A Flexible Framework for Hydraulic of Stormwater Green Modeling Infrastructure performance

In this paper, the application of an integrated modeling system GIFMOD is presented for evaluating the effect of a green infrastructure composed of a dual rain garden system on reducing the volume and peak flow of storm runoff collected from a steep wooded area and a parking lot. The model is composed of the catchment where the runoff is generated and a detailed representation of the rain garden system. Flowrates entering the upper rain garden and leaving the upper and lower rain gardens measured over a period of 45 days (including six major rain events) were used to calibrate the model. The calibration was done probabilistically by using Bayesian Markov Chain Monte Carlo. Two different schemes were used for modeling the catchment to obtain the storm runoff inflow to the gardens, the first scenario considered the entire pervious catchment as a single reservoir, and a second scenario, discretized the pervious catchment into three connected sub-catchments. Each rain garden was represented by five vertically connected layers (blocks), where each layer represents a different spatial feature including zones for surface storage, engineered soil, storage, native soil, and groundwater. The agreement between the modeled and measured flow rates was acceptable as evaluated based on Nash-Sutcliffe coefficient with the range between 0.71 to 0.86. With the parameterized and calibrated model the long-term effectiveness of the green infrastructure at reducing the peak flow and the volume of storm runoff entering the sewage/storm collection systems can be evaluated.