Heterogeneous water provision in Dar es Salaam: The role of networked infrastructures and alternative systems in informal areas

Introduction

Urban waterscapes in developing world cities are notoriously fragmented, resembling archipelagos with islands of service rather than continuous networks (Bakker, 2003; Kjellén, 2006). Graham and Marvin’s (2001) concept of splintering urbanism offers one explanation by linking this service fragmentation to the collapse of integrated infrastructure networks over the past few decades. They stress the fundamental importance of these networks in creating modern, cohesive, and equitable cities. Yet Graham and Marvin have been criticized for applying Northern theory and experiences to Southern cities. Kooy and Bakker (2008) note that these ideas of splintering urbanism assume that the presence of modern infrastructure networks such as universal piped water networks is a stage of development. Conversely, the ideas also assume that the absence of these networks is a sign of failed modernity where the rich are integrated into the network while the poor are excluded creating landscapes of spatial apartheid. In their case study of Jakarta, Kooy and Bakker (2008) instead describe a pattern of splintered urbanism that has been a constant in the city since the colonial era rather than a recent phenomenon of splintering.

One way to situate this splintered urbanism is through urban political ecology (UPE), which provides a framework to understand how social, historical, political, and ecological processes work together to create uneven and unjust landscapes and waterscapes. UPE has been especially useful in considering the links between injustice and the colonial legacies of unequal water provision in developing world cities. Present-day problems are traced to the colonial development of piped water networks that supplied urban residents unequally (for example see Bell, 2015; Gandy, 2008; Kooy and Bakker, 2008). Inequalities between urban fringes and the urban core also contribute to these unjust landscapes as these areas are not fully incorporated into political systems (for example see Bontianti et al., 2014; Ranganathan, 2014). These water inequalities and injustices are explained through unequal power relations. Ranganathan and Balazs (2015) suggest water access has both proximate dimensions in regard to quantity and quality but also processual dimensions from multi-scaled and historical power relations that result in differing benefits from water.

A further critique of splintering urbanism is that Graham and Marvin privilege the modern infrastructural ideal and do not acknowledge the importance of non-networked utilities in developing world cities. Bousquet (2010) points out that while splintering urbanism sees alternative providers as reinforcing fragmentation and reducing cohesion in the developed world, in the developing world they may be a benefit. In cities where the network does not reach everyone, these alternative suppliers help to realize the spatial equity of water provision (even if they do not achieve equity in price). At the same time, however, these alternative systems of delivery can contribute to fragmentation through the process of splintering from below (Kooy and Bakker, 2008).

The unfair privileging of modern technology has been highlighted by scholars who acknowledge and even celebrate the heterogeneous nature of water provision in cities. Even though some people use these alternatives because of a lack of network access, others purposefully choose them over the network (Furlong and Kooy, 2017). This idea of urban heterogeneity acknowledges that piped networks have often failed in Southern cities, but this does not mean that the cities are failures. Instead it is an opportunity to recognize the “vitality and multiplicity of actual delivery systems which … survive and contribute to the functioning of cities” (Jaglin, 2014: 434). Thus, it is useful to think of these diverse delivery systems as possibilities rather than only as problems (Lawhon et al., 2018). For example in Quibdó, Colombia rather than working to expand the piped water network, the network coexists with other alternative forms of water provision, especially rainwater collection (Furlong, 2015). Likewise, Truelove (2019) describes Delhi’s gray zones that defy dualistic notions of formal/informal and instead “account for the multiplicities of water regimes beyond that network that profoundly shape the uneven experience of urban space” (15). It is important to see these alternative delivery systems as filling a service gap in developing cities but also to not romanticize these heterogeneous services (Lawhon et al., 2018). They help to serve the vast oceans between the archipelagoes of networked service provision, but they often come with a financial or quality cost.

In presenting two case studies of water provision in informal areas, this paper builds on the concept of intra-urban comparisons put forth by McFarlane et al. (2017) and seeks to highlight the diversity within cities and in urban life. They conceive of cities not as singular spaces but instead as spaces inhabited by “many urban worlds” (1394) and caution against oversimplifying cities and instead push to examine the complexity, divergent experiences, varied experiences, heterogeneity, and contradictory trajectories of cities. Simone (2010: 42) highlights these same ideas, noting that “districts that would on the surface appear to look the same and be subject to the same social and economic forces may end up with very different futures.” McFarlane and Silver (2017) stress that both the challenges of urban life—including securing water—and their solutions are radically different even within a single neighborhood and note that effective policy and scholarship requires a grasp of this spatial differentiation.

The two areas of Dar es Salaam discussed here—Buguruni and Vingunguti—share similar origins but their development trajectories have diverged since then. Colonial legacies of uneven development are an important part of understanding piped water provision in Dar es Salaam, but the story is more complex than a dichotomy of former European and former African residential areas. This comparison contributes to UPE with this situated small-scale approach. It also furthers the discussion on the modern infrastructural ideal in several ways. It agrees that a lack of a network does not mean failed development and documents the value—and necessity—of water supplied beyond the network. Yet it also acknowledges the desires of some Dar es Salaam households to connect to this network. This intra-urban comparison of two different informal areas shows that despite similar beginnings, these areas have diverged to become distinct spaces.

This paper approaches networks through Cleaver’s (2001, 2012) ideas of institutional bricolage, the process of mixing formal and informal as well as modern and traditional systems of natural resource management. She notes the false dichotomy of modern versus traditional and does not suggest one is better than the other but rather advocates for a complex, innovative, and adaptive institution. Originally applied to resource conflicts in Tanzania’s Usangu Basin, Liddle et al. (2016) use these ideas to demonstrate the value of informal water supplies alongside formal in urban Zambia. In the context of water supply in Dar es Salaam, these networks take various forms. At their most basic level, these are physical networks that transport treated water through Dar es Salaam. Yet social networks are an important part of water provision. Those households without a direct connection to the city’s piped water network must rely on others to secure water including by purchasing resold water by the container or negotiating a piped connection to a neighbor’s private well. People also are themselves infrastructure and fill gaps in the physical network by delivering water (Simone, 2004).

Inequities in Dar es Salaam’s water provision

Like other developing world cities, the fragmented and heterogeneous nature of Dar es Salaam’s water network dates to the colonial era. Both the German (1887–1919) and British (1919–1961) colonial administrations planned a racially segregated city. Although the administrations never enacted a legal policy of segregation by race, they used building ordinances and construction codes to achieve the same result. For example, the zone that ultimately became the African residential area had no specific building requirements and allowed for native-style construction while the zone that became the European area required large sturdy homes with flush toilets (Brennan, 2007; Smiley, 2009). The rationale behind these construction codes was that strict building standards would ensure only Europeans could afford to live in that specific zone. Underlying this rationale was the belief that only Europeans would want those types of toilets.

Beyond having different types of homes constructed from different building materials, these areas also had varying levels of service provision. The city’s piped water system initially served only the European residential neighborhoods and the commercial area and distributed groundwater extracted from boreholes (Kjellén and Kyessi, 2014). Importantly, the European homes received direct household water connections to the piped network. Over time, the British administration constructed small groundwater distribution schemes that were later connected into a modern piped network (Kjellén and Kyessi, 2014). Eventually these systems shifted from supplying groundwater to treated river water. The British opened the Mtoni Water Treatment Plant on the Kizinga River in 1952 and the Upper Ruvu Water Treatment Plant on the Ruvu River in 1959.

Yet it was not until the 1950s that water service expanded beyond the European and commercial areas (Rugemalila and Gibbs, 2015). Once the system reached African residential areas, those households did not receive direct connections. Instead the colonial administration provided these areas with shared public water kiosks. Since the piped network delivered water to the kiosks, it would have been possible to directly connect the homes, but the government chose to rely on the cheaper shared kiosks. As Hungerford and Smiley (2016) note, “ultimately, British colonial discourse on water supply in Dar es Salaam questioned whether better water supplies were needed for Africans, whether better supplies were deserved, and whether the expenses of providing such water access were justified” (77).

Water inequalities did not end in the post-colonial era. The piped water network saw few meaningful expansions during that time. In 1975, a third water treatment plant was added—the Lower Ruvu Plant on the Ruvu River. This limited piped water supply occurred in spite of rapid population and spatial growth in the city. Between 1967 and 1978, the city had an annual population growth rate of 10% and it is now the third fastest growing city in Africa (Briggs and Mwamfupe, 2000; McGranahan et al., 2016). In addition to inadequate expansions in the quantity of water, few expansions occurred to the network itself so it did not extend to newly built-up areas. Although all three water treatment plants have been rehabilitated since their initial construction to increase production capacity, they still do not produce enough water to meet current or future demand. To overcome this water supply deficit, the Dar es Salaam Water and Sewerage Corporation (DAWASCO) rations water delivery so that not all areas of the city receive water at the same time; unsurprisingly the informal and low-income areas receive the most rationing.

These continuing inequities contribute to the city’s heterogeneous water supply, with some areas receiving DAWASCO water primarily through direct household connections, other areas receiving DAWASCO water indirectly (from kiosks or household resellers), and still other areas lacking any DAWASCO water service at all. Those areas lacking service entirely must rely instead on groundwater or on vended water delivered from other parts of the city. Groundwater helps to fill two gaps in the city’s water supply. Not only does DAWASCO supplement its treated river water with groundwater, but it is also an economical source of water for households. The reliance on groundwater, however, is problematic as high levels of extraction have lowered the water table and led to saltwater intrusion in aquifers.

As Kjellén (2006) notes, water access in Dar es Salaam is determined both by location and income. Some middle- and high-income areas are not connected to the piped network because of their distance from the city center. Their incomes, however, allow for the installation of large underground storage tanks that are filled with DAWASCO water delivered by tanker trucks. An undercurrent of the inequitable water supply relates to electricity. The country’s electricity provider does not produce enough power to satisfy demand, leading to regular interruptions in service. These interruptions impact the supply of both groundwater and piped treated water. Electricity is needed to operate pumps at wells so when there are power cuts, households lose water access. The city’s two water treatment plants on the Ruvu River have dedicated electricity supplies but still experience interruptions; each cut requires several hours to resume water flow and up to a day to regain full pressure (Smiley, 2016).

Still while geography and history are important in understanding Dar es Salaam’s water inequalities, politics also contributes to these inequalities. Tanzania is officially a multi-party democracy, but the Chama cha Mapinduzi (CCM) party has dominated the political sphere since independence. Although opposition parties have made some gains in recent elections, CCM holds a supermajority in Parliament and in locally elected positions. More than just occupying offices, CCM exerts control over citizens, a phenomenon that Campbell (2009) calls a standard grid. Although he suggests that citizens may attempt to challenge the grid, including through the informal economy, their abilities to do so are limited. Corruption is an issue in Tanzania, including in Dar es Salaam’s water supply where households have been asked for additional payments in return for piped connections (Smiley, 2019).

Carlitz (2017) describes the role of the central Tanzanian government in allocating funds to local authors for rural water provision. These local authorities often display political favoritism, providing areas with more support for CCM with more water infrastructure; wealthier and better-connected communities also tend to receive more infrastructure. This favoritism occurs despite Tanzania’s Water Sector Development Program that sought to increase access to clean water, promote decentralization, and encourage public participation. Decentralization should make service delivery more efficient but instead creates more opportunities for this local-level corruption. Bourque (2010) describes how informal water providers in Buguruni shifted allegiance to an opposition party since they felt CCM was not addressing local needs. This party change came with consequences as CCM limited its assistance to the area.

This paper compares water provision in two areas of Dar es Salaam that experience these different types of water service. One contains more DAWASCO connections while the other relies almost exclusively on alternative delivery systems. Certainly this paper does not suggest that having a household connection guarantees access or that a connection is necessarily preferable to resold or vended water. One impediment to access is the existence of non-functioning piped connections. Those households with dry pipes—whether caused by deteriorating infrastructure, rationing, or disconnection for non-payment—do not have real access.

Methods and study sites

This paper uses 106 household surveys—49 from Buguruni Ward and 57 from Vingunguti Ward—conducted in October and November of 2017. These surveys focused on aspects of water access, specifically the primary type of water used, its delivery, the amount of time involved in collection, water availability, and drinking water source. Households were purposefully selected using convenience sampling to ensure an adult head of household respondent was home. Efforts were made to ensure spatial coverage of each ward. In Dar es Salaam, wards are divided into smaller administrative units, and surveys were conducted in each of these subwards (four each in Buguruni and Vingunguti). All respondents provided verbal informed consent, and surveys were conducted in Kiswahili with the assistance of a research assistant from the University of Dar es Salaam. This project received ethical approval from the researcher’s university and the Tanzanian Commission for Science and Technology.

It is important to acknowledge some challenges with data collection and research in Tanzania. A recent law restricts the publication of statistics that disagree with official data. This shift might make government officials and even private citizens cautious to share information. Although I do not believe that households provided false information, it is certainly possible that some were reluctant to offer criticisms of their water supply. A specific challenge for this project is obtaining details on the piped water network. Although DAWASCO has a mapping office, it does not currently make its data on the location of kiosks or pipes publicly available. A few older maps do exist, including one in Bayliss and Tukai (2011) that shows Buguruni with only a few DAWASCO pipes and Vingunguti with none, but these do not reflect recent service expansions although they do provide a general idea of inequitable service. Thus while a current and accurate map of water pipes or kiosks for the study areas would enhance this paper, the lack of available data precludes its inclusion.

Both Buguruni and Vingunguti share some characteristics so were purposefully selected for this project. They are located in Ilala District, with Vingunguti bordering Buguruni to its southwest (Figure 1). These areas are centrally located, with Vingunguti located just 6 kilometers southwest of the City Center (Bayliss and Tukai, 2011). According to the 2012 National Census, Buguruni Ward had a population of 70,585 people, and Vingunguti Ward had a population of 106,946. Both wards are considered informal settlements, and Lugalla (1997) characterizes them as overcrowded high-density residential areas that house the urban poor. In both areas, residents lack formal employment and there are high levels of unemployment (Kombe et al., 2015; Lugalla, 1997). These wards are also characterized by poor housing conditions and high degrees of socio-economic inequalities, especially in education and health (Awinia-Mushi, 2013). Buguruni is the older of the two areas, having been settled in 1921. At first Vingunguti was considered a rural extension of Buguruni, but its prime location along a major road and proximity to industries drew immigrants from other areas of the city (Kombe et al., 2015).

OPEN IN VIEWER

Figure 1.

Location of Buguruni and Vingunguti Wards. Source: Nathaniel Lee, University Libraries.

Although these wards play similar roles in Dar es Salaam’s political economy in terms of the importance of informal employment and the presence of some industry, they differ in urban form. Vingunguti feels remote and neglected. It may be centrally located within Dar es Salaam, but it is still marginalized. Ranganathan and Balazs (2015) suggest that the urban fringe is bound to the city both in spatial and relational contexts. They continue that fringe areas are physically proximate but their “position of exteriority from that city for all practical purposes, means that everyday life is saturated with anticipation and hope, as well as exasperation and despair about the prospects for improved conditions” (409). This marginalization is felt when traveling to Vingunguti. It is necessary to pass through Buguruni to reach it by public bus, and there is a noticeable change from a paved main road in Buguruni to an unpaved one in Vingunguti. Vingunguti previously housed the city’s landfill and is known for environmental contamination from the many abattoirs located there. Its marginalization might best be illustrated by an observation made during this fieldwork. I watched several tanker trucks marked maji taka (Kiswahili for sewerage) drive to a large pool of standing water. Although they did not empty their tanks in my presence, the smell of the pool suggested that was the intent.

Lugalla (1997) ties many of the present-day problems in these wards to the legacies of their colonial development. As part of the de facto African residential area, they received few services and relied on shared public water kiosks. Eventually the city transitioned away from the zones based on construction codes and building standards into more formal planning. Under three master plans (created in 1949, 1968, and 1979) the city was divided into high-, medium-, and low-density areas. The change in terminology did not change the reality of these areas; as high-density areas, Buguruni and Vingunguti continued to rely on communal public kiosks for water provision, which were deemed sufficient to fulfill the needs of the area (Armstrong, 1987).

In spite of the similarities in their origins, the waterscapes of Buguruni and Vingunguti have taken divergent paths in more recent times. Buguruni’s waterscape has improved over the past three years, including increased DAWASCO service hours and increased numbers of DAWASCO household connections (Smiley, 2019). Prior to that time, however, it had few household connections and very limited service hours (Bourque, 2010; Rewetta and Sampath, 2000). Its residents were forced to find alternative water sources but Bourque (2010) notes that there were few places to purchase well water and as a result, many residents relied on vendors to deliver water sourced from other areas of the city. Vingunguti, on the other hand, has extremely low levels of DAWASCO water supply, both in service hours and number of connections. Yet unlike Buguruni, Vingunguti’s piped supply was better in the past than it is today; its DAWASCO service became unreliable in the 1990s as growth elsewhere in the city and aging infrastructure negatively impacted water provision (Bayliss and Tukai, 2011; Wandera, 2000). The DAWASCO water that is provided there today is limited in hours and has very low pressure (Kombe et al., 2015). On paper the similar characteristics of these areas might suggest comparable levels of service provision, but in reality they have substantial heterogeneity in water provision.

Heterogeneous water access in Dar es Salaam

Many of the differences between water access in Buguruni and Vingunguti can be at least partially attributed to the primary source of water used by households. All surveyed households in each ward use either DAWASCO networked water or water from a protected well as their primary source. Although some households in Dar es Salaam do use unprotected wells or surface water, none of the surveyed households do so. Yet while these surveyed households are using just two types of water, the proportion of sources is very different. In Buguruni, most households use DAWASCO (59%) while in Vingunguti most households use well water (91%). Furthermore, not only are households in these wards using different types of water, but they also obtain this water in different ways. Fetching is the most common delivery mode in both wards, but many households do have direct household connections. In Buguruni, nearly all of those connections are to DAWASCO (87% of all piped connections) while in Vingunguti nearly all are to wells (82% of all piped connections).

Beyond different types of water, these piped connections differ in their levels of formality. All of the DAWASCO connections are formal connections obtained directly through the water provider. In Vingunguti, most of the connections are to privately owned and operated wells. In these cases, the owner allows their neighbors to connect to a micro-network for a fixed monthly price. One additional difference between these two areas is the more visible presence of local government-run wells in Vingunguti. Households both fetch from these wells and obtain piped connections to their micro-networks. Since CCM has ruled Tanzania since independence, the presence of government sources may help to reward supporters.

It is these differences in primary source type and delivery method that help to explain four further water access inequalities between these wards. First, the collection time for water from the households’ primary source varies. Ideally, households should have a source of water located directly on-premises; if they must fetch from an external source, the total collection time should take no more than 30 minutes including round-trip walking, waiting, and the filling of the container (WHO and UNICEF, 2017). In Buguruni, none of the households that fetch water indicated a total collection time over 30 minutes while in Vingunguti five households did, and all were attributed to long queues. As both are high-density areas, the central explanation for these differences is the total number of sources. More direct DAWASCO connections in Buguruni put water closer to home for fetchers. New wells are not being dug at the same speed in Vingunguti so more people must use fewer sources, which leads to longer queues.

Second, the availability and reliability of water differs. In Tanzania, the water regulatory body identifies 24 hours as the ideal amount of service, but this remains an elusive goal (EWURA, 2017). A minority of households (31% in Buguruni and 16% in Vingunguti) indicated that they had 24 hours availability for their primary source of water. The differences become more pronounced when also considering households that are only without water for a few hours a week (49% versus 21%, respectively). Buguruni receives more reliable water provision, largely because of the increased DAWASCO production that has lessened the need for water rationing. The wells in Vingunguti are limited by electricity outages which affect pumps. As noted earlier, inadequate electricity impacts water access. Although power interruptions have decreased over the past few years, they still occur and do not impact all areas equally. Households in Vingunguti note more frequent cuts than those in Buguruni. Water availability is also dictated by the well owner who sells or distributes water on their own schedule. As a result, most households in Vingunguti indicate that their water is available only for certain limited hours each day.

Third, the wards use different sources of drinking water. Issues of collection time, availability, and reliability specifically asked about the primary source of water. In Buguruni, most households (69%) also use this primary source for drinking; the other households either buy DAWASCO from a neighbor or purchase bottled water. In Vingunguti, however, the primary source is reserved solely for household purposes such as laundry, cleaning, bathing, and cooking; there a minority (25%) also drink from their primary source. This number does include five households that use DAWASCO as their primary source (four from piped connections and one from fetching).

All of the remaining households in Vingunguti use well water for the primary source but purchase DAWASCO specifically for drinking. Although most of this DAWASCO drinking water is purchased from mobile vendors, three households obtain their DAWSCO drinking water from an unusual source: their own pipes. These households have direct piped connections to DAWASCO water but only use that water for drinking; for their primary water source they all fetch well water because it is cheaper and more reliable than their DAWASCO connection. For those households purchasing drinking water from vendors, the mode of delivery is unique. Unlike in other parts of the city where vendors use pushcarts to deliver water, the vendors in Vingunguti use modified three-wheeled passenger vehicles. Where the backseat would normally be is instead a large water tank with an attached hose that customers use to fill their containers. Customers may wait along the side of the road for a vehicle to pass or come to the road when they hear the vehicle approaching. These vendors are not employed by DAWASCO but instead sell water sourced from a private business. The vendors state that the water comes from a DAWASCO connection, and households indicate they are drinking DAWASCO water when purchasing from these vehicles. There are potential concerns with the quality of this vended water. The hose is exposed to the air so could become contaminated from dust and perhaps even vehicle exhaust. The tank itself often has a dirty exterior, and it is impossible to know how often the inside is cleaned.

Fourth, related to the source of drinking water, the cost of this drinking water varies. For those household who purchase DAWASCO from an external source or mobile vendor, the prices are much higher in Vingunguti than Buguruni. In Buguruni those households who fetch this drinking water from neighbors with connections pay 100–150 Tanzanian Shillings for 20 liters. At this time of this research, 100 Tanzanian Shillings was equivalent to 4.5 US cents. In Vingunguti, vendors sell 20 liters of DAWASCO water for 500 Tanzanian Shillings. To put that price in context, households purchase 20 liters of well water for 50–100 Tanzanian Shillings. Although vendor delivery is more convenient for households, they also have limited choices for drinking water so are forced to pay this premium price, which potentially limits the amount purchased and consumed. This price is also standard across all of these vendors. Bayliss and Tukai (2011) suggest rather than using competition to lower prices, some vendors in Vingunguti practice collusion to keep prices uniform. Beyond price collusion, prices may be artificially inflated. Pushcart vendors generally sell DAWASCO water in other parts of Ilala District for 200–400 Tanzanian Shillings per 20 liter container, much lower than in Vingunguti. Smiley (2016) documents inflated pricing in the city at DAWASCO kiosks where posted signs give the price as 50 Tanzanian Shillings for a 20 liter container but water is actually sold for 100 Tanzanian Shillings.

Situating Buguruni and Vingunguti in the UPE literature

Buguruni and Vingunguti share many commonalities and were once considered to be one area but now have markedly different waterscapes owing to their heterogeneous delivery configurations. Although Buguruni has more household connections to the DAWASCO network, not all of its residents are connected. A large part of the surveyed households still obtain DAWASCO water informally through a neighbor. This water still comes from the formal network even though it is obtained informally. Of those using well water, a few have personal wells and one uses a government well, but the rest purchase it informally from neighbors. So even in an area where network expansions have occurred, heterogeneous provision still exists. An additional layer of diversity occurs when you look at Buguruni and Vingunguti side by side. When looking at piped connections, on the surface both areas have similar statistics yet one is linked to the modern infrastructural ideal while the other is connected to what Bousquet (2010) calls micro-networks.

These heterogeneous water delivery systems vary in their functionality. In Buguruni, those households that are not connected to the network are still fairly satisfied with their water access. Even if they have not personally gotten a household connection, their neighbors likely have. As one respondent noted, if one house is not selling water for some reason, there are now many other options. Others commented that they used to walk far distances to fetch water but now only walk to the next house. Respondents in Vingunguti described very different situations. Some noted that while there are sources close to their homes, they are not always available whether because they are broken or inoperable because of electricity shortages. Thus, Vingunguti residents were more likely to walk further to fetch water in those situations and they indicate less satisfaction with their water access.

This comparison of water provision and access makes two important connections to UPE. One attempts to fill a gap identified in the literature while the other extends ideas of the modern infrastructure ideal. First, this research seeks to address the critiques of UPE that it does not pay enough attention to scale and everyday spaces (Ahlers et al., 2014; Zug and Graefe, 2014). Following those concerns, Lawhon et al. (2014) point out that “UPE tends to overlook the situated understandings of the environment, knowledge and power that form the core of other political ecological understandings as well as recent work in Southern urbanism” (498). In telling the stories of six residents of Kampala’s Namuwongo neighborhood, McFarlane and Silver (2017) illustrate how people navigate everyday life in the margins and cope with poverty and call for more research on the dialectics of everyday urbanism. In particular, feminist political ecology (FPE) presents a framework that focuses more attention on inequalities within urban spaces, specifically on how gender operates at these different scales (Truelove, 2011). Although water differences in Buguruni and Vingunguti are not attributed to gender, the broader point is still applicable. It is important that UPE considers a wide range of inequalities and differences in water access at smaller intra-urban scales. “FPE approaches to urban water help to illuminate how and why social inequality continues to be tied to water even when water quantities and access points are improved” (Truelove, 2011: 150).

These case studies here show the complexity of water provision even in the face of improved network delivery. As Truelove (2019) notes in her study on Delhi, there is an important differentiation within informality that produces different outcomes. Thus, one benefit of this small-scale case study is the ability to highlight and describe the range of water supply that occurs within Dar es Salaam. The alternative supply systems in the city include micro-networks where households obtain piped connections to private wells and vended drinking water delivered by adapted three-wheeled vehicles. There is both inclusion and exclusion in this diversity: micro-networks help people get closer to their goal of piped network access but costly vended water prevents others from exclusively using their ideal source. This small-scale approach helps to illustrate that water supply is “neither the centralized piped-water network nor its absence, but the range of practices and technologies that unite people, nature and artefacts in a complex socio-ecological politics of water” (Furlong and Kooy, 2017: 888).

This research illustrates the importance of considering scale when examining urban water access. If Dar es Salaam’s informal areas were analyzed as a uniform space, it would overlook the spatial variations of networked service and informal delivery systems. In Vingunguti, in the absence of reliable and consistent DAWASCO service, informal providers continue to fill the water provision gap. Although nearly all households use wells as their primary source of water, only four households used a government well; the vast majority use private wells that are dug by individual households who then allow their neighbors to either fetch water or connect directly to the well via pipe. For drinking water, informal vendors provide the majority of water delivery via mobile trucks. Although this is DAWASCO water, it is resold by informal providers. The social networks that allow for this informal water provision exemplify Cleaver’s concept of institutional bricolage. As Wutich et al. (2018) find in their review, water sharing is more than a charitable act. It also illustrates complex social relations and unequal power relations that are shaped by broad political ecologies of water.

The role of informal suppliers in the city is not new. Today these informal suppliers continue to play an important role in water provision in Dar es Salaam, with estimates suggesting that they supply about half of the city’s population (Wamuchiru, 2017). One significant benefit of these suppliers is that they extend water access to those areas where the piped network providers have not yet reached or do not intend to reach (Kjellén and McGranahan, 2006). Yet they also charge prices up to 30 times the cost for DAWASCO water delivered directly to the home and even prices higher than in some developed world countries (Bayliss and Tukai, 2011). Thus, while the heterogeneity literature highlights the benefits of the diversity of delivery methods, the situation is complicated. On one hand, vendors and private individuals fill the gap left when DAWASCO service stopped. On the other hand, a lack of formal options—whether through networked pipes or government operated selling points—gives residents little actual water choice.

The differences in water access observed between Buguruni and Vingunguti illustrate both the dichotomies of formal/informal provision and of function/dysfunction. Many actors participate in water provision in Dar es Salaam, and the relationships between these actors vary substantially. As Mapunda et al. (2018) describe, informal water providers operate independently of formal providers since the Water Supply and Sanitation Act of 2009 does not recognize these small-scale providers. In their survey of peri-urban communities, the majority acknowledged the importance of informal providers but the majority of government officials interviewed saw very limited or no contribution from these providers in improving water access and did not see them as potential actors in the water sector. Likewise, these officials did not support strengthening the interaction between formal utility providers and these small-scale providers.

Bourque (2010) also describes uneven relationships between actors in Buguruni. While relationships between residents and water sellers were fairly cooperative, relationships between water sellers and both community organizations and local governments were harmed by a lack of transparency, legitimacy, and perceived corruption. Relationships between residents and DAWSCO were the worst because of poor service and lack of community outreach and involvement. The surveys discussed here raised similar mistrust between residents and DAWASCO including with meter readers inflating bills. One respondent in Buguruni produced an abnormally large bill of almost 4.5 million Tanzanian Shillings (nearly US $2000). Other respondents raised concerns about the lack of a contract stating a connection fee and with technicians requiring households to purchase additional materials on the spot.

Second, UPE critiques the idea that the modern infrastructure ideal and networked city is the development objective and instead argues for a more heterogeneous approach that evolves differently based on the place (Lawhon et al., 2018; Monstadt and Schramm, 2017). As Schramm (2016) notes, piped infrastructure extended to all is the ideal even though attempts to achieve it, especially in the colonial era, led to water shortages and inequalities. In valuing urban heterogeneity, UPE still sees the role that the modern network can play in water supply and acknowledges that for some urban residents, these connections are the ultimate goal. For those living at the margins in informal areas, connecting to a network and paying for water is a way to gain recognition, respect, benefits from the state, and citizenship and belonging (Anand, 2017; Ranganathan, 2014). It is important to not discount these desires.

In Vingunguti, similar desires for DAWASCO connections exist. Most households surveyed have little experience using it beyond drinking, but it is still their ideal water even if they currently use well water for most purposes. In fact, 81% of households indicated it would be their water of choice if they had no constraints. These households noted that DAWASCO tastes better than well water and that it is actually better for household purposes; for example when using it for laundry, they can use less, and even cheaper, soap and their clothes look cleaner and brighter. For those households that would still prefer well water over DAWASCO, most cited concerns with availability rather than displeasure with the actual water.

Although some households might choose to purchase this preferred DAWASCO water informally from neighbors because of reservations about cost, many indicated a desire to have their own direct connections. DAWASCO has told residents in the area that service will eventually be extended there. Some residents have even been approached—either directly by DAWASCO representatives or by their local government—to gauge their interest in receiving a household connection and most of them are indeed interested. One resident even stated that he did not know the benefits of DAWASCO water but he still signed up for a connection with his local government. Another stated that she will connect because “everyone wants DAWASCO” and indicated a benefit was more hours of availability than the well he currently uses. An additional benefit highlighted was the importance of clean and safe water, with one respondent stating that DAWASCO is a “must” in this regard. For those households who signed up for a connection, they had little information about timing or whether they had a binding commitment from DAWASCO for service, but they still want this connection. For them, Graham and Marvin’s modern infrastructure ideal is truly their ideal form of water delivery. Rugemalila and Gibbs (2015) argue that poor water access in Dar es Salaam is caused largely by governance failure which they use to include formal and informal institutions and practices. They highlight a lack of effective community participation in all stages of water projects, especially the planning. This failure is evident in Vingunguti as households are clearly not involved in any planning for future DAWASCO projects.

Conclusion

Improving access to water remains a major development challenge. Not only do Sub-Saharan African cities lag behind cities elsewhere in the world, access also varies within these cities; it is the poor and most marginalized populations who often live in informal areas that have the worst access. UPE traces these inequities to colonial policies that privileged the small European population at the expense of the large indigenous population. These inequities persist, with piped water networks serving only portions of cities while large parts of the city lack network access and instead rely on alternative systems. Critiques of the modern infrastructure ideal object to equating formal networks with development. In fact, some households may choose not to connect to these systems or to utilize both formal and informal supplies. Furlong and Kooy (2017) advocate for the worlding of water’s urbanization to consider this range of water access. Doing this does not make piped networks less important nor does it discount the desires of people to connect to them but instead “underscores the need to look beyond the network for a fuller understanding of water’s urbanization and for the development of meaningful policy responses in turn” (889). UPE advocates for this wider focus on water provision so rather than seeing a lack of network, we are instead engaging with the everyday life in developing world cities. Some have even called for new models that integrate alternative systems with formal networks (see for example Furlong, 2014; Jaglin, 2014).

Yet the case studies of Buguruni and Vingunguti presented here show that water inequities are not entirely due to the colonial history of segregation and unequal amenity provision. These two wards share many socio-economic and historical characteristics; they are both informal, were once considered one larger space, and are centrally located rather than part of the urban fringe. Despite these similarities, their water systems have developed along different trajectories. Buguruni receives more DAWASCO service, and notably this service has vastly improved over the past few years. Vingunguti’s residents primary use well water and must purchase DAWASCO drinking water from mobile vendors even though it once had more reliable networked water service. It has been promised DAWASCO service in the future, but it remains unclear when and if it might experience the same water improvements as Buguruni. It is these differences in places that make this comparison interesting and important. McFarlane (2010) advocates for indirect learning through comparisons of differences rather than focusing on similarities. Ranganathan and Balazs (2015) do this with a north–south comparison to use two unlikely cities to prompt reflection of new policies, theories, and models.

It is useful to return to Cleaver’s ideas of bricolage to understand the complexities of water provision in these two areas. It is the mixing of formal and informal along with modern and traditional that creates this heterogeneous supply. The result is a complex combination of a modern network that is desired by some and innovative and adaptive alternative systems. The gray zones that result help to illustrate why Graham and Marvin’s splintering urbanism theory does not fully explain the everyday realities of water access challenges. Dar es Salaam has been a fragmented city since its colonial origins, and these divisions will not be erased with the implementation of a modern infrastructure network nor will the city be a failure without this network. As Wamuchiru (2017) demonstrates in her study on the city’s Chamazi settlement, although networked water access was the initial end goal, residents instead made the conscious choice to be “beyond the network” rather than connect to the formal network.

This intra-urban comparison shows the complexity of water delivery services in Dar es Salaam. It also demonstrates the key role that alternative delivery systems can play in bringing water to underserved areas. It is important to remember, however, that while they offer potential benefits, these alternative systems are diverse. Some are innovative such as the well micro-networks in Vingunguti. On the other hand, the area’s mobile vendors may be exploiting a community with no other drinking water choices. The modern infrastructure ideal has been rightly criticized for seeing Southern cities as failed or undeveloped. Still these cities have multiple water delivery systems that are creative and truly benefit people. In this way, the networked delivery system is just one part of a larger heterogeneous waterscape.

This intra-urban comparison seeks to contribute to the call for UPE to pay greater attention to scale and everyday life. In doing so, it shows substantial water provision differences in Dar es Salaam’s informal areas. Only by understanding the inequities at these smaller scales is it possible to work toward greater justice. Ranganathan and Balazs (2015) argue for a greater connection between UPE and environmental justice literatures. Rather than view these as a binary, they suggest areas of cross-fertilization including in the issues of informality, the everyday state, and the power relations inherent in water access. This comparison also challenges the critiques of splintering urbanism by including the voices of residents who consider networked connections the ideal source of water. For them, the alternative delivery systems do serve them but do not provide the water they truly want. The heterogeneity seen in Dar es Salaam’s waterscape exists both between and within areas and demonstrates the possibilities of alternative suppliers but also the need to continue to improve the water network infrastructure.

Highlights

Dar es Salaam’s water is provided through the networked infrastructure and alternative informal systems.