Joshua A Steele ^1* Adriana Gonzalez-Fernández ¹ John Griffith ^1* Darcy Ebentier Mccargar ² Sierra Wallace ² Kenneth Schiff ¹

• ¹ Southern California Coastal Water Research Project, Costa Mesa, United States
• ² WSP (United States), New York, United States

Inflow and infiltration are well-known issues for sanitary sewer collection systems, but exfiltration is understudied and rarely empirically quantified. The goal of this study is to estimate the potential human fecal contribution from wastewater exfiltration from sanitary sewers to stormwater in an urban watershed with separate sanitary sewer and storm sewer systems. This study uses newly developed techniques to empirically measure sanitary sewage exfiltration, then compares these exfiltration rates to human fecal pollutant loading in stormwater runoff from multiple urban catchments without other sources of human inputs (i.e., no septic systems, no homeless encampments, no reported sanitary sewer overflows) to estimate the amount of exfiltrated sewage that reaches stormwater. The human-specific genetic marker HF183, which is highly concentrated in raw sewage, was used as a surrogate for human fecal pollution and was measured in nearly every stormwater sample collected. We extrapolated measured exfiltration to the entire 419 km 2 watershed and estimated up to 4.25 x 10 6 liters exfiltrate each day. This is 0.6% of the average daily volume of sewage treated in this sewer collection system and is similar in scale to exfiltration allowed by design standards. Based on ratios of exfiltration loading predictions vs stormwater loading measurements, the proportion of exfiltrated human fecal load that is estimated to be transported via subsurface pathways (i.e. the subsurface transfer coefficient, STC) to stormwater in the studied catchments is 8.27 x 10 -5 (95% CI: 6.30 x 10 -5 to 1.37 x 10 -4 ). Human fecal pollution loads from exfiltration via subsurface transfer during a storm event were calculated to be 1.5 x10 13 (95% CI: 1.79 x10 12 to 3.59 x 10 13 ) HF183 gene copies per storm. This estimate is similar in scale to the measured mass loading estimates in stormwater for the studied watershed and comparable to independently-measured tracers of sewage. Future work is needed to better understand subsurface transport mechanisms of exfiltrated sewage and to test this approach, and the assumptions used, in other watersheds and sewer systems.