Modeling RDII Without the R: Measuring Return Periods from Flow Monitoring Data

Urban sanitary sewer systems routinely face significant uncertainties associated with Rainfall-Derived Inflow and Infiltration (RDII). Conventional design approaches, often based on fixed RDII allowances or rainfall return-period assumptions, may fail to reflect the observed variability and scale-dependent behaviour of wet weather flows, especially in small-to-medium sewersheds. This study presents the results of a statistical frequency analysis of flow monitoring data to characterize actual RDII observations and offers an alternative approach to estimating such flow contributions for design purposes, thereby circumventing the need to infer wet weather response solely from rainfall return periods.

Leveraging an extensive dataset from over 250 monitored sewersheds in York Region, Ontario, RDII events were identified using an extraction method which considers rainfall quantity thresholds, interevent times, and the subtraction of dry weather flow as determined through a moving window average. Peak hourly RDII responses were normalized by contributing area and analyzed across a wide range of basin sizes and, for each monitoring site, four probability distributions (GEV, Gumbel, Pearson III, and Log-Pearson III) were fitted to the RDII series. Goodness of fit performance metrics, such as the KS test, RMSE/MAE, and QQ-plot consistently indicated that the Log-Pearson III best captures the heavy-tailed and skewed nature of RDII. Confidence intervals and return-period quantiles were determined through statistical bootstrapping, allowing for estimation of rare but impactful wet weather events.

Results reveal a clear non-linear scaling relationship between RDII magnitude and sewershed area: small and medium sewersheds (e.g., < 400 ha) can sometimes exhibit disproportionately high RDII compared to large basins and with much higher variability, and the uniform deterministic allowance of 0.26 L/s/ha is observed to often under-predict RDII for many such drainage areas. Based on these findings, a power-law RDII design curve calibrated to observed monitoring data is proposed, offering a scale-aware, risk-informed alternative that reflects actual system behaviour.

This methodology enables municipalities and engineers to design sanitary infrastructure grounded in observed wet weather performance rather than assumptions on the rainfall-to-RDII transformation process(es). By directly modelling RDII return periods, the data-driven approach allows for capacity planning that duly accounts for scale-dependent variability. Given the widespread problem of inflow and infiltration and the continuously increasing collection of monitoring data, this approach offers a transferable framework for characterizing RDII for sanitary (and combined) sewer analyses and design in any jurisdiction.

 Click here to download a static PDF version of the presentation.