Combining inflow and infiltration assessment methods to enhance insights in data-scarce contexts: application to 46 Brazilian systems

Abstract

Excessive inflow and infiltration (I&I) can compromise sanitary sewer safety and resilience. I&I assessments support the design and operation of resilient sanitary sewers, but data scarcity in low- and middle-income countries hinders many traditional I&I assessment methods. This paper introduces a two-stage approach, leveraging multiple I&I assessment methods. First, I&I assessment methods are compared to reveal network and data concerns. Second, I&I methods are combined in inverse proportion to their sensitivity to variable (mis)estimation to improve accuracy. When combining water balance and pollutant mass flux methods, higher I&I assessments receive larger weights since they are less sensitive to variable misestimation. Our proposed combination of these methods (the hybrid method) had negligible bias (<0.2%) and improved accuracy (reduced mean average percentage error by >10 percentage points). We applied this hybrid method to assess quarterly I&I over two decades in 46 sanitary sewer networks in Brazil, combining water balance and pollutant mass flux methods (using biological oxygen demand). I&I consumed 24% of total sewer capacity and >50% of capacity in four networks. Unlike previous I&I research, we propose and demonstrate how to learn from and combine multiple I&I assessment methods; our approach proved well-suited for data-sparse environments, increasing accuracy and insights.