Consumer satisfaction on quantity and quality of water supply: a study in Matara, Southern Sri Lanka

People living in South Asian countries face many problems in obtaining safe drinking water. Some South Asian countries frequently experience outbreaks of diarrhea due to microbial contamination of drinking water (Ashbolt 2004; Khurana & Sen 2008). In addition, high arsenic concentration in drinking water encountered in Bangladesh has caused many health issues to people, including skin cancers and cancers in main organs, such as kidney, lungs, cardiovascular disease and problems in neural development in children (Mosler et al. 2010). Among the South Asian people, Sri Lankan people have access to relatively pure water for drinking, and experience diarrhea out breaks very rarely. However, contaminated drinking water is considered as one possible culprit for chronic kidney disease encountered recently in Sri Lanka (Chandrajith et al. 2010). At present, increased awareness about environmental pollution (Turgeon et al. 2004) and increase in detection rate of non communicable diseases such as chronic kidney disease has increased public concern over health and subsequent concern about their drinking water.

National Water Supply and Drainage Board (NWS&DB) is the main agency responsible for pipe borne water in Sri Lanka. They supply water to 34% of the country's population, while local authorities and community water supply schemes (WSS) accounts for another 10%. Most of the cities and their suburbs are supplied with treated water by NWS&DB. Rural population use their own water supplies, most frequently sourced from groundwater. Shallow wells and deep wells are used commonly in Sri Lanka (Chandrajith et al. 2010). People living in the dry zone of Sri Lanka (E.g. Rural population in Anuradhapura and Polonnaruwa districts) use water in irrigation tanks for bathing and household consumption. Drinking water quality needs of them to be obtained via reverse osmosis treatment to achieve high level of microelement organic and inorganic removals. Rainwater harvesting is also popular in the dry zone (E.g. Hambantota district) and government promote rainwater harvesting by providing technology and funding. However, rainwater collected from the roofing area may contaminate microbiologically (Lye 2002). Heavy metals and trace organics also may increase the danger of using rainwater as the drinking water source (Meera & Ahammed 2006). River water is the most common raw water source for drinking water supply in Sir Lanka. However, there are some areas in which irrigation tanks or boreholes are the source of raw water. Nilwala River is the source of raw water for Matara WSS.

Providing a reliable water supply is a challenge. Reliability is not only providing the demanded quantity of water. Reliable supply should provide water that can be consumed directly from the distribution network at anytime (Genius et al. 2008). WHO guidelines ensure that consumers are supplied with safe water (WHO 2011). Sri Lanka standards of potable water (2013) has been prepared by Sri Lanka Standards Institution. NWS&DB provide water in compliance with above standards. Water quality and quantity is interrelated. When supply is not continuous, water quality may deteriorate (Genius et al. 2008). Until early 1990s water quality standards were accepted as a scientific measurement ensuring water quality in global level (Doria 2010). Later, it is noticed, ‘standards should be based on the protection of human health and consumer acceptability’ (IWA 2004). From 1960s, researchers were interested in evaluating public perception of water quality (Doria 2010). Usually aesthetic quality is the basis of the judgment of a consumer; hence dissatisfaction on organoleptics (i.e. Sensorial information on taste, colour, odour and turbidity) may increase the risk perception (Jardine et al. 1999; Doria 2010). It has been observed that consumers’ age, income and level of education is influential on their risk perception of tap water (Turgeon et al. 2004). Levallois et al. (1999) and Turgeon et al. (2004) identified that dissatisfaction with the taste and the awareness of the raw water source increases the risk perception. Further, Turgeon et al. (2004), has been observed that geographic location and residual chlorine levels are important on consumer perception of the quality of tap water.

A major challenge faced by NWS&DB, Sri Lanka is providing demanded quantity of water. Rapid urbanization increases the number of water connections although the treatment plant (TP) capacities are not increased at the same rate. Therefore it leads to intermittent water supply or water shortages in local elevated areas when distribution reservoirs do not reach full supply level. Dissatisfaction on service supply may lead to water quality issues (Genius et al. 2008; Wedgworth et al. 2014). For example, intermittent supply may lead storage of water and hence water quality may deteriorate in storage. Therefore, the goal of this study is to examine the influence of the level of service supplied, water quality variations, geographical location of consumers and socio economic factors, on their perceptions on service delivery, satisfaction and risk perception. The study was limited to Matara WSS, Southern Sri Lanka.

The study was carried out in Matara city and its suburbs. Matara WSS has two water TPs located in Nadugala and Malimbada. Both TPs have distribution towers, which supply water for the proximity of each TP. Remaining water from both TPs is collected into a ground sump at Uyanwatta and then pumped to the distribution towers located at Nupe and Gabadaweediya and a ground sump located at Brownshill, which is the highest elevation in the distribution area. Both TPs receive raw water from Nilwala River. Nadugala intake is located downstream of the river and Malimbada intake is located several kilometers upstream. Treatment processes in both TPs are the same while the capacities of TPs are different (Nadugala TP – 6,500 m³/day and Malimbada TP – 45,000 m³/day). Matara WSS provides water for more than 100,000 connections. Even when the TPs are running at full capacity the production is not enough for supplying treated water to the supply area continuously. Water demand rises especially in dry and warm periods of the year due to increased consumption. At the same time, water levels at the intakes are low, hence reducing the production of treated water. Sometimes, even when a particular area is supplied, relatively higher elevations may not receive supply due to high consumption at relatively lower elevations in the same area. Subsequently, level of service may be different for the consumers living in the same area.

Distribution system length and water retention time in the distribution system vary with the area. Booster stations are located in the distribution system in order to maintain free residual chlorine (RCL) in water at desirable levels. Consumer awareness of the water treatment process and distribution as well as the socio economic characteristics of the consumers may also vary depending on the location and hence resulting in different perceptions. Hence, six areas were selected for the study based on their geographic location relative to the water distribution network and to represent urban and suburban areas of the city.

In order to investigate consumer perceptions on tap water quality and service delivery, a questionnaire was conducted from July to September 2015 (n = 583). The survey consisted of 25 questions about consumer perception, knowledge of their water utilities, pretreatment tendency and some socio demographic attributes. Hypothesizing that the effect of water age, chlorine re-boosting and socio economic variation in the urban and suburban areas will influence the consumer behavior, urban and suburban areas were selected for the study. In addition, attention was paid to select areas near the TP and distribution network extremity because pressure and RCL may vary. The gender, age, education level and income of the respondents were chosen as socio economic characteristics (Table 1). Further, number of family members, capacity of the household storage tanks, if any, and availability of alternative water sources such as dug wells were recorded.

Public perception and satisfaction on water supply service and water quality was studied in terms of several aspects. Organoleptics were directly evaluated with direct questions with yes/no type answers. Public satisfaction on water safety was evaluated with rated responses. Risk perception on chlorination and chemical addition is analyzed in the same way. Tendency to use home purification methods were analyzed with a direct question with answers (Table 2). In addition, details on perceived reasons for intermittent water supply, perceived pressure, average monthly water consumption, and the availability and capacity of water storage tanks and the availability of alternative water resources were gathered in the survey.

Logistic regression was used in statistical analysis of data. Answer options were dichotomized for using binomial logistic regression analysis, which minimizes the complexity of the model. Table 2 shows all the details on dichotomization of dependent variables. Explanatory variables were categorical variables with 2–6 categories (Table 3). This type of analysis is very useful in predicting the probability of occurrence of an event. Stepwise logistic regression was used in this analysis and SPSS 16.0 software package was used for analysis. In using stepwise procedure it evaluates the effect of each explanatory variable on the dependent variable being predicted (Hosmer & Lemeshow 1989). Explanatory variable selected at each step is the variable that contributes mostly to the increased significance of the regression based on the maximum likelihood ratio until all explanatory variables that have significant influence on the dependent variable is selected (Hosmer & Lemeshow 1989). In the current study, service satisfaction, taste satisfaction and perception on chemical addition and tendency of applying home purification methods were modeled against the selected independent variables (Table 3). Last category of each categorical explanatory variable was selected as the reference category in logistic regression (denoted REF). Model results are presented in the Table 4.

The impact of socio economic characteristics and geographic location was apparent on the consumer perception of the water quality. Satisfaction on water supply service delivery is dependent on the residing area of the consumer and perceived water pressure. It is surprising that it is not dependent on the level of service. People living in urban areas are less satisfied about the water supply service rather than people living in suburbs. However, percentage of houses having alternative water sources is higher in suburban areas, which may minimize the level of expectation of the consumer. In contrast, urban population has no alternative sources hence reliability of water supply is necessary.

Consumers are satisfied on the taste and perceive less risk when they are residing away from the TP, except the consumers residing in Mirissa area. It is surprising to observe that people living in Mirissa area are less satisfied about the taste of their tap water and perceive more risk in chlorination and chemical addition in water. Since Mirissa is a coastal area, the influence of tourism and subsequent observation of using bottled drinking water by foreign tourists may have an influence on their low satisfaction on taste and higher perceived risk. Influence of such external factors has been observed elsewhere also (Doria 2010). Further, taste satisfaction is dependent on the income and comparatively lower income group satisfies more. Same trend has been observed in Quebec City, Canada (Turgeon et al. 2004). Among the consumers who encounter a taste in their tap water, 84% claim that the taste is due to the presence of chlorine, while 13% claim that it is due to salinity. Further, it is important to notice that approximately 95% of respondents are satisfied on the safety of water, odour and clarity. Therefore these parameters were not included in multivariate analysis.

Risk perception on chlorine addition and chemical addition in water purification were separately evaluated because the awareness of people on these two may be different. However, people residing near the TP perceive more risk in chlorination and chemical addition. Their observations in residual chlorine levels and the awareness about the chemical added into water in the treatment process may influence this perception (Turgeon et al. 2004). While lower age groups perceive more risk in chlorination and chemical addition, gender is influential only on the perceived risk of chlorination.

In Sri Lanka people tend to apply home purification methods for their tap water, although the NWS&DB provide potable water comply with Sri Lankan standards and WHO guidelines. Common home purification methods applied in Sri Lanka is boiling and/or filtering. Application of home purification may improve the taste experienced due to the presence of chlorine. Although, consumers in lower age groups perceive more risk, the tendency to apply home purification methods is less. It is very important that consumers have an understanding of water treatment and water supply processes, so that perceived risk reflect real image. More respondents with a university degree are applying home purification methods than those who do not have. There were 11 respondents (out of 583), who do not drink tap water and they were removed from the analysis. Their drinking water source is groundwater wells located in their own garden. Availability of dug wells is high in suburban areas compared to the urban areas, because the size of the land per household is larger in suburbs compared to urban areas.

Geographical location is highly influential on the satisfaction and risk perception by the consumers in Matara WSS. Dissatisfaction on service delivery is a burning issue regarding Matara WSS. Matara WSS- Stage IV project is proposed by NWS&DB with the aim of increasing the capacity of water TPs. Unavailability of reliable continuous supply has resulted increased use of household storage tanks. The capacity of household storage tanks varied from 500–2,000L. Average per capita storage is 235 L, minimum being 71 L and maximum being 1,000 L. If average daily per capita consumption in Sri Lanka is considered to be 165 L (Gamini 2015), thus over storage in household storage tanks may create quality issues.

Associations between the risk perception and gender are often encountered with women perceiving higher risk than men (Doria 2010). Further, young respondents are more likely to be dissatisfied with tap water (Park et al. 2001, Doria 2010). Although this study fails to verify the relationship between education and income in risk perception, some authors have found that those are inversely associated with the risk perception (Doria 2010). In some countries it has been observed that bottled water consumption increases when public perception of drinking water supply is negative (Anadu & Harding 2000; Hu et al. 2011). In Sri Lanka, usually people do not use bottled water as their primary drinking water source. However, increase of bottled water consumption is not preferable, because it increases the pollution load on environment and water wastage (Linden 2015). Therefore it is necessary to evaluate the consumer satisfaction and implementation of corrective measures for improving the satisfaction.

Based on the survey and complementary information this study shows that area of residence and other socio economic parameters are highly influential on the perception of pipe borne water quality. Perceived pressure has a major impact on the satisfaction on service delivery and water consumption, hence indicating that perceived pressure has a close relationship with the discharge. Since perceived pressure is influenced on the service satisfaction, it is important to increase water supply pressure. Allowing distribution tanks to reach full supply level can increase pressure. Therefore intermittent supply with full supply level of the distribution tanks may increase the service satisfaction. However, this may have negative influences in the maintenance of distribution network and water quality because of the suction of air and/or groundwater into the system. It should be noted that the level of supply (continuous or intermittent) is not influencing the service satisfaction.

The results support the observation by other researchers that consumers living at the distribution extremities are less aware of water quality and water treatment process (Turgeon et al. 2004). Instead consumer satisfaction may be influenced by other factors, such as tourism related behaviours. However this study confirms the influence of socio economic factors on perception of drinking water quality. The analysis and comparison can be improved if these parameters at each respondent's location are measured in future studies.

The authors would like to thank NWS&DB for providing necessary information for the study. The study was partially funded by the University of Ruhuna.