A Modeling Framework for Bioretention Analysis: Assessing the Hydrologic Performance under System Uncertainty

Abstract

Bioretention systems are one of the most common low-impact development (LID) facilities. In this paper, we develop a modeling framework that combines the one-dimensional Green and Ampt model with the outlet modeling of weirs, underdrains, bottom, and lateral exfiltration. This framework is specially designed to be applied in poorly gauged watersheds or where continuous simulations are intractable due to a lack of data. First, we calibrate and validate the model using field data. The Nash–Sutcliffe efficiency (NSE) indicator comparing observations with simulations varies from 0.58 to 0.83, while Pearson’s correlation coefficients (r2) vary from 0.91 to 0.98, indicating good agreement. Following, we assess various hydrologic performance indicators using the model by carrying out different analyses such as (1) defining critical rainfall duration for design storms, (2) assessing the flood performance of bioretention designed with predesign methods, and (3) assessing the optimal drainage areas that a bioretention could receive inflows from. Subsequently, we used the model to (4) perform a one-at-time sensitivity analysis identifying the single most sensitive parameters in bioretention hydrologic performance; and (5) assess the combined influence of parameters by performing Monte-Carlo simulations. The results indicate that the optimal impervious area to the drainage area is approximately 6% for the climate of São Carlos - SP. Furthermore, the critical rainfall duration for lot-scale bioretention with a subtropical climate is between 60 and 120 minutes. The bioretention surface area, the depth of the surface layer, and the saturated hydraulic conductivity of the medium were the parameters most sensitive to hydrologic performance. The results of the model comparison with the SWMM software show that both models are similar in infiltration modeling but fundamentally different in percolation conceptualization, causing a difference in the hydrographs between the developed model and the SWMM model. Finally, the results of the design optimization modeling show a simple mathematical procedure to optimally design bioretention by minimizing construction, maintenance, use of land, and other associated costs while guaranteeing expected design hydrological performance.