Evaluating Green Infrastructure Stormwater Capture Performance under Extreme Precipitation

Abstract

The use of green infrastructure (GI) for urban stormwater management has become a big industry, with cities like New York and Philadelphia planning to invest more than a billion dollars over multiple decades into this distributed approach to runoff reduction. Throughout the northeast US, GI systems are typically sized to fully capture all runoff generated within their tributary areas during approximately 90 percent of all wet weather events occurring annually (e.g., [Formula: see text] 25–30[Formula: see text]mm of precipitation). Though many claim that retrofitting such GI systems into urban landscapes will also help cities adapt to climate change, few researchers have actually attempted to document GI facility performance during more extreme precipitation. In this study, the stormwater capture performance of a bioretention facility located in Queens, New York City was evaluated under non-extreme and extreme precipitation conditions occurring between 2011 and 2014, including Hurricane Irene and Superstorm Sandy. Performance was found to be highly variable from event to event. The site rarely ponded, and overflowed only once (during Irene), for a short time (e.g., 10[Formula: see text]min), generating an insignificant volume (0.085[Formula: see text]m3) of overflow, likely because of the high infiltration capacity of in-situ, sandy soils and the facility’s low hydraulic loading ratio (3.8:1). Though the facility was able to infiltrate nearly all the runoff that it receives through its inlet, field monitoring suggests that site performance is often hindered by inlet bypass, not soil saturation or overflow. The site captures 70, 77, and 60 percent of all runoff generated within its tributary area during all events ([Formula: see text]), just the non-extreme events ([Formula: see text]), and just the extreme events ([Formula: see text]). The facility also regularly receives and infiltrates runoff originating outside of its designed tributary area. A regression analysis suggests that storm duration, total amount of precipitation and peak-hourly intensity are significant predictors (p-value [Formula: see text]) of, and negatively correlated with, the facility’s stormwater capture performance. The analysis also anecdotally suggests that performance might improve as vegetation gets established and/or as the frequency of maintenance visits are increased. Though this study suggests that this facility does attenuate a significant amount of runoff extreme precipitation, future work will investigate factors other than the climate which could be triggering inlet bypass occurrence, in particular GI maintenance frequency.