Physiochemical and bacteriological quality of tanker vended water supply chain in Lodwar Town, Kenya Open Access

Water access and quality challenges are the most common limiting factors which have impaired safe water supply efforts globally, especially in sub-Saharan African countries like Kenya according to a UNICEF and WHO (2023). By 2022, 2.2 billion people were reported to still lack safely managed drinking water according to a Joint Monitoring Program (JMP) report by UNICEF and WHO (2023). Salehi (2022) in a global review observed that factors such as climate change, population growth, high water demand, prolonged drought and poor management contribute to water shortages in most parts of the world. These factors in addition to water quality and inadequate infrastructure have impacted urban water supply resulting in water utilities being unable to meet the demand for water services in informal or peri-urban settlements. Water vendors have emerged as an alternative water service provider. This is due to their flexible nature and general aim in making profits. The flexibility of vendors means the ability to use various water sources, the use of tanker tractors for alternative businesses such as the delivery of water for construction purposes, delivery of hardcore stones, ballast or sand, and moving to other towns when the water supply demand in Lodwar town is low. According to the Water Services Regulatory Board (WASREB) report (WASREB 2020), the underlying concerns are the quality of the water supplied and prices attributed to human activities and climatic variations. Without monitoring, water safety cannot be guaranteed to the consumers because it is difficult to tell what went wrong, where, and when along the supply chain. Water safety monitoring in Kenya involves several agencies such as WASREB, water utility (Turkana Urban Water and Sewerage Company (TUWASCO)), National Environment Management Authority (NEMA), Water Resources Authority (WRA), and public health department.

Water vending is prohibited in many countries such as Ghana, Senegal, and Mali because of the public health risks associated with it (Zuin et al. 2011). In sub-Saharan Africa, access to safe water is at 57% according to other studies (Asefa et al. 2023; Baddianaah et al. 2024). Furthermore, 35% of the urban population (308 million) had access to household piped water connection, while 49% relied on shared water (standpipes and kiosks) as reported by Zuin et al. (2011). Mobile vendors (pushcarts and tankers) obtain water from sources not approved by the water utility, especially in times of water shortage, thus compromising the quality of water they sell (UNDP 2011). Kjellen & McGranahan (2006) noted that such businesses result from utilities' limitations and failure to reach all customers. Water storage facilities and water vendors' handling practices provide a conducive environment for microbial and nutrient contamination due to poor hygiene and sanitation, improper cleaning of transporting and storage containers, providing an environment for diseases and illness-causing enteric pathogens to thrive (Okoko & Idise 2014; Kwami & Sawyerr 2018). Mobile water vendors' supplies are more expensive than fixed point water suppliers (public taps and kiosks), identified to be safer and more affordable in Kenya (UNDP 2011). For example, in Lodwar town, we noted during fieldwork that vended water goes for 1 Kenya Shilling per litre compared to piped water by the water utility which goes for an average of 5 Kenya Shillings per 20-L container.

Once a water source has been constructed or improved/rehabilitated, water quality in most cases is not given much attention or monitored due to financial and logistical constraints (Rossiter et al. 2010). In the cases where water quality is taken into consideration, studies have suggested that the focus is mainly on the degree of bacterial contamination and the health of consumers leaving out physiochemical properties (WHO 2003; Silva et al. 2008). The microbiological and chemical characteristics of vended water remained uncharted in most countries (Ayalew et al. 2014). The analysis of thermotolerant coliforms across supply chains of vended water in Kisumu (Kenya) and Addis Ababa (Ethiopia) showed deterioration of water quality in consumers' storage facilities compared to taps, standpipes, and borehole water (Ayalew et al. 2014). Additionally, Okoko & Idise (2014) affirmed that parameters of water quality deteriorated with distance from the water source. To maintain water quality during transportation means treating water at the source or regular cleaning and addition of water disinfectants at the tankers to avert water-related diseases.

Lodwar town was the best candidate for this study due to inadequate water access, water scarcity, and stress associated with arid and semi-arid land (ASAL) climatic conditions within the Sahel region. Coupled with the growth of commercial activities and population, these conditions allowed for alternative water supply systems, including water-vending activities which are of interest in this study. Lodwar town, the headquarters of Turkana County, is rapidly growing with a total population of 82,970 (KPHC 2019). Water coverage in Lodwar town is at 59% with water quality rating/approval at 66% (WASREB 2020). The existing water supply gap in the town may expose its populations to waterborne diseases (cholera, typhoid, and dysentery) reported in the area (WHO 2008; IFRC 2020). Moreover, Ayalew et al. (2014) asserted that the water quality of vendors has not been studied to determine the safety of the water supplied. This is a gap this study sought to address by analysing the physical, chemical, and bacteriological quality of water tankers in Lodwar town across the supply chain, from the source to the consumer point of use. This study aligns with one of the pertinent questions posed by Garrick et al. (2019); ‘do informal water markets deliver dirty water?’. Interestingly, Ayalew et al. (2014) reported that the perceived good quality of vended water was one of the key drivers of its demand by the consumers. This is because the consumers base their perceptions on organoleptic properties such as colour, odour, and taste together with vendor's reputation even though these organoleptic properties do not clearly indicate the accurate quality or safety of the water according to Garrick et al. (2019).

A total of 35 samples for physiochemical and bacteriological characteristics analyses were collected across the tankers’ supply chain with two replicates each totalling 70 samples. All the samples were analysed in triplicates at each sampling point with the third sample being a composite of the two replicate samples collected. Furthermore, 20 representative composite samples were selected for fluoride, chloride, lead, and iron analysis across the supply chain, as shown in Table 1.

All glassware and sampling bottles used were thoroughly pre-washed with acid water and then rinsed with distilled water before use for the collection of water samples. In addition, for microbiological analyses, the outlet taps for boreholes, tankers, and kiosks were wiped with 70% alcohol-soaked tissue paper during sampling. This was followed by a significant amount of water flushed out before sample collection to avoid contamination (Awere & Anornu 2016). The water samples were collected using 1,000-mL plastic bottles for chemical analysis, 500 mL for microbial analysis, and a further 500 mL for fluoride, chloride, lead, and iron analysis. Samples for in situ measurements (temperature, pH, EC, DO, TDS, turbidity, and salinity) were collected in a bucket, thoroughly pre-washed, and rinsed with the sample to avoid water recontamination. The readings were taken in triplicates from the field using respective multimeters calibrated prior, and electrodes were cleaned with distilled water to avoid errors as recommended in the manufacturer's instructions. They include the following: (i) HACH HQ 40d (EC, pH, temperature, salinity, and TDS); (ii) HACH HQ 30d (DO), and (iii) HACH HQ 11d (turbidity).

The samples were preserved in cool boxes and transported to the Turkana University laboratory for bacteriological tests within 24 h. Samples for chemical analysis (nitrate, nitrite, ammonium, sulphate, total alkalinity, chloride, fluoride, iron, and lead) were kept in the refrigerator at 4 °C until analyses at the Nakuru Water and Sanitation Services Company (NAWASSCO) laboratory in Nakuru City and Egerton University water quality laboratory. Free Residual Chlorine (FRC) was analyzed in situ using a pool tester in the DelAqua water testing kit. This was done by dropping one DPD (N, N-diethyl-p-phenylenediamine) No. 1 tablet at the right-hand cell (C12) of the pool tester. The sample was shaken until the tablet dissolved and the results were then compared by matching the colour with the pool tester standard colours representing different residual chlorine concentrations. All these parameters were analysed, as summarised in Table 2.

Table 2

Standard methods used for the determination of chemical characteristics of water quality

Parameter .	Method (APHA 2005) .
Ammonium-nitrogen (NH4-N)	Colorimetric method using reaction between sodium salicylate solution and hypochlorite solution
Nitrate-nitrogen (NO3-N)	Sodium salicylate method
Nitrite-nitrogen (NO2-N)	Colorimetric method using the reaction between sulphanilamide and N-Naphthyl-(1)-ethylendiamine-dihydrochloride
Fluoride (F−) ions	SPADNS (Sulphanilic acid-azochromotrop. 1, 8-Dihydroxy-2-(4-sulphophenylazo) naphthalene-3, 6-disulphonic acid trisodium salt) colorimetric method
Chloride (Cl−) ions	The Argentometric method (also known as the Mohr method) using silver nitrate as the titrant
Total alkalinity	Acid titration methods
Sulphates (⁠⁠) ions	The Turbidimetric method
Iron (Fe) and Lead (Pb) ions	The Atomic Absorption Spectrophotometry (AAS) direct air–acetylene method

Parameter .	Method (APHA 2005) .
Ammonium-nitrogen (NH4-N)	Colorimetric method using reaction between sodium salicylate solution and hypochlorite solution
Nitrate-nitrogen (NO3-N)	Sodium salicylate method
Nitrite-nitrogen (NO2-N)	Colorimetric method using the reaction between sulphanilamide and N-Naphthyl-(1)-ethylendiamine-dihydrochloride
Fluoride (F−) ions	SPADNS (Sulphanilic acid-azochromotrop. 1, 8-Dihydroxy-2-(4-sulphophenylazo) naphthalene-3, 6-disulphonic acid trisodium salt) colorimetric method
Chloride (Cl−) ions	The Argentometric method (also known as the Mohr method) using silver nitrate as the titrant
Total alkalinity	Acid titration methods
Sulphates (⁠⁠) ions	The Turbidimetric method
Iron (Fe) and Lead (Pb) ions	The Atomic Absorption Spectrophotometry (AAS) direct air–acetylene method

View Large

The faecal coliform densities were determined through the Membrane Filtration Technique (MFT) using the Oxfam – DelAqua water testing kit (Oxfam-Delagua 2012). This procedure is differential and selective for Escherichia coli and doesn't require further identification.

The data were processed using Statistical Packages for Social Sciences (SPSS) for analysis at a significance level of 0.05. A one-way ANOVA test was used to test for the relationships in the water quality parameters across the supply chain. FRC across the supply chain was analysed using the Independent-Samples Kruskal–Wallis test at a 0.05 significance level. The Pearson Correlation analysis test measured the correlation between assessed parameters across the supply chain.

The trends for physical and chemical characteristics of vended water values (mean scores ± SD) are represented in Table 3.

The temperature increased significantly (p < 0.05) across the tankers’ supply chain from the borehole (28.37 ± 0.32 °C) being the lowest to the tankers (30.26 ± 0.35 °C) being the highest with a slight decrease at the consumers (Table 3).

Turbidity across water tankers’ supply chain increased from the lowest at the standpipe (2.04 ± 0.39 NTU) to the highest at the consumer (5.90 ± 3.64 NTU) (Table 3). The value recorded at the consumers' point of use exceeded WHO-recommended guidelines of 5 NTU for drinking water quality (WHO 2017). Consequently, turbidity across the supply chain was significantly different (p < 0.05) (Table 3).

The concentration of EC across the tankers’ supply chain showed a slight increase from the borehole (271.33 ± 3.27 μS/cm) with a peak at the tankers (325.13 ± 36.18 μS/cm) (Table 3). The consumers' storage facilities recorded a slight decrease to 279.56 ± 7.44. However, EC across the supply chain was not significantly different (p > 0.05) (Table 3).

Salinity across the tankers’ supply chain recorded constant values at the borehole, standpipe, and consumer endpoints (i.e. 0.05‰ [parts per thousand]), as shown in Table 3. However, there was an insignificant slight increase at the tankers (0.06 ± 0.01‰).

Dissolved oxygen (DO) values increased across the tankers’ water supply chain ranging from 7.27 ± 0.19 mg/L being highest at the consumers and lowest (3.66 ± 0.49 mg/L) at the borehole, as shown in Table 3. The DO levels across the supply chain were significantly different (p < 0.05) (Table 3).

Total dissolved solids (TDS) increased across the supply chain ranging between 52.18 ± 0.47 mg/L at the borehole (lowest) and 62.60 ± 7.18 mg/L at the tankers (highest) with a slight decrease at the consumers (Table 3). Nevertheless, the TDS across the tankers’ supply chain was not significantly different (p > 0.05).

FRC was only detected at the consumers (0.19 ± 0.40 mg/L) and was significantly different (p < 0.05) using independent-sample Kruskal–Wallis tests across the supply chain (Table 3).

The pH ranged between 7.63 at the borehole (lowest) and 8.65 at the consumers’ storage tanks (highest) across the supply chain (Table 3). The pH values recorded at all sampling sites were above 7, implying the water was alkaline. There was a significant difference (p < 0.05) in the pH values across the supply chain (Table 3).

The concentration of total alkalinity across the supply chain of water tankers increased from the borehole (3.80 mEq/L) to the tankers (4.92 mEq/L) with a slight decrease at the consumers' storage tanks (Table 4). However, total alkalinity across the supply chain was not significantly different (p > 0.05).

Across the supply chain, it was observed that most parameters increased at the tankers. This was a result of corrosion of tanks, sediments, formation of biofilms in the tanks, and poor hygienity of tankers, as observed during fieldwork. Dissolved ions with an increase in EC, TDS, and temperature were observed at the tankers’ level. The high standard deviation at the tankers was an indication of high variance in the range of values reported in individual tankers' water which was away from the mean. Hence, this explains the difference in how vendors handled water across the supply chain. It points to other factors such as cleaning of tankers, presence of covers, distance travelled, and lidding among others.

The concentrations of ammonium, nitrite, and nitrate were relatively constant across the supply chain. The concentration of nitrate-nitrogen (NO₃-N) across the tankers' supply chain decreased from the borehole and standpipe both recording 0.058 mg/L towards the consumers' storage tanks (0.055 mg/L), as shown in Table 4. On the other hand, ammonium–nitrogen (NH₄-N) and nitrite-nitrogen (NO₂-N) recorded 0.043 and 0.004 mg/L, respectively, across all sampling points (Table 4). However, ammonium and nitrite values recorded across the supply chain were significantly different (p < 0.05) using One-Way ANOVA and Independent-Sample Kruskal–Wallis tests. This could be explained by the consistency of recorded and slight variances witnessed across the sampling sites, as well as values recorded by individual tankers.

Sulphate levels across the water tankers' supply chain ranged between 0.019 mg/L at the borehole being the lowest and 0.021 mg/L at the consumer endpoints being the highest. Thus, sulphate levels slightly increased towards the consumers, as shown in Table 4. The concentration of sulphate across tankers’ supply chain was a significant difference (p < 0.05) (Table 4).

Chloride levels across the tankers’ supply chain increased from 7.9 mg/L at the standpipe (lowest) to 17.5 ± 14.57 mg/L at the tankers (highest), as shown in Table 4. Nevertheless, chloride concentrations across the water tankers’ supply chain were not significantly different (p > 0.05) (Table 4).

The concentration of fluoride ranged from 0.08 mg/L at the standpipe (lowest) to 0.18 ± 0.10 mg/L at the tankers, being the highest with a slight decrease at the household level (Table 4). However, there was no significant difference (p > 0.05) in the concentration of fluoride levels across the sampling points (Table 4).

Iron and lead concentrations across the tankers’ supply chain were below detectable levels of 0.02 and 0.004 mg/L, respectively, at all sampling points (Table 4).

The faecal coliform densities increased across the water tankers' supply chain. They ranged between 6.4 ± 0.21 CFUs/100 mL at the borehole (lowest) and 199.4 ± 48.11 CFUs/100 mL at the tankers' (highest) with a slight decrease at the consumers (Table 4). There was a significant difference in the mean scores of faecal coliforms recorded across the supply chain (p < 0.05) (Table 4).

Pearson correlation analysis across water tankers’ supply chain revealed a significant correlation (p < 0.05) between faecal coliforms with DO (positive), turbidity (positive), nitrite (negative), ammonium (positive), and sulphate (positive) (Table 5). Temperature significantly (p < 0.05) correlated with nitrite (negative), ammonium (positive), and chloride (positive). DO correlated significantly with faecal coliforms (positive) and turbidity (positive). Electrical conductivity, salinity, TDS, chloride, and total alkalinity correlated positively with pH, whereas fluoride significantly correlated (p < 0.05) positively with turbidity.

The borehole and standpipe had the lowest temperature values due to the cold nature of borehole water witnessed in the field. Tankers recorded the highest temperature values attributed to sampling time, distance, material of tanks (steel), and influence by ambient temperature as observed in the field and a global review (Salehi 2022). Across the supply chain, temperature varied significantly inferring the factors above provided an ideal scenario for water to heat up at the tankers during transportation. At the consumers, the slight decrease in temperature was because most storage tanks were under a shade lowering water temperature, as similarly reported in a water quality review of domestic water storage tanks (Slavik et al. 2020).

Water tankers and consumer households recorded increased turbidity values across the supply chain. Turbidity levels significantly varied from the source to the consumers due to the unhygienic handling of tankers and consumer storage tanks. The presence of sediments and the formation of biofilm were observed in some water tankers, and customers’ storage tanks and containers. As the water was transported to consumers, dust and airborne substances entered the tankers/tanks as some tankers and tanks were not well-lidded, as observed in the field. This contributed to increased turbidity at both the tanker and consumer levels. Prolonged storage of water encourages the formation of biofilms or resuspension of TSS during discharge by the water tankers, as observed in the field and reported by Hoko (2008). Turbidity values obtained were within WHO's recommended levels of 5 NTU (WHO 2017) except for the consumers which recorded a slightly higher value of 5.90 NTU. Turbidity beyond the accepted guideline of 5 NTU was reported to favour high levels of E. coli counts and was associated with disease-causing organisms like pathogenic bacteria, viruses, and protozoans (Schafer et al. 2009; Sorlini et al. 2013; Awere & Anornu 2016). Their explanation agrees with the findings of this study due to the positive correlation between turbidity and faecal coliform densities.

The low DO levels at the borehole were due to high depth (negatively correlated), with less interaction with atmospheric oxygen. Groundwater DO depends on borehole depth, groundwater temperature, and oxidation-reduction potential of water, as reported in Hong Kong, China (Zan et al. 2019). The DO increased across the supply chain due to water mixing during transportation coupled with aeration by atmospheric oxygen since some tankers were not well covered. Enough DO in water indicates good water quality, good aeration, and less pollution according to Muhammad et al. (2023). The results obtained in this study were contrary to observations in Hong Kong, China which suggested that cold water had higher levels of DO, whereas warm water had lower levels (Zan et al. 2019). The expectation in this study was decreased DO across the supply chain compared with recorded increasing water temperature. According to the WHO (2017) report, high levels of DO exacerbate the corrosion of metal pipes or tanks, which was a risk water tankers in Lodwar town were exposed to. However, this observation was contrary to the result of total iron recorded in this study which was below detectable levels of <0.02. DO positively correlated with faecal coliforms in this study, contrary to the findings of Shamsudin et al. (2016) that a high concentration of DO increases microbial inactivation in water, hence inversely correlated. DO increased as faecal coliforms increased, which was attributed to the surrounding high temperatures and hence accelerated microbial activity.

Electrical conductivity across water tankers' supply chain slightly varied but was not significantly different. The EC values recorded across the supply chain were below the WHO acceptable guideline of 400 μS/cm (WHO 2017), suggesting the water was good for drinking. This finding was contrary to some boreholes water in Turkana which were reported in other studies to exceed WHO-acceptable guidelines (WHO 2017; Rusiniak et al. 2021). Tankers recorded higher EC values across the supply chain than other sampling sites. This was due to the dissolved ions leachates in the tanks, favoured by high temperature and pH recorded at the tankers responsible for chemical reactions like corrosion and oxidation of substances at the tankers. These results compare with other studies that EC increases as temperature and pH increase in borehole water in Bindura District, Zimbabwe (Hoko 2008).

The concentration of salinity slightly increased at the tankers by 0.01 ‰ across the supply chain although was not significantly different. This was attributed to leachates from the tanks as an indicator of some level of corrosion in the tankers. Furthermore, it was linked to increased concentrations of TDS, EC, and chloride recorded across the supply chain, as similarly reported by Hoko (2008) in borehole water in Bindura District, Zimbabwe.

The TDS recorded across the supply chain varied among the sampling sites but were not significantly different. Higher TDS levels at the tankers were due to the presence of leachates as revealed by increased sulphate, chloride, fluoride, EC, salinity, pH, and high temperatures at the tankers like results reported in other studies (Lukubye & Andama 2017). TDS concentration at the tankers was also influenced by the TDS concentrations recorded in individual tankers, as explained by a high standard deviation of 27.80 mg/L. The TDS concentration in this study was within WHO's desirable value of less than 600 and 1,500 mg/L for Kenyan standards in drinking water, hence of good quality (WASREB 2008; WHO 2022). Reduced TDS among the consumers was due to the mixing of water from other sources, as observed in some consumers' households.

Increased pH recorded at the tankers and consumers was attributed to increased hydroxide ions (OH⁻) from dissolved minerals such as calcium, manganese, and iron in the tankers during transportation. The pH values recorded at all sampling points varied significantly with tankers (7.71–8.36) and consumers (7.78–8.65) recording high ranges. This was an indicator of a significant influence of water pH from oxidation processes during transportation due to aeration and leachates in the tanks. According to Meride & Ayenew (2016), the level of pH in water was determined by the amount of dissolved carbon dioxide in water, forming carbonic acid. This may not be the case in this study as the pH recorded at all sampling points was slightly alkaline, above 7. Other factors reported to influence water pH are water sources, water storage tanks or vessel materials, temperature, mineral absorption, dust, the level of bacterial activity in a vessel, and duration of water storage before use (Packiyam et al. 2016; Manga et al. 2021). The pH at all sampling sites was within the WHO recommended range of 6.5–8.5 similar to the observation in Bengaluru, India where pH was within accepted standards in tanker water (Joseph & Sibi 2020).

There was no FRC detected at the borehole, standpipe, and tankers since there was no treatment of water undertaken at the sampling sites. This exposed water to contamination across the supply chain. FRC of 0.19 ± 0.40 mg/L was detected at the consumers' level attributed to individual household's ability to afford disinfectants such as Aqua tabs, water guards, or P&G purifiers of water, as observed in the field. The results obtained in the consumers' households were within the recommended FRC levels of 0.2 mg/L, effective in protecting water against microbial reinfection from the point of chlorination to the point of use (WASREB 2008; WHO 2017).

The slight increase of sulphate concentration across the water tankers’ supply chain (0.0189–0.0207 mg/L) was attributed to aeration which oxidizes sulphide to sulphate according to WHO (2017). The significant variance of sulphate concentration at the sampling points was due to the mineral deposits from storage tank leachates or re-suspended sediments which would result in bad taste, as reported in Uttarakhand, India (Kothari et al. 2021). Sulphate levels recorded in this study were below the WHO-recommended value of 250 mg/L in drinking water. This implies that the water delivered by tankers was safe, agreeing with results reported in Molo Town, Kenya (Chebet et al. 2020).

Total alkalinity increased across the water tankers’ supply chain with a slight decrease at the consumers attributed to vendors' water handling. At the tankers, high concentration of alkalinity was a result of contaminants containing carbonate, bicarbonate, and hydroxide ions such as calcium, magnesium, and microbial decomposition of organic matter, coupled with rusting of metal tanks forming iron oxide (Fe₂O₃). A similar observation was reported in Zamora, Mexico where increased alkalinity in water was associated with environmental factors or the presence of contaminants that contained carbonate, bicarbonate, and hydroxide compounds (Reyes-Toscano et al. 2020). The presence of these ions in water influences its buffering capacity to neutralize acids and bases, impacting the stability of pH levels. This was witnessed in increased levels of TDS, pH, EC, and salinity in this study. Similar findings were reported in Tamil Nadu, India where an increase in carbonates, bicarbonates, and hydroxide ions positively influenced total alkalinity in water (Jothivenkatachalam et al. 2010).

The presence of chloride in water increases water EC leading to an increase in its corrosivity and bad taste (WHO 2022). This assertion agrees with this study's finding as both EC and chloride increased at the tankers possibly exacerbating corrosion in tankers. Public drinking water standards require chloride levels not to exceed 250 mg/L (WASREB 2008; Kumar & Puri 2012; WHO 2022). The results obtained in this study were within permissible guidelines, according to Okoko & Idise's (2014) observations at Delta State University, Nigeria where chloride was within permissible standards in pipe-borne water. The slightly higher levels of chloride at the tankers were associated with tankers’ leachates, mixing of water from different sources and possible corrosion due to rust, according to the result of tanker water in Bengaluru, India (Joseph & Sibi 2020). Furthermore, they observed that 18% of water samples they collected from tankers had higher chloride content and hence did not qualify for drinking purposes, unlike this study.

Fluoride levels slightly varied across the supply chain but were not significantly different. At the borehole, the presence of fluoride was associated with the chemical weathering of minerals containing fluorine in parts of Turkana (Rusiniak et al. 2021). The results obtained in this study were within WHO's acceptable guideline value of 1.5 mg/L for drinking water (WHO 2022). Thus, vendors' handling of water (filling at the standpipes, transportation, cleaning of tankers, and transfer to consumer tanks) across the supply chain did not contribute significantly to fluoride level fluctuations in water. In other studies, as is the case in Turkana, fluoride levels in water were associated with the geology of the area and depth of the borehole, as observed in Hawassa, Ethiopia (Abdurahman & Zewdie 2018). Alarcón-Herrera et al. (2013) suggested that fluoride concentrations were influenced by alkalinity, bicarbonate concentrations, and electrical conductivity in groundwater, especially in semi-arid regions in Latin America unlike in this study. In Turkana, a previous study observed that some boreholes/wells had fluoride concentrations (range of 0.15–5.87 mg/L) exceeding WHO and WASREB recommended levels of 1.5 mg/L contrary to the findings of this study (Rusiniak et al. 2021). They noted that high levels of fluoride in drinking water may cause health-related risks like dental or skeletal fluorosis or hypocalcaemia when calcium fluoride (CaF₂) precipitates in human tissues.

Iron (Fe) and lead (Pb) concentrations recorded in this study were below WHO permissible guideline values of 0.03 mg/L (iron) and 0.01 mg/L (lead). This observation agrees with the findings of Kothari et al. (2021) in Uttarakhand, India. The findings of this study are contrary to the observations of Awere & Anornu (2016) in Cape Coast Metropolis, Ghana where higher levels of iron were associated with rust in water tankers. This study suggests that water tankers in Lodwar town had minimal rust, and hence were of good quality since iron concentrations were below detectable levels. The major source of lead in drinking water was mainly reported in plumbing materials, coupled with alkalinity and pH influence on lead and iron in water, as observed in drinking water pipes in Egypt unlike in this study (Lasheen et al. 2008).

Nitrogen derivatives (ammonium, nitrite, and nitrate) recorded low values across tankers’ supply chains. Temperature correlated significantly with ammonium and nitrite across the supply chain favouring oxidization of ammonia to nitrite revealing microbial activity that favours nitrification in water according to Rantanen et al. (2018). Nitrite levels in water are influenced by microbial action which reduces nitrate to nitrite in reduced oxygen conditions attributed to its significant distribution across the supply chain (WHO 2017). However, the nitrite levels obtained in this study were within WHO and WASREB permissible levels of less than 3 and 0.5 mg/L in drinking water, an indication of safe water.

Nitrate concentrations across the tankers’ supply chain were not significantly different, implying vendors’ water handling had minimal influence on nitrate concentration. There was a correlation between temperature with nitrite (negative), ammonium (positive), and chloride (positive), probably due to the oxidation of nitrogen compounds from increased temperature which was reported as a driver of nitrite seasonality (Rantanen et al. 2018). The nitrate levels across the supply chain were within WHO permissible levels of 50 mg/L in drinking water, according to the findings of Adimalla & Qian (2019) in Telangana, India. Similar results were reported by Meride & Ayenew (2016) in a water supply in the Wondo Genet campus, Ethiopia. Contrary to this study, other researchers reported some borehole water in Turkana exceeded WHO permissible standards for nitrate which may lead to health risks such as methaemoglobinaemia and thyroid effects (WHO 2017; Rusiniak et al. 2021).

Faecal coliform densities across the tankers’ supply chain increased from the borehole to the tankers with a slight decrease in the consumers attributed to management at the households indicated by the presence of FRC. The low levels of faecal coliforms at the borehole were attributed to protection from extraneous contamination by proper casings. In the case of incidental borehole contamination from flooding as reported in Anambra state, Nigeria and Southeastern Nigeria (Onuorah et al. 2019; Nwaiwu et al. 2020), inhibition of bacterial growth due to its cold temperatures was expected compared to other sampling points, as reported in sachet water in Accra, Ghana (Stoler et al. 2014). Contamination at the standpipe was due to poor handling and contaminated filling pipes as seen in the field. According to Opryszko et al. (2013), water quality degrades during transportation and storage agreeing with the findings of this study. Any type of unhygienic handling of water was expected to increase faecal coliforms along the supply chain unless appropriate interventions were made. This was due to recontamination by increased suspended solids, poor handling or inadequate residual protection that may exist across the transportation chain, as clearly portrayed by increased faecal coliforms in tankers. Researchers reported bacterial contamination in tanker water due to inadequate water disinfection in Makkah Al-Mokarama, Saudi Arabia and Bengaluru, India (Mihdhdir 2009; Joseph & Sibi 2020). This observation may apply to Lodwar town since water supplied to tankers and subsequently delivered to consumers was treated neither at the source nor by the tanker vendors.

Contamination among the consumers occurred when the tank was not cleaned, encouraging biofilm formation or recontamination when transferring water from one container to another as was reported by Awere & Anornu (2016) in Cape Coast Metropolis, Ghana. The decline of faecal coliform densities at the consumers' outlet was due to the treatment of water in some households through boiling or the addition of purifiers like water guards, aqua tabs, or/and filtration. The presence of faecal coliform densities in water associated with subsequent poor handling poses a public health threat (Elimelech 2006). The deterioration of water quality at consumers' storage facilities compared to the taps, standpipes, and borehole water was previously reported in Kenya and Ethiopia (Ayalew et al. 2014). Faecal coliforms across the supply chain were significantly different indicating that vendors' activities exacerbated contamination due to improper cleaning and poor sanitation, as reported in Gombe Metropolis, Nigeria (Kwami & Sawyerr 2018). Poor management at the source either or both by the water utility and tanker operators has been identified as the cause of poor quality of water delivered by tankers as reported by Cape Coast Metropolis, Ghana (Obeng et al. 2010).

Physiochemical characteristics and faecal coliforms were significantly different across the water tankers' supply chain in Lodwar town. This was attributed to water handling and hygiene, transportation, and storage between the source and consumers, which influenced changes in the levels of parameters assessed. The parameters varied with distance across the supply chain with lower values at the borehole/source and increased towards the consumers. Most parameters had higher levels at the tankers except turbidity, DO, pH, and FRC, which had higher levels at the consumers. There was no treatment of water at the source by the water utility or at the tankers by the water vendors. This study revealed inadequate water treatment processes and handling by the water utility, vendors, and consumers across the supply chain. Microbial contamination of water was identified as a key issue making the water unsafe for drinking; hence good handling of the water during transportation was necessary. Turbidity at the consumers and faecal coliforms at each sampling point across the supply chain exceeded permissible limits for drinking water. This poses a public health risk of waterborne diseases to consumers, especially when water was not treated before consumption as was the case in Lodwar town.

• A water quality monitoring system should be developed by the relevant County Departments like Public Health and Water Services within the tankers’ supply chain.

• The water utility in Lodwar town (TUWASCO) needs to treat water at the source for safe water supply and delivery by the water vendors.

• Adoption of a regulation for management of informal water vendors like tankers in Lodwar Town, and Turkana County.

Our sincere gratitude to the Rotary Club of Vienna (RCV) for the financial support through Florian Demmer (RCV Scholarship Counsellor). Appreciations to Michael Lopeyok, Boniface Onyango, Eng. Jackson Mutia, Erick Owino, Japeth Tembo, and Marvin Mukoma for their support in different capacities during the study. We thank Engineer Sepharinus Owino Onyango of Oxfam for the provision of the DelAqua water quality testing kit. Thanks to individual vendors and consumers who participated in this study; the County Government of Turkana and Turkana University College for their valuable support.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.