A fuzzy inference-based index for piped water supply service quality in a complex, low-income urban setting

Abstract

Sustainable Development Goal (SDG) 6 calls for universal access to safely managed drinking water services. We studied the evolution of the water supply service between January 2017 and December 2021 in the town of Uvira (South Kivu, Democratic Republic of the Congo) where large investments were made to improve the water supply infrastructure during this period, including a new 2,000-m³ tank, 56 community taps and 1,191 private taps. Across 16 geographic clusters in the town, we assessed water service accessibility, water quantity, continuity, and affordability, based on data provided by the construction team and the utility. We combined these dimensions into a single index ranging 0-100% according to rules defined using the fuzzy inference Mamdani method. Our results show that despite substantial increases in accessibility (i.e. proportion of households with a private tap or within 200 m of a community tap), overall service quality remained unsatisfactory, with a maximum index value of 38.1%, and worsened in many parts of the town due to limitations of the water production capacity after major flooding events and persistent electricity supply issues. The estimated amount of water supplied per user per day remained under 20 L during >95% of the observation period, with a decreasing trend. Pumps operated 58% of the time on average and the frequency of days without electricity supply increased over time. Our study highlights the important gap between upgrades in water supply infrastructure and improvements in the quality of service. The analysis of potential future scenarios for Uvira indicates that increasing production capacity is priority to improve overall service quality. Our results demonstrate that meeting SDG6 will be challenging in complex urban settings and will not only require sustained investments in water supply infrastructure but also in systems management and in energy supply.

Keywords: Cholera; Fuzzy inference; Infrastructure; Service; Wash; Water supply.