Introduction

Low- and middle-income countries (LMICs) in Africa, Asia, and Latin America have experienced fast paced population growth and urbanization over the past decades (UN 2014, 2017). For example, the urban population on the African continent has increased from less than 15% in 1950 to over 40% in 2014 (UN 2014, 2017). While urbanization creates remarkable livelihood opportunities for urban dwellers in newly established industries, it comes often with a negative trade-off for the environment and public health (UN Habitat 2014).

Around these rapidly growing cities in LMICs, appropriate treatment facilities for the emerging micropollutants, which are released with industrial effluents, such as heavy metals, pharmaceuticals, or pesticides, are often missing (Nyenje et al. 2010). As a consequence, contaminated industrial wastewater is released into the environment without appropriate treatment (Nyenje et al. 2010). Several studies have shown that the heavy metals (e.g., arsenic (As), chromium (Cr), lead (Pb), and mercury (Hg)) are among the pollutants of high concern for the environment and public health (Wasi et al. 2013; Khan et al. 2015; WHO AFRO 2016). In urban areas, these contaminants can occur from natural sources but are also released through industrial processes, such as the production of batteries, paints, and plastics or urban mining (Calamari and Naeve 1994; Weiss et al. 2016; WHO AFRO 2016). Exposure to heavy metals can be multifaceted through air, food, water, or soil and can lead in severe cases to organ damage or distort neurodevelopment in fetuses and children (Järup 2003; Tchounwou et al. 2012; WHO AFRO 2016). Furthermore, As and Cd are classified carcinogenic to humans (WHO IARC 1993, 2006, 2012). Exposure to Hg and Pb compounds has toxic effects on fetal growth and neurodevelopment in children leading to reduced intelligence, attention deficit hyperactivity disorder (ADHD) and autism (Vrijheid et al. 2016). Moreover, some heavy metals like copper (Cu), Cr, or Hg can exhibit toxicity to the environment and accumulate in plants and animal species consumed by humans (Islam et al. 2007; Wasi et al. 2013; Baran and Antonkiewicz 2017).

In Uganda, the National Environment Management Authority (NEMA) sets the maximum permissible limits for heavy metals in industrial wastewater in the “Standards for Discharge of Effluent into Water or on Land” (NEMA 1999). In Kampala, the capital city of Uganda, the “Pollution Task Force” consisting of the NEMA and the most important public stakeholders and private industries was formed. Their aim is to reduce industrial pollution by enforcing the discharge regulations and to create public awareness for sustainable production (https://www.kcca.go.ug/pollution-control-task-force). For this endeavor to succeed, a monitoring scheme to track industrial activities and pollution levels in Kampala’s major wastewater management system is needed (Weiss et al. 2016). This requires a thorough understanding of the spatial and temporal variations of various chemical pollutants along the channel system.

However, there are currently no comprehensive details of industrial activities involving heavy metals and the contents of their effluents in Nakivubo sub-catchment (main wastewater drainage channel in Kampala). Furthermore, published data on heavy metal concentrations are inconsistent and only provide point estimates of the environmental contamination from industrial sources along the Nakivubo Channel (Muwanga and Barifaijo 2006; Mbabazi et al. 2010; Sekabira et al. 2010a, b; Fuhrimann et al. 2015). For example, a study conducted between 2006 and 2008 reported Pb levels of between 1.3 and 16.0 mg/L (Mbabazi et al. 2010). These concentrations were considerable different to another study 5 years later which measured Pb levels between 0.1 and 3.0 mg/L, even though the samples were taken at the same location and season (Fuhrimann et al. 2015). These differences may be explained by a number of factors, such as varying rainfall patterns and long-term changes in industrial activities, but could also be due to short-term peaks introduced by daily and diel fluctuations of industrial wastewater generation or discharge (Rule et al. 2006; Nimick et al. 2011; Meghea et al. 2012; Camacho-Muñoz et al. 2014; Kao et al. 2017; Swain and Sahoo 2017).

Several studies have studied and described the long-term variations of industrial heavy metals in river water systems (Mbabazi et al. 2010; Zhen et al. 2016; Batbayar et al. 2017; Eliku and Leta 2018). While these studies point at considerable seasonal fluctuations of these pollutants, little is known about their daily variation. The few studies assessing daily variations of heavy metals in urban wastewater have a strong focus on cities in high- and middle-income countries in Europe (Rule et al. 2006; Meghea et al. 2012; Deycard et al. 2014). Hence, there is a lack of published literature on the daily variation of heavy metals in urban wastewater streams in low-income settings, such as Kampala, Uganda.

The objective of this paper is to guide future monitoring efforts, to set reasonable contextual pollution thresholds, and to contribute to the limited literature on the short-term variation of industrial pollution in urban wastewater in low-income countries. Specifically, the study aimed at assessing the daily variation of heavy metal contamination of the Nakivubo Channel over the course of 1 week, to compare heavy metal pollution levels to national standards and to map and rate industries discharging heavy metals along the Nakivubo Channel system.

Materials and methods

Study area

Kampala, the capital city of Uganda (geographical coordinates 0° 18′ 49.18″ N latitude and 32° 36′ 43.86″ E longitude), has an estimated population of 1.5 million people according to the 2014 census (UBOS 2016). The city is built on 21 hills at approximately 1140 m above mean sea level and is bordered by the Lake Victoria in the south. Kampala has tropical climate and rainfall is expected throughout the year. Rainfall peaks between March and May (short rainy season) and September and November (long rainy season) (UBOS 2017). The driest months are January and July, with 51 and 50 mm precipitation, respectively (UBOS 2017). The rainfall data for the sampling period measured at the Bugolobi Sewage Treatment Works (BSTW) near the city center is given in Fig. S1 in the Electronic Supplementary Material.

This study focuses on the catchment area of the Nakivubo Channel. The channel stretches over 12.3 km through the whole city transporting industrial and domestic wastewater including the effluents of the BSTW. The Inner Murchison Bay (IMB) is important for local fisheries and the source of Kampala’s drinking water is located about 4 km from the mouth of the Nakivubo Channel (Fuhrimann et al. 2015).

Industries active in the Nakivubo Channel catchment area include manufacturers of paints, plastics, metal products, textiles, as well as warehouses and garages. A more comprehensive overview of the industrial activities is shown in Table S1 in the Electronic Supplementary Material. After passing the Nakivubo wetland, the channel discharges into the IMB of Lake Victoria.

Sampling strategy

To understand industrial water pollution within the catchment of the Nakivubo Channel, 16 sampling points were strategically selected to capture the variation along the channel as well as the contribution of waters feeding the main channel (Fig. 1). The points were grouped into three categories: six points along the Nakivubo Channel (C1–C6), four tributaries (T1–T4), and six discharge points from industries (D1–D6). More details on the sampling points (GPS coordinates and industrial activities in the proximity) are given in Table S1 in the Electronic Supplementary Material.

Fig. 1
figure 1

Sixteen selected sampling points along the Nakivubo Channel, Kampala, Uganda. C1–C6 = six points in the main Nakivubo Channel; T1–T4 = four tributaries; D1–D6 = six discharge points of industries

Full size image

At each sampling point, three water samples (0.5 L each) were collected in the mornings between Thursday, December 10 and Wednesday, December 16, 2015. All samples were taken in clean plastic bottles and stored in an ice chest. Within 3 h, the samples were transferred to the storage facility and put in a refrigerator at 4 °C. The samples that were used for heavy metal analysis were acidified with 0.5 mL conc. nitric acid (15.6 M) to a pH of < 2. This is to maintain the heavy metals in solution until laboratory analysis were conducted.

Sample analysis

To identify the relevant heavy metals, data from previous studies in the area (Mbabazi et al. 2010; Fuhrimann et al. 2015) were considered and discussed with local experts on industrial pollution. Furthermore, in a pilot study, a range of different heavy meals were analyzed in samples from four locations in the Nakivubo channel. Based on these results, Cr, Pb, Hg, and Cu were selected for this study. The analysis was done by a local laboratory (Chemtest (U) Ltd.) following the American Public Health Association (APHA) standard methods (APHA, AWWA, WEF 1995); the same procedure which was previously used to measure heavy metals in the area (Fuhrimann et al. 2015). Electrical conductivity (EC), pH, and total dissolved solids (TDS) were directly measured in situ using the CyberScan PC 300 multiparameter water analyzer (Eutech Instruments, Landsmeer, The Netherlands). Biochemical oxygen demand (BOD5) and chemical oxygen demand (COD) analysis was done at the National Water and Sewerage Corporation (NWSC) central laboratory at BSTW on the first and the last sampling day. BOD5 was analyzed with the azide modification of the Winkler method (APHA, AWWA, WEF 1995). For COD, the closed reflux colorimetric method was chosen (APHA, AWWA, WEF 1995).

Mapping and rating of industries

To map potentially polluting industries, 36 industries located in the industrial area in central Kampala were selected based on their potential to discharge heavy metals into the environment and georeferenced. Among those, a subset of 25 industries were selected for key informant interviews based on the type of products produced and relevance for heavy metal pollution. During the interviews information on production, raw materials, effluent management, and monitoring were obtained. The companies were then grouped into three groups:

  • “Class 1” industries using raw materials that are known to contain heavy metals: This class comprises manufacturers of paints, textiles, batteries, or corrugated iron sheets

  • “Class 2” industries that are suspected to release heavy metals but information on the chemical characteristics of their effluent is not available such as vehicle servicing firms, electric poles processing firms, soap, and plastic manufacturers and other metal processing industries (pipes, galvanized products)

  • “Class 3” are industries that produce goods which are not directly associated with heavy metals but contribute to organic and nutrient loading (high BOD5, COD) of the Nakivubo Channel (e.g., production of food and beverages)

Data analysis

All measured water quality parameters were visualized using box plots or scatter plots grouped by their source or sampling day and compared to Ugandan and other international standards (see Table 1). For analyzing the overall spatial variation (i.e., differences between sampling points in the main channel, tributaries, and discharge points), the Kruskal-Wallis equality-of-populations rank test was used. To further explore which individual type of source differed, Wilcoxon rank-sum tests were performed. Temporal variations in heavy metal levels and physicochemical parameters were analyzed using the Skillings-Mack test. A p value below 0.05 was considered as statistically significant and p values below 0.001 as highly significant. All statistical analyses were done using Stata 13.1 (Stata Corp, College Station, TX, USA). Content analysis was used to extract qualitative and semiquantitative information from the key informant interviews.

Table 1 Limits of detection and Ugandan guideline values for measured water quality parameters set by the Ugandan National Environment Management Agency (NEMA) (NEMA 1999)
Full size table

Results and discussion

One hundred seven water samples were taken at 16 locations over the course of 7 days. Heavy metal analysis was done for 78 samples from 11 sites as some sampling points were in close proximity and thus little variation was expected. At the discharge point at the 6th Street Bridge (D4), samples could only be taken on two occasions (December 10 and 13, 2015) due to the low water flow. Of these, one was used for BOD5/COD (December 10, 2015) and the other for heavy metal (December 13, 2015) analysis. The samples taken from the main channel on December 14, 2015 had to be excluded from the heavy metal analysis due to a calibration error.

Levels of chemical pollution

Figure 2 shows the levels of physicochemical parameters along the Nakivubo Channel. Measurements of pH and TDS are all within the NEMA standards. On December 12 and 13, 2015, the EC of the effluent from BSTW (D3) exceeded the NEMA standard (> 1500 μS/cm). BOD5 and COD concentrations exceeding national thresholds (above 50 and 100 mg/L, respectively) were measured at seven and eight sampling points, respectively (Fig. 3). Extremely high COD values were found in the effluent of the local slaughterhouse (1111 mg/L) and at the discharge point by the 6th Street Bridge (891 mg/L).

Fig. 2
figure 2

Box plot showing spatial and temporal variation of the pH, electrical conductivity (EC), and total dissolved solids (TDS) measured at 16 sites along the Nakivubo Channel, Kampala, Uganda between December 10 and 16, 2015. Left: across sampling points (N = 7); right: across sampling days (N = 17). The p values indicate differences between categories of sampling points derived from Wilcoxon rank-sum tests. Red bars indicate National Environment Management Agency (NEMA) thresholds

Full size image
Fig. 3
figure 3

Box plot showing BOD5 and COD concentrations stratified by the three categories of sampling points along the wastewater system in Kampala, Uganda. Red-dashed line indicates National Standards for BOD5 (50 mg/L) and black solid line COD (100 mg/L). The p values indicate statistical differences between sampling categories derived from Wilcoxon rank-sum tests

Full size image

The Cr, Hg, Pb, and Cu concentrations at 11 sampling points are illustrated in Fig. 4. The concentration for Cr ranged from 0.0018 to 0.2090 mg/L (median, 0.101 mg/L), for Hg from 0.0069 to 0.0103 mg/L (median, 0.080 mg/L), for Pb < 0.0045 (below LOD) to 0.6919 mg/L (median, 0.0569 mg/L), and for Cu < 0.0001 (below LOD) to 0.0627 mg/L (median, < 0.0001 mg/L). For Cr, no sample exceeded the National Standards for discharge of effluent into water or on land set by NEMA (1.0 mg/L). The NEMA standard for Hg (0.01 mg/L) was exceeded once at the discharge point near Kiboko Enterprises (D6). In total, the NEMA standard for Pb (0.1 mg/L) was exceeded in 29 samples taken from all sampling points except for the Fire Brigade Bridge (C1), the Access Road Bridge (C2), and the Namuwongo Bridge (C4). By far, the highest concentration of Pb was measured in the effluent of BSTW (D3; 0.6919 mg/L) on December 11, 2015. Cu concentrations were all well below the NEMA threshold (1.0 mg/L).

Fig. 4
figure 4

Spatial and temporal variation of chromium (Cr), mercury (Hg), lead (Pb), and copper (Cu) concentrations measured at 11 sites along the Nakivubo Channel system, Kampala, Uganda between December 10 and 16, 2015. Left: across sampling points; right: across sampling days. The p values indicate differences between categories of sampling points derived from Wilcoxon rank-sum tests. Red solid lines indicate NEMA thresholds (standards for Cu and Cr (1.0 mg/L) beyond scale)

Full size image

The elevated Pb concentrations measured along the Nakivubo Channel were in line with those reported by Muwanga et al. (in December 2003–February 2004), Sekabira et al., and the Ugandan Directorate of Water Resource Management (DWRM, unpublished data collected in July 2013) (Muwanga and Barifaijo 2006; Sekabira et al. 2010b). On the other hand, considerably higher Pb concentrations were measured in samples taken in Decembers in 2006 to 2008 (> 16.0 mg/L) (Mbabazi et al. 2010) and in November 2013 (3.0 mg/L) (Fuhrimann et al. 2015).

Reasons for the deviations of heavy metal concentrations from some studies can be attributed to dilution with rainfall and run-off during the wet season and concentration during the dry season, varying patterns of effluent discharge by industries, remobilization of heavy metals from sediments, or even periodic removal of sediments during dredging (Kayima et al. 2008). There were indeed heavy rains in the week prior to sampling (i.e., 234 mm at Bugolobi weather station, unpublished data). Industries holding back their wastewater in retention tanks may have used this opportunity to discharge contaminated effluents. In addition, sediments potentially containing heavy metals had been removed from the channel when parts of the channel bed were desilted by Kampala Capital City Authority (KCCA) in the week prior to sampling of this study (KCCA, unpublished data) potentially leading to reduced resolubilization.

Spatial variation of chemical pollution

Statistical analysis of the spatial variation revealed with high significance that the pH was lower in the tributaries, and that TDS and EC were higher in the samples from the discharge points (Fig. 2). Figure 3 shows that tributaries showed significantly less organic pollution than the main channel. Among the measured heavy metals, particularly Pb concentrations varied spatially when comparing all samples taken on the same day (Fig. 4). For example, on December 16, 2015, Pb concentrations in the samples from the main channel (C) ranged between 0.0059 and 0.1971 mg/L. Statistical analysis revealed that for all measured heavy metals, there were significant differences between the types of sources. More specifically, concentrations in the main channel were lower than in the tributaries for Pb and Cr but higher for Hg and Cu.

In contrast, no reduction in heavy metal concentrations through the Nakivubo wetland was observed. The wetland is located in the downstream part of the Nakivubo channel and acts as a flood plain and a buffer to pollution of Lake Victoria. It is anticipated that the wetland ecosystem would provide tertiary treatment for the wastewater transported by the channel. However, also Mbabazi et al. found a reduction in the capacity of the Nakivubo wetland to filter heavy metals from the urban wastewater channel (Mbabazi et al. 2010). This may reflect the altered hydrology of the wetland system through channelization and continued degradation resulting in 62% vegetation loss between 2002 and 2014 (Isunju and Kemp 2016).

The lack of tertiary treatment causes flushing of pollutants directly into the lake while threatening the local fish production and recreational activities in the IMB (Mbabazi et al. 2010). Indeed, traces of Pb exceeding permissible limits in vegetables and fish from the downstream area were found (Nabulo et al. 2012; Fuhrimann et al. 2015). Similarly, in a study assessing public health hazards from heavy metals in Kampala’s drinking water conducted in September to November 2015, Pb was identified as the most significant pollutant (Bamuwamye et al. 2017). Strikingly, all drinking water samples taken at public water taps and groundwater-fed protected springs were above national and international standards for Pb (Bamuwamye et al. 2017).

Temporal variation of chemical pollution

High temporal variation and spikes in chemical pollution were detected. For example, the differences for pH, TDS, and EC between the sampling days were statistically significant (pH: p = 0.002; TDS: p < 0.001; EC: p < 0.001). As shown in Fig. 2, the medians for those indicators were lowest on Friday, December 11, 2015. Similarly to the spatial variation, Pb showed also the highest temporal variation which was larger at the discharge points (D) compared to the main channel and tributaries (C and T) (see Fig. 4). As an example, the median in Pb levels on Friday was almost seven times higher than on Monday (0.1008 vs. 0.0146 mg/L). Despite these variations, there were no statistically significant differences between the weekdays indicating an absence of a clear temporal pattern. Additionally, some peak values were found when samples exceeded the weekly median of the respective sampling point by a multiple. For example, the Cu concentration measured on Tuesday in the effluents of the slaughterhouse was more than 11 times higher than the median across the whole week. Also, while during four of the days no Pb (i.e., < 0.0045 mg/L) was detected in the effluent of BSTW (D3), the concentration measured on December 11, 2015 exceeded the NEMA threshold almost seven times (0.6919 mg/L).

There are only few studies looking at the temporal variation of heavy metal in urban wastewater at such high resolution. Rule et al. assessed daily heavy metal concentrations in wastewater in a city of the UK and found only little variations for Hg, Pb, and Cr (Rule et al. 2006). Similarly to this study, they also found peak values that spiked from 1 day to the other, mostly on Mondays (Rule et al. 2006). In another study of heavy metal pollution of urban rivers in Romania, a 7-day periodicity with lower Hg levels on the weekend has been found (Meghea et al. 2012).

The presence of high variations and various peak concentrations but a lack of a clear pattern over the week as found in the present study point at several issues. Firstly, factors that would affect the whole channel system, such as dilution from the little rainfall during the sampling week or the varying contribution of heavy metals released from natural sources, were marginal. Secondly, industries may hold back their effluents in retention tanks and release them in bulks which are only detected locally. Indeed, some industries indicated in the key informant interviews to discharge their effluents irregularly, ranging from daily to once a month (see following section).

Industrial manufacturers in the Nakivubo catchment area

In total, 36 industries were georeferenced along the eight streets of the industrial area in proximity to the Nakivubo Channel (see Fig. 5). Eight industries were classified into Class 1 among which three produce paints, three textiles, one car batteries, and one iron sheets. Further, 16 industries were grouped into Class 2 and seven into Class 3. Five industries were not classified as they did not belong to any of the pollution classes.

Fig. 5
figure 5

Classification of the 31 industries in the Kampala industrial area classified according to raw materials and manufacturing processes

Full size image

Of the 25 industries included for key informant interviews, seven (28%) have an effluent treatment plant in place. Eight (32%) industries discharge continuously during operations while others collect effluents and discharge them either daily, one to two times per week, or even only once a month. At ten industries (40%), effluents are monitored by an external agency while two (8%) rely on internal controls.

The key informant interviews further revealed that there is currently no extensive understanding of industrial activities in the Nakivubo sub-catchment. Furthermore, there is a lack of monitoring and management systems for industrial wastewaters. These factors taken together may contribute to the inconsistent but continuously high heavy metal pollution of the Nakivubo channel system (Mbabazi et al. 2010; Sekabira et al. 2010b; Fuhrimann et al. 2015).

Strengths and limitations

This study is the first to investigate the daily variation of heavy metals in an urban wastewater stream in Africa and adds to the limited global literature on that topic. The high temporal resolution allowed detection of peak levels that would otherwise remain undetected by other monitoring efforts. Furthermore, as many samples were taken on the same day, statistical comparison for both spatial and temporal variations of the heavy metal contamination in the whole channel system was feasible. Another strength of the study was the consistent schedule according to which samples were collected daily which reduced the risk of confounding from different flow intensities resulting from human and industrial activities.

A limitation of this study is the lack of flow measurements. Urban wastewater channels, such as the Nakivubo Channel, are mainly fed by rainwater runoff and wastewater generated by domestic and industrial activities (Barrenha et al. 2018). Hence, the quantity of water in the channel and thus the dilution of heavy metals may vary over the course of 1 week (Koklu et al. 2010; Barrenha et al. 2018). However, rainfall in the week prior and during the sampling period was little and limited to the first two sampling days. Finally, the inventory of industries focused on large-scale industries. Small-scale industries, like car servicing garages and recyclers of scrap metal, may combined contribute considerably to the heavy metal load of the Nakivubo Channel.

Recommendations

The high variation of heavy metal concentrations measured in this study in combination with evidence from the literature on Kampala’s wastewater management system leads to the following recommendations:

  1. 1.

    Longitudinal measurements over a short time interval are necessary to monitor industrial activities and their environmental pollution (Elzwayie et al. 2017). Ideally, monitoring schemes should be designed to capture the peak values and short-term variations of heavy metal levels but should also consider the financial and technical constraints of the local authorities (Weiss et al. 2016).

  2. 2.

    As an alternative or complementary to water samples, river sediments could be analyzed for assessing long-term variation of heavy metal pollution (Dafforn et al. 2012). Furthermore, biological samples, such as algae, crops, or fish, could be analyzed to not only reflect environmental pollution but also indicate potential human exposure to heavy metals (Stankovic et al. 2014).

  3. 3.

    New innovative methods for heavy metal monitoring, such as passive sampling or remote sensing, could be explored to not only reach a high temporal resolution but also to measure the pollution load over time (Mikkelsen et al. 2005; Roig et al. 2011; Stewart et al. 2016; Edenborn et al. 2017; Swain and Sahoo 2017).

  4. 4.

    The current NEMA standards do not consider the different pollution management technologies of industries. Thus, it could be advisable to set industry-specific standards as done in other countries (United States Environmental Protection Agency 2002).

  5. 5.

    For a sustainable reduction of industrial pollution, industries need to be supported in managing their wastewater. As a basis, industrial activities including potential pollutants should be registered by local authorities. Focus of support should be on the identified primary polluting industries but the role of small-scale industries should also be considered. In Kampala, the “public-private wastewater dialogue,” bringing together KCCA, NEMA, NWSC, and major industries, aiming at reducing industrial pollution in the Nakivubo sub-catchment, should be reinforced in facilitating capacity building and knowledge transfer.

  6. 6.

    Considering the threat of the high measured Pb and Hg concentrations to environment and public health via bioaccumulation in fish or drinking water pumped out of the Murchison Bay, natural ecosystem services from wetlands should be preserved through extension and enforcement of conservation areas (Tilley et al. 2014; Isunju and Kemp 2016; Bamuwamye et al. 2017).

  7. 7.

    Finally, adequate wastewater treatment systems need to put in place both centrally and at the source that are capable of reducing the Pb load of the Nakivubo Channel system (Barakat 2011; Tilley et al. 2014). For example, this could include technical measures such as organic polymer based coagulation/flocculation and hydroxyl precipitation basins or natural measures, such as attached growth stabilization ponds and constructed wetlands (Fu and Wang 2011; Dhokpande and Kaware 2013).

Conclusions

High temporal and spatial variations in heavy metal levels were detected in water samples taken in the Nakivubo Channel area in Kampala, Uganda. Together with reports from key informant interviews, this indicates that industries might discharge contaminated wastewater in bulks without appropriate treatment. Pb and Hg concentrations peaked at various occasions and their measured concentrations above the National Standards are a concern for environmental and public health.

To efficiently enforce environmental standards, isolated sampling campaigns are not sufficient to get a full picture of heavy metal pollution of wastewater systems by industries. Along the line of the provided recommendations, we note that a registry of industrial activities is needed to guide capacity building activities. Outlined technical and natural treatment measures should be implemented at the main polluting industries to treat effluents at source and in the Nakivubo wetland to avoid downstream bioaccumulation in fish and the pollution of drinking water pumped out the IMB in Lake Victoria.