2020 journal article

Enhanced bioretention cell modeling with DRAINMOD-Urban: Moving from water balances to hydrograph production

JOURNAL OF HYDROLOGY, 582.

By: W. Lisenbee*, J. Hathaway*, L. Negm n, M. Youssef n & R. Winston*

co-author countries: United States of America πŸ‡ΊπŸ‡Έ
author keywords: Bioretention; DRAINMOD; Modeling; Stormwater; Urban hydrology
Source: Web Of Science
Added: March 30, 2020

Bioretention systems have become a leading stormwater control measure for mitigating urban hydrology. Although these systems have performed well in many site-scale field studies, less investigation has been directed toward effectively modeling these systems. This is critical, as modeling of bioretention systems provides an avenue for evaluating their effectiveness prior to devoting time and resources into installation. Many hydrologic models capable of simulating bioretention consist of lumped parameters and simplifications that do not fully account for fundamental hydrologic processes such as soil-water interactions. DRAINMOD has shown promise for obtaining detailed daily water balances within bioretention systems under continuous simulations. One significant advantage of DRAINMOD is that it uses the soil-water characteristic curve to account for fluctuations in soil moisture instead of assuming saturation; however, the model historically only produces daily outputs. For this study, DRAINMOD was modified to develop DRAINMOD-Urban, which allows high temporal resolution inputs and outputs, more closely matching the residence time of runoff in urban systems. DRAINMOD-Urban simulations of a bioretention cell in Ohio, USA, revealed that DRAINMOD-Urban could effectively produce hydrographs with a cumulative Nash-Sutcliffe Efficiency (NSE) of 0.60 for the 12 events that produced drainage over a 7-month monitoring period. Overflow was also modeled by DRAINMOD-Urban, but additional overflow data are necessary to derive conclusions about model effectiveness in predicting this hydrologic component. Input parameters previously calibrated for the DRAINMOD model did not translate well to DRAINMOD-Urban with the top-down approach applied in this study (NSE = 0.31 for drainage and NSE = βˆ’1.83 for overflow), but the bottom-up approach showed that parameters calibrated with DRAINMOD-Urban (NSE = 0.60 for drainage and NSE = βˆ’0.1 for overflow) could be used in DRAINMOD to obtain reasonable drainage volumes (25.6% error compared to measured values). This study suggests DRAINMOD-Urban is an effective tool for modeling bioretention hydrographs and demonstrates the importance of temporal scale in bioretention modeling by illustrating multiple model calibration approaches. Despite the promising results of this study, additional studies are recommended where validation of the model is performed at more sites, in particular for events with overflow. Further, sensitivity analysis of input parameters and comparison of DRAINMOD-Urban to other commonly used bioretention models would inform future modeling efforts.