysis of 18 heavy metals and biotoxicity at 22 sampling points in December 2017 and January 2018 (figure 3). Samples were collected from a boat at 2m depth, followed by field analysis of temperature, pH, dissolved oxygen, oxidation reduction potential, electrical conductivity, and total dissolved solids using HACH HQ40d multiparameter device and turbidity with VELP Scientifica TB-1 portable Turbidimeter.

Laboratory analysis of colour, alkalinity, dissolved organic carbon, ammoniacal nitrogen, nitrate, phosphate, iron (ferrous), chloride, sulphate, and sulphide ions, and pathogens (total coliforms and E. coli) were conducted at Bangladesh University of Engineering and Technology (BUET). Concentrations of heavy metals were analysed at the Department of Earth Sciences, Oxford University through Inductively Coupled Plasma Mass Spectrometry (ICP-MS) using a PerkinElmer NexION 350D ICP-MS. The heavy metal data was previously used for a separate study by Oxford Molecular Biosensors, which developed bacterially derived sensors to detect the bioavailability of specific chemicals and their overall cell damage toxicity (refer to Rampley et al (2020) for further details). Speciation analysis of few samples during the Ijtema preparatory phase in January 2018 were done using Visual Minteq ver. 3.1 (May 2020 Edition).

We conducted a household survey in December 2017, in collaboration with the University of Dhaka, to collect baseline data on socio-economic and demographic profiles, indicators of multidimensional poverty, the state of water and sanitation facilities, itemised household expenditures and general development concerns. The survey was conducted in Bangla by 15 trained local enumerators using a semi-structured questionnaire on ONA, a mobile data collection platform (https://ona.io/). We covered all settlements within 1km buffer zone along the river, using a random walk method to select every 10th household within each settlement. If no consenting adult was present at the time of visit, the next household was chosen instead. A total of 1826 households were surveyed, of which 446 households were in Zone-1, 730 in Zone-2, 510 in Zone-3, and 140 in Zone-4. Data analysis involved generation of zone-wise descriptive statistics and calculation of a multidimensional wealth index through principal component analysis (PCA) of 11 variables on housing materials, education, and asset ownership (see table B1 in appendix (available online at stacks.iop.org/ERC/3/095003/mmedia)). A k-means cluster analysis was then applied to the factor scores of the first principal component (PC₁) to disaggregate the households into four wealth classes.

We designed and implemented a 'structured direct observation' study to monitor the gender- and age-disaggregated daily river water use practices in relation to the spatial and temporal variations in water quality risks. Direct observation is an established method in social science research whereby the researcher uses a pre-designed questionnaire to collect standardised quantitative information on the research subjects in their usual environment without any alterations. Early examples of direct observation can be found in the medical anthropology literature, where researchers observed the patterns of human contact with water bodies to evaluate the pathways of schistosomiasis transmission (Dalton 1976, Dalton and Pole 1978, Slootweg et al 1993) with more recent applications is monitoring handwashing behaviour in rural South Asia (Halder et al 2010, Ram et al 2010).

We conducted our study in two phases. The first phase was carried out over an 18-day period in the dry season (9–26 Feb 2019) which coincided with the Bishwa Ijtema held in two groups of three consecutive days. The second phase was conducted for a 15-day period in the wet season (20 Aug–3 Sept 2019). We selected ten observation points, two of which were in Zone-1, seven in Zone-3, and one in Zone-4. The site selection involved multiple scoping visits to the household survey areas to identify spots with observable river use activities, as well as ensuring spatial distribution, diversity of river interactions, accessibility and security for field team, and alignment with river water quality monitoring points. Zone-2 was excluded as we did not find any interactions with the river during our scoping visits, as the short river branch flowing through the bottom part of Zone-2 remains dry for part of the year. During analysis, as shown in figure 1, we merged results from the ten observation points into six sites in three zones, namely, Konabari (Zone-1), Ijtema field, Tongi slum, Abdullahpur and Railway Bridge (Zone-2) and Mausaid (Zone-3).

Each observation day comprised three 3-hour slots: 7–10 am (morning); 10.30–1.30 pm (midday); and 2.30–5.30 pm (afternoon) during which enumerators recorded their observations in an electronic form in a tablet, whereby an 'observation' is defined as any activity conducted by an individual or group visible within the enumerator's field of view (refer to appendix C for detailed observation schedule). Thus, each slot comprised several observations, with each observation including one or more activities. Over the 33 study days, we recorded about 7900 observations for 852 slots. River users were visually categorised as children and adolescents (<16 years of age) and adults, as male and female, and also as groups and individuals. Differentiation of age was based on enumerators' judgment; some adolescents and young adults may have been misclassified. Activities were listed as drinking, food washing, water collection, dish washing, laundry, washing oneself, bathing, using hanging latrine, open defecation, urination, fishing, swimming, boating and other. These activities were listed as multiple-choice questions, as multiple tasks were often conducted by different individuals in a group. Enumerators were given clear guidelines for certain activities; for example, if fishing was done on a boat, the activity was recorded as fishing and not boating. Enumerators also took a photo for each observation, following approved ethical guidelines, which were inspected by the lead author to resolve issues in data entry.

Monthly analysis of water quality along the Turag River and Tongi Khal shows extreme spatial and temporal variation in physiochemical parameters and heavy metal concentrations, with Zone-3 and Zone-4 exhibiting the worst conditions (figure 2). During the dry season, DO levels are so low that the waters and sediments may be presumed to be anoxic, with some metals from bound sediments and gases such as methane, ammonia, and hydrogen sulphide being released. Organic pollution, as indicated by the total organic carbon (TOC), is highest in January owing to the dumping of faecal and food waste in the Tongi Khal during Ijtema. TOC reduces as monsoon approaches, albeit a spike in May from dredging activity undertaken during the sampling period. Increase in temperature and DO, coupled with addition of different genre of coliform bacteria from the monsoon runoff, also reduces the relative proportion of faecal coliform (E. coli) to total coliform. Likewise, the concentrations of ammoniacal-nitrogen, nitrates, and phosphate decreases to below national standards in monsoon, except for the spike in nitrates in June from rainwater accumulating on the surface during the sampling. There is, however, diffuse runoff from the waste tips and polluted soils adjacent to the rivers and, during storm events, there is significant local storm runoff from urban and industrial areas into the streams and rivers, which is evident from elevated turbidity levels in rivers throughout the monsoon season.

Zoom In Zoom Out Reset image size

The concentrations of heavy metals increased from December 2017 to January 2018 in general (figure 3) due to combined effects of decreased river flow, resuspension of sediments caused by the disturbance of riverbed during Ijtema preparatory work, and subsequent dissolution of metals due anaerobic condition. This sediment disturbance continues throughout the Ijtema event with addition of different metals, such as aluminium, chromium and iron from washing utensils, selenium from the likely consumption of medicines, lithium from battery disposal, and manganese from fish residues. Chromium, lithium, lead and selenium contribute to cellular toxicity to most microbes, while copper shows toxicity to bacteria and algae in mostly free ion form.

Zoom In Zoom Out Reset image size

The heavy metal concentrations presented here, however, indicate the total concentration of each metal instead of their availability as free cations. Most metals are bound in organic and inorganic complexes with dissolved organic matter, sulfide and ammonia dictating complexation. Chromium and copper bind almost exclusively to organic complexes, and aluminium, manganese, cadmium, lead, selenium and lithium all show considerable inorganic complexation to oxyhydroxides, chloride and ammonia. Lithium, selenium and zinc also tend to be in free ionic forms. A speciation analysis of samples in January 2018, through Visual Minteq 3.1 confirmed high levels of organic complexation for copper and chromium near Ijtema sites. Similar organic complexation was found for most metals except for selenium and lithium which had free ions as dominant species. A parallel study using this dataset and involving the co-authors of this paper found that high level of cellular toxicity was particularly associated with increased concentrations of selenium, zinc and chromium (Rampley et al 2020). Hence, despite the overall increase in metal concentrations, the simultaneous dumping of organic waste results in lower availability of free ions, which in turn lowers the toxicity in the samples collected during the Ijtema.

Zone-1 comprises land privately-owned by a few families descending from early settlers, as well as long-term lease holders, who built and rented out houses to families and individuals working in nearby brick kilns, garment factories, and as day labourers. In comparison, settlements in Zone-3 are on government-owned land, with most tenants and homeowners lacking legal ownership, putting them at risk of eviction. Our survey reveals a high clustering of poverty in Zone-3, with 41% of the households belonging to the poorest wealth class, compared to 23% and 18% in Zone-1 and Zone-2 respectively (table 2). About 20% of adult women in Zone-3, equivalent to half of those in paid employment, work in one of the many garment factories nearby, while the men are engaged in small businesses (22%), casual labour (22%) and garment factories (19%). Zone-4 is a relatively well-off area, with 66% in the top wealth quartile and having the highest proportion of adults working in the service sector (14% men and 6% women). The average household monthly expenditure in Zone-4 is BDT 25,500 (USD 307), which is significantly higher than Zone-1, Zone-2 and Zone-3, with mean expenditures of BDT 13,500 (USD 162), BDT 16,300 (USD 196), and BDT 14,700 (USD 177) respectively. When disaggregated by wealth class, the average monthly expenditure ranges from BDT 12,000 (USD 145) for the poorest to BDT 25,000 (USD 301) for the richest class across all zones (Table B3 in appendix).

Households in Zone-1 and Zone-2 access water from motorised tubewells installed by NGOs, either through pipelines drawn inside their dwelling or through shared taps in communal spaces. Those living on the southern bank of the Tongi Khal in Zone-3 have access to piped water from DWASA, whether legally or illegally, while people on the northern bank rely on community tubewells. While they have access to an improved source within a few minutes of walking distance, water is often only available at certain times during the day, creating long queues for a limited quantity that is prioritised for drinking and cooking. About 68% of the households in Zone-3 reported sharing their water source with at least five other households. Households in Zone-4 usually have their own tubewells, with only 25% sharing it with 5–10 households in the same compound.

Although none of the households are connected to formal sewerage infrastructure, 79% households in Zone-4 use a flush toilet with a septic tank and 86% do not share their toilet with others. The sanitation conditions are worst in Zone-3, with 27% reportedly using hanging latrines or practicing open defecation, with the faecal waste discharging into the river. About half of the households in Zone-3 share their toilets with at least five households, suggesting the lower living standards and increased likelihood of spreading pathogens. Improved pit latrines are more common in Zone-1 (68%) and Zone-2 (45%), followed by flush toilets with septic tanks (19% and 41%).

The variations in living standards are also reflected in the concerns expressed by the surveyed households. For households in Zone-3, clean environment (25%) ranked as the topmost concern, followed by water services (17%) and sanitation (16%). In comparison, households in Zone-1, Zone-2 and Zone-4 prioritised roads and transportation (35%, 33% and 37% respectively), followed by healthcare services or gas supply. However, when surveyed households were specifically asked about their environment-related concerns, 50%–60% across all zones mentioned the river being dirty as their main concern.

The types and intensities of river use activities were closely linked to the socio-demographic contexts of the studied populations and the water quality. Direct contact with river water, through dish washing, laundry, cleaning fish and vegetables, and personal washing was high in Zone-1 and Tongi slum in Zone-3 (figures 5 and 6), with averages of 14 and 15 people observed per 3-hour slot during the dry season, of which 10 people in each site were women and girls (Table C3 in appendix). These sites were close to residential clusters, where overcrowding at community water points and restricted supply only at certain times during the day, led to high usage of river water for domestic activities. We also observed several boat dwellers along the Tongi Khal in Zone-3, who used the river for almost all purposes other than drinking (figure 4). Domestic activities were relatively fewer in Zone-4, owing to the better socio-economic status of residents in this area.

Zoom In Zoom Out Reset image size

During the wet season, we observed a slight decrease in domestic activities in Zone-1 and Tongi slum in Zone-3 (figure 5). In Zone-1, construction of new houses along the bank which partially deterred the local residents from accessing one of the observation sites, while in Zone-3, around 60 houses were evicted as part of the government's drive to control river encroachment. However, we recorded a steep increase in swimming and bathing activities in Zone-1 and Tongi slum in Zone-3, which changed the overall observation counts to 11 and 27 people observed per 3-hour slot. Recreational swimming was more prevalent among adult men and male children, as cultural norms sometimes refrained women from sharing the same public space with men for recreation (figure 6). This increased river interaction by men, along with the reduction in women carrying out domestic activities, led to equal proportions of both genders observed per 3-hour slot. The rise in bathing and swimming, which was also prevalent in other sites, can be attributed to the warmer weather and perceived improvement in water quality.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Small-scale productive uses were also observed at some sites in Zone-3. These included washing and dyeing denim, washing fish baskets or plastic sheets, collecting plastic waste and fishing. Denim washing was mostly carried out by men, under the pillars of the railway bridge on the Tongi Khal (figure 4). During the wet season, the platforms of the pillars were flooded which resulted in a decrease in this observed activity, which likely shifted to another site not included in our study. Informal waste pickers could be spotted wading through the river on a boat, collecting plastic bottles for resale while others were seen washing plastic sheets either on the banks or in waist deep waters. These indigo tainted sheets are waste products from dye packaging, which serve as an income source for these marginalised citizens. One of our observation sites in Zone-3 was near the local fish market, where traders regularly used the river to wash their fish and the fish baskets. Fishing, with or without a boat, was commonly observed across all sites during the wet season. In Zone-1, the abundance of fish along the banks meant that women and children could easily catch these with their bare hands.

The river was heavily used for boating, mainly for transportation and sometimes for recreation. Boating increased significantly in Abdullahpur and Ijtema field in Zone-3 and in Zone-4 during the Ijtema period in February 2019 for transportation of people, food and construction materials. While temporary sanitation facilities with faecal sludge containment was provided for the Ijtema devotees, urinating into the river was commonly practiced, along with the disposal of organic food waste (figure 4).

Regardless of sampling location and season, the rivers have severe pathogen pollution, as indicated by the faecal coliform count that is several hundred times higher than the recommended national and international standards of <200cfu/1000ml for bathing and swimming (MoEF 1997, UNEP 2016). Swimming and bathing activities, which increase in monsoon, are likely to cause infectious diseases from accidental ingestion and contact with mucous membranes in eyes and ears. Results from a Quantitative Microbial Risk Assessment across three rivers in southwestern Bangladesh show that the overall risk of illness for a single exposure of pathogens during bathing ranged from 9%–19% for children and 7%–16% for adults, which exceeds the USEPA acceptable risk of 3–6 illnesses per hundred bathing events (Islam and Islam 2020). Similarly, Hamner et al (2006) found significant associations between the use of the Ganges River in Varanasi, India for bathing, laundry, and washing utensils and the incidence of water-borne enteric diseases, with poor sanitation, low income and low education as confounding factors.

The bioaccumulation of heavy metals in fish and vegetables have also been documented in previous studies (Ahmed et al 2016, Rashid et al 2017). Our observations show that fishing activities are prevalent in the monsoon season, indicating potential consumption of accumulated heavy metals by families consuming fish and nearby populations through the sale of caught fish at neighbourhood markets. Our study shows that women in low-income settlements are regularly exposed to contaminated water, even during the dry season, through washing clothes and dishes. Similarly, informal waste recyclers, standing in waist deep water to wash dye packaging, and casual workers washing denim in the river, are risking exposure on a daily basis. While Hasan et al (2019) found significantly higher concentrations of chromium in nails and hair samples of tannery workers, there has been no similar research on bioaccumulation of heavy metals in these riverine dwellers, despite the measured cytotoxicity potential of the river water. Systematic observation of human interaction with polluted river water is unprecedented, providing opportunities for adapting our method to quantify health risks. Recent studies also show high likelihood of heavy metal exposure through dermal contact, the risk being dependent on exposed body surface area and duration of activity (Zhao et al 2019). However, these assessments of 'total human exposure' are based on retrospective snapshots of human activity patterns from surveys, which are prone to recall and social desirability bias. In such cases, our method can be adapted in the future to include recordings of the proportion of body surface exposed and duration of activity.

The riverbank settlements studied here are home to thousands of factory workers and their families, living in dense, unsanitary conditions that they afford with their low wages. Though the aggregate benefits of economic growth are enjoyed by the wider population, the negative externalities of unregulated industries and unplanned urbanisation are disproportionately borne by those directly contributing to the growth process. While rural migrants are often habituated to use surface water for domestic activities, the lack of adequate water services and the resulting overcrowding at community waterpoints are key drivers of river usage for washing and bathing purposes. Such unequal water-citizenship in Dhaka slums 'is a daily reminder to the urban poor of their abjection and undesired status in the city' (Sultana 2020, p.1413). While aggregate statistics by JMP (2019) suggest that open defecation is non-existent in urban Bangladesh, the high prevalence of hanging latrines, as observed in this study and previous work by Haque et al (2020), reveals that the challenges of marginalised communities may be masked or ignored in the bigger picture.

While the export-oriented economic growth agenda has dominated the Government of Bangladesh's political and institutional approach since the 1990s, actions are being taken to curb river encroachment, reduce industrial and municipal pollution, and regulate water use, with the High Court bestowing rivers with the same rights as 'legal persons' in July 2019 (Islam and O'Donnell 2020). The ongoing construction of new sewerage treatment plants under the 'DWASA Masterplan 2011', and the recently approved 'Dhaka Sanitation Improvement Project', which includes plans to upgrade toilets and provide communal septic tanks in low-income communities (The World Bank 2020), are two of the planned large-scale interventions to reduce municipal pollution. The Water Resources Planning Organization (WARPO) in collaboration with United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP) is drafting the Industrial Water Use Policy (UN ESCAP and WARPO (2019)), including the Greater Dhaka Watershed Restoration project to promote industrial wastewater treatment and reuse (2030WRG 2017). The relocation of the tanneries has been another major intervention, though in absence of a functioning centralised effluent plant, the pollution has merely shifted from the Buriganga to Dhaleshwari River (Hasan et al 2020).

The Government of Bangladesh has ambitious goals to dramatically increase ready-made garment exports; however, no equivalent bold environmental policies have been enacted to counter the potential harm from increased wastewater (Hossain et al 2018). Government plans to achieve SDG 6.3 tend to rely on the technocratic logic of dominant institutions, driven by physical interventions in manipulating river morphology through dredging (General Economics Division 2015: p. 423). Industry efforts, largely pursued by international brands sourcing from Bangladesh and donor institutions such as the World Bank and International Finance Corporation, focus on creating zones of 'cleaner production' that end up excluding the worst offenders that supply only to the domestic market from external scrutiny (Selim 2018).

However, even if all these initiatives are executed effectively, the portfolio of engineering, judicial, and institutional initiatives is likely to take years to translate into reduction of pollution risks for the most vulnerable. In the short-term, responses to curb pollution exposure, through provision of safely managed water and sanitation facilities and approaches like risk communication can minimise health risks of those living in riverine communities.