Simulation of the cumulative hydrological response of green infrastructure

As a first product of post-doc Pedro Avellaneda’s work at Kent State, we were pleased to submit the following abstract to the 2016 Low Impact Development Conference. A short version of the work will be published as a conference paper in August, and we are preparing a more detailed version for journal submission this summer.

Simulation of the cumulative hydrological response of green infrastructure

P.M. Avellaneda, A.J. Jefferson, and J.M. Grieser

The performance variability of green infrastructure is connected to the changing dynamics in rainfall-runoff processes. Because of these dynamics, a green infrastructure facility experiences a range of rainfall-runoff events that are difficult to fully capture during a monitoring program. In this study, we evaluated the cumulative hydrologic performance of green infrastructure in two residential areas of the city of Parma, Ohio, both draining to a tributary of the Cuyahoga River. Green infrastructure involved the following spatially distributed devices: 19 street side biorententions, with surface area ranging from 26 to 44 m²; 5 rain gardens, with surface area less than 25 m²; and 43 rain barrels. The engineered soils for the bioretentions and rain gardens consisted of sand (~72%), organic matter (5-28%), and clay (~10%). Data consisted of rainfall and outfall flow records for a wide range of storm events, from 0.5 mm to 89 mm of measurable precipitation, monitored over three years, including pre-treatment (~1 year) and post-treatment periods (~2 years). The Stormwater Management Model (SWMM) was calibrated and validated to predict the hydrologic response of green infrastructure. Optimized parameters, in the form of Posterior Probability Distributions (PPDs), were used to estimate flow attenuation over multiple years of precipitation data in order to capture the complex rainfall-runoff dynamics. The hydrologic performance of green infrastructure was evaluated by statistically comparing the non-exceedence probability plot for pre-treatment and post-treatment outfall flow rate scenarios. In addition, the following probability plots were estimated: (1) ratio of pre-control and post-control total runoff volume, and (2) ratio of pre-treatment and post-treatment maximum peak flow rate. Parameter and predictive uncertainties were inspected by implementing a Bayesian statistical approach. Overall, the cumulative hydrological response of green infrastructure was positive: reductions of up to 33% of peak discharge and 40% of total run-off volume were estimated.