| Bioretention cells are widely used as an infiltration-based stormwater control measure to reduce the negative impacts of urban stormwater runoff. Two sets of cells were monitored at Rocky Mount and at Nashville to measure the effects that underdrain configuration, media depth, surface storage volume, and underlying soil type had on hydrologic and water quality performance. Both sites are located in the Upper Coastal Plain of North Carolina, where insitu soils tend to have a high sand content.;The two bioretention cells at Rocky Mount were designed with an internal water storage (IWS) zone and had varying degrees of sandy underlying soils. The underlying soils for these two cells were sand (Sand cell) and sandy clay loam (SCL cell). After the first year, the IWS zone depth was reduced by lowering the outlet. While the Sand cell, with its sandy underlying soils and deep IWS zone provided greater outflow reduction, it had minimal nitrogen treatment because of a short hydraulic residence time (less than three hours). On the other hand, the SCL cell had a longer hydraulic residence time (up to seven days), and it had significant concentration reductions for all forms of nitrogen, including nitrate.;Fill media is a major expense, so a study objective from the Nashville site was to evaluate the impact of varying media depth (0.6 m versus 0.9 m). A post-construction objective was to analyze the impact of under-sizing the surface storage zone. Construction and design errors resulted in the surface storage volumes of the bioretention cells to be approximately one-third of the design volume; moreover, the surface was clogged, which limited infiltration. After one year of monitoring, the clogging layer was removed, which doubled the surface storage volume. Deeper media depth promoted more exfiltration and met a low impact development goal of outflow reduction twice as often. Also, despite being relatively undersized, the repaired cells were able to treat nearly 90 percent of runoff, suggesting that current design guidance may be over-sizing the surface storage zone. Also in Nashville, another bioretention cell was installed in series with a pervious concrete system that included a subsurface storage zone. These two practices in series had excellent peak flow and outflow reduction. For low impact development (LID) practices in series, serious consideration should be taken to balance the returns of flow rate and outflow reduction vis-avis cost. The bioretention cell was installed at a site with a high water table, so this impact was quantified. Because of the intercepted groundwater, the site exported 63 percent more total nitrogen than what was present in the runoff load.;Overall, bioretention cells can be designed and constructed with a variety of specifications, among them are media depth, underdrain configuration, media composition, drainage area to bioretention area ratio, and surface storage volume. One way to quantify various designs is using a long-term model. The hydrology data from the field sites were used to calibrate and validate DRAINMOD, a widely-accepted, long-term drainage model. The measured and predicted (modeled) results were in good agreement during both the calibration and validation periods; Nash-Sutcliffe coefficients for runoff, drainage, overflow, and exfiltration/evapotranspiration commonly exceeded 0.8. These results proved that DRAINMOD can be reliably used to simulate the hydrologic response of runoff entering a bioretention cell. With a reliable long-term bioretention model, designers and regulators will be able to shift from the current "one size fits all" design approaches and establish a "flexible" bioretention design methodology that is based on underlying soil type, design specifications, and climate.;Finally, in an attempt to improve construction practices, the effects of construction activity on underlying soils were explored. The effects of soil type, soil moisture, and excavation technique were tested. The results showed that excavating using the teeth of the bucket to scarify the surface (rake method) would maintain a more permeable surface than using the back of the bucket (scoop method). |
