Calibration and validation of DRAINMOD to model bioretention hydrology

SummaryPrevious field studies have shown that the hydrologic performance of bioretention cells varies greatly because of factors such as underlying soil type, physiographic region, drainage configuration, surface storage volume, drainage area to bioretention surface area ratio, and media depth. To more accurately describe bioretention hydrologic response, a long-term hydrologic model that generates a water balance is needed. Some current bioretention models lack the ability to perform long-term simulations and others have never been calibrated from field monitored bioretention cells with underdrains. All peer-reviewed models lack the ability to simultaneously perform both of the following functions: (1) model an internal water storage (IWS) zone drainage configuration and (2) account for soil–water content using the soil–water characteristic curve. DRAINMOD, a widely-accepted agricultural drainage model, was used to simulate the hydrologic response of runoff entering a bioretention cell. The concepts of water movement in bioretention cells are very similar to those of agricultural fields with drainage pipes, so many bioretention design specifications corresponded directly to DRAINMOD inputs. Detailed hydrologic measurements were collected from two bioretention field sites in Nashville and Rocky Mount, North Carolina, to calibrate and test the model. Each field site had two sets of bioretention cells with varying media depths, media types, drainage configurations, underlying soil types, and surface storage volumes. After 12 months, one of these characteristics was altered – surface storage volume at Nashville and IWS zone depth at Rocky Mount. At Nashville, during the second year (post-repair period), the Nash–Sutcliffe coefficients for drainage and exfiltration/evapotranspiration (ET) both exceeded 0.8 during the calibration and validation periods. During the first year (pre-repair period), the Nash–Sutcliffe coefficients for drainage, overflow, and exfiltration/ET ranged from 0.6 to 0.9 during both the calibration and validation periods. The bioretention cells at Rocky Mount included an IWS zone. For both the calibration and validation periods, the modeled volume of exfiltration/ET was within 1% and 5% of the estimated volume for the cells with sand (Sand cell) and sandy clay loam (SCL cell) underlying soils, respectively. Nash–Sutcliffe coefficients for the SCL cell during both the calibration and validation periods were 0.92.

► Bioretention hydrology was modeled with DRAINMOD, a widely accepted drainage model. ► Four bioretention cells monitored for 2 year periods were used in calibration. ► DRAINMOD can model an internal water storage (IWS) zone configuration. ► It can be used to predict bioretention hydrology on a continuous, long-term basis. ► In the validation period, Nash–Sutcliffe coefficients commonly exceeded 0.7.