Journal of Irrigation and Drainage Engineering

The ability of catchbasin sumps and hydrodynamic separators to remove sediment must be balanced with their ability to retain the previously captured material by preventing scour. The sediment scour process in these storm-water structures differs from the unidirectional scour and sediment transport process that occurs in pipes and channels. The hydrodynamics is affected by the particular characteristics of the hydraulic structure. The study of sediment scour in these devices requires incorporation of all the factors involved in the scour phenomenon and is best supported through computational fluid dynamic modeling (CFD) verified experimentally. Scour can be documented in the effluent as suspended sediment or total suspended solids concentration, a parameter of critical importance in storm-water quality management. This paper presents two simplified models for estimating effluent suspended sediment concentration attributable to scour of previously captured sediment. These models are based on results obtained from full-scale physical experimentation and calibrated and validated CFD modeling over a wide range of operating conditions where different particles sizes would be scoured. This paper also shows the effects of the armoring layer and of homogeneous and heterogeneous sediment sizes on the effluent concentration patterns.

A two-dimensional variable saturated flow model was developed to simulate subsurface flow in bioretention facilities employing the Richards’ equation. Variable hydrologic performances of bioretention are evaluated using the underdrain outflow hydrographs, outflow volumes for 10 storms with various duration and depth, and flow duration curves for 25 different storms. The effects of some important design parameters and elements are tested, including media type, surrounding soils, initial water content, ratio of drainage area to bioretention surface area, and ratio of cell length to width. Model results indicate that the outflow volume via underdrain is less than the inflow; the flow peak is significantly reduced and delayed. Underdrain outflow volume from loamy sand media (with larger $K_s$) is larger than that from sandy clay loam media. The saturated hydraulic conductivity, storage capacity, and exfiltration into surrounding soils contribute to the hydrologic performance of a bioretention cell. Initial media storage capacity is affected by the hydraulic properties of media soils, initial water content, and bioretention surface area. The exfiltration volume is determined by the surrounding soil type and exfiltration area, dominated by flow through the bottom of the media.

Fill media and excavation volume are the main costs in constructing bioretention cells, but the importance and impact of media depth in these systems is relatively unknown. Two sets of loamy-sand-filled bioretention cells of two media depths (0.6 m and 0.9 m), located in Nashville, North Carolina, were monitored from March 2008 to March 2009 to examine the impact of media depth on their performance with respect to hydrology and water quality. Construction and design errors resulted in the surface storage volume being undersized for the design event (2.5 cm). The actual surface storage volume was only 28% and 35% of the design volume for the 0.6-m and 0.9-m media depth cells, respectively. Overflow (bypass) occurred at least three times more frequently than intended. The exfiltration volume was much higher in the deeper media cells, presumably because of greater storage volume in the media and more exposure to side walls. Evapotranspiration (ET) plus exfiltration accounted for 42% of the inflow runoff in the 0.9-m media cells, while ET and exfiltration accounted for only 31% of the inflow runoff in the 0.6-m media cells. With the increase in exfiltration, the deeper media depth met a previously defined low-impact development (LID) hydrology goal of volume reduction more frequently than the shallower media system (44% of events compared to 21%). Larger outflow reduction consequently increased the reduction in pollutant loads. Estimated annual pollutant load reduction for total nitrogen, total phosphorus, and total suspended solids were 21, 10, and 71% for the 0.6-m media cells and 19, 44, and 82% for the 0.9-m media cells, respectively. Overall, nitrogen reduction was poor owing to suspected export of nitrate from the fertilizer use, and phosphorus removal was hampered because of irreducible concentrations in the inflow. Pollutant reduction was limited because the cells were undersized as a result of construction and design errors.

Recent research indicates that phosphorus (P) retention by bioretention systems comprising sandy media may not be effective for even a decade of urban runoff loads. To improve P retention for longer durations, this paper present findings from bioretention mesocosms using media amended with red mud (RM), a by-product of bauxite processing; water treatment residuals (WTRs), a by-product of water treatment; or Krasnozem soil (K), a highly aggregated clay soil. All treatments were vegetated except for one (K20nv). All treatments had outlets to restrict outflows except for one (WTR-Knr). To simulate the effect of long-term nutrient loads, the mesocosms were loaded weekly with secondary treated effluent with P concentrations averaging $3.3\mathrm{mg}\mathrm{\text{-}}{\mathrm{L}}^{-1}$. Over 80 weeks, this comprised hydraulic loads from 24.5 to $29.3\mathrm{m}\mathrm{\text{-}}{\mathrm{year}}^{-1}$ at a flow-weighted average between 2.8 and $3.2\mathrm{mg}\mathrm{\text{-}}{\mathrm{L}}^{-1}$ ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$, or mass loads from 1,115 to $1,284\mathrm{kg}\mathrm{\text{-}}{\mathrm{ha}}^{-1}$ ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$. These cumulative P loads represent the equivalent of over three decades of runoff loads. After 80 weeks, cumulative ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ retention in the K and RM soil treatments ranged from 79% to 95%, whereas ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ retention in the WTR treatments ranged from 95% to 99% of the input load. At 6-month intervals, the treatments were dosed with at least four sequential dosing runs of synthetic storm water with concentrations less than $0.8\mathrm{mg}\mathrm{\text{-}}{\mathrm{L}}^{-1}$ ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$. After 56 weeks of effluent loading, removal of ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ from storm water was negative 109% in the unvegetated K20nv treatment, compared with 33% retention in the corresponding vegetated K20 treatment. The K40 treatment with the most K retained 69% ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$, while the RM10 treatment with the most RM retained 78% ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$. After 80 weeks of effluent loading, removal of ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ from storm water was negative in both the K20nv and vegetated K20 treatments. The K40 treatment retained 76% ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$, and the RM10 treatment retained 55% ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$, while the total dissolved P (TDP) retention was 72% and 52%, respectively. After 110 weeks of effluent loading comprising $1,598\mathrm{kg}\mathrm{\text{-}}{\mathrm{ha}}^{-1}$ ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$, equivalent to 48 years of bioretention loads, ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ retention from storm water by the K40 treatment increased to 85%, and retention by the RM10 treatment increased to 91%. TDP retention also increased to 78% and 75%, respectively. These observations of P retention increasing after exposure to additional loads are uncharacteristic of typical sorption responses. After 80 weeks of effluent loading equivalent to 32 years of bioretention loads, the flow-restricted WTR-K treatment removed 99% of the storm-water ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ load while the corresponding free-discharge WTR-Knr treatment retained 94%. The WTR-K treatment was less effective than the WTR-Knr treatment in the earlier storm-water runs. The restricted outlet WTR30 treatment, which contained the most WTRs, retained 99% of the storm-water ${\mathrm{PO}}_4\mathrm{\text{-}}\mathrm{P}$ load. These high rates of P retention from storm water after accelerated P loads indicate that these amendments can provide effective P retention for the expected lifetime of bioretention facilities.

This research presents an initial study of the impacts of storm-water infiltration on a shallow unconfined aquifer at a bioinfiltration best management practice (BMP) on the campus of Villanova University. The study site is a vegetated infiltration basin with a 0.52 ha drainage area consisting of parking areas and recreational fields and features approximately 35% directly connected impervious area. The research utilized continuous monitoring of precipitation, groundwater elevation, and groundwater temperature in conjunction with surface water hydrologic modeling to assess the duration, magnitude, and extent of groundwater mounding at a storm-water infiltration BMP. Results indicate that precipitation greater than 1.80 cm causes increased mounding at wells adjacent to the site. In addition, it was found that precipitation less than approximately 1.80 cm leads to larger increases in groundwater elevation at an upgradient control well located near the edge of a large grass field. The extent of groundwater mounding is observed to be localized to the BMP and does not extend a significant distance downgradient. In addition, the magnitude and duration of groundwater mounding is related to both infiltration rate and groundwater temperature, such that cooler temperatures correlate to increased mounding. This study demonstrates the utility of groundwater monitoring for the purpose of BMP hydraulic performance assessment, and recommends that additional research be conducted in the future and that groundwater monitoring be considered for site monitoring plans.

Roof surfaces represent a significant portion of the impervious area associated with urban development. Storm-water runoff from those surfaces causes stream degradation in receiving waters attributable to excess volume of water runoff. This paper investigates the influence of roof surface type on storm-water runoff and specifically considers the benefits of a vegetated roof, or green roof, as a storm-water best management practice (BMP). Runoff data were collected over a 6-month period from three full-scale roofs, which were retrofitted with flow meters and automated water-quality samplers. The roof surfaces included an asphalt roof (for control purposes), a vegetated extensive green roof, and a stone ballasted roof. Both the green roof and stone roof were effective at reducing runoff volume and attenuating peak discharge, with the green roof being more efficient for rainfall events less than 2.54 cm. Overall, the green roof retained 68.25% of rainfall volume and reduced peak discharge by an average of 88.86%. Water-quality results were inconclusive, but did provide some indication that green roof systems could reduce nutrient loadings.

An assessment of the performance of four level spreader–vegetative filter strip (LS-VFS) systems designed to treat urban storm-water runoff was undertaken at two sites in the Piedmont of North Carolina. At each site, a 7.6-m grassed filter strip and a 15.2-m half-grassed, half-forested filter strip were examined. Monitored parameters included rainfall, inflow to, and outflow from each LS-VFS system. A total of 21 and 22 flow-proportional water quality samples were collected and analyzed for the Apex and Louisburg sites, respectively. All studied LS-VFS systems significantly reduced mean total suspended solids (TSS) concentrations ($p<0.05$), with the 7.6 and 15.2-m buffers reducing TSS by at least 51 and 67%, respectively. Both 15.2-m VFSs significantly reduced the concentrations of total Kjeldahl nitrogen (TKN), total nitrogen (TN), organic nitrogen (Org-N), and ${\mathrm{NH}}_4\mathrm{\text{-}}\mathrm{N}$ ($p<0.05$), whereas results were mixed for the 7.6-m VFSs. Significant pollutant mass reduction was observed ($p<0.05$) for all nine pollutant forms analyzed in Louisburg, which was caused by infiltration in the VFSs. The effects of VFS length and/or vegetation type are very important for pollutant removal, as effluent pollutant concentrations were lower (with one exception) for the 15.2-m VFSs. The median effluent concentrations for TN and total phosphorus (TP) for the four LS-VFSs were nearly always better than fair water quality benchmarks for the Piedmont of North Carolina, but only met good water quality metrics in one-half of the studied storm events.

Tools were developed for estimating costs of vegetative roofs, rainwater catchment systems, and bioretention facilities. These tools provide a detailed framework to facilitate cost estimation for capital costs, operation and maintenance costs, and life-cycle net present value. The tools can provide users with planning-level cost estimates and serve as a format for cost-reporting for past, current, and future projects. Very little cost data was available in the public forum, and prolific inconsistencies of supporting details were found in the available cost data. To address this, design assumptions were established for each facility type and professionally prepared cost estimates based on these design assumptions were used. Electives in design, such as plant selection and media depth, also greatly affected costs. To make the user aware of these effects, the model separates each option into line items that can be elected or excluded as appropriate. To facilitate collecting future cost data, best management practice (BMP) designers and builders should use these tools to record actual costs and report them to a clearinghouse such as the BMP Database.

Low impact development (LID) strategies are being encouraged in many communities as an approach to reduce potential adverse impacts of development on receiving streams. Many questions exist regarding how well various LID strategies perform in different settings, just as similar questions have been raised regarding performance of traditional stormwater best management practices (BMPs). Whereas historical focus on BMP performance has been water quality concentrations or loads, characterization of volume reduction benefits for both conventional and LID practices is increasingly an objective of researchers and stormwater managers. More than a decade ago, Urban Water Resources Research Council (UWRRC) members worked to develop a set of standardized monitoring and reporting protocols for traditional BMPs and to establish a master database for the purpose of evaluating BMP performance and the factors affecting performance. This effort culminated in the International Stormwater BMP Database (www.bmpdatabase.org), which contains data for more than 360 BMPs and continues to operate as a clearinghouse for stormwater BMP data and performance analyses. During 2008–2009, the International Stormwater BMP Database project expanded to better integrate LID into the database and develop a set of metrics that can be used to characterize BMP performance with regard to surface runoff volume reduction. This paper provides a condensed overview and progress report on the LID-focused effort, including the following topics: (1) monitoring guidance for LID at the overall site development level, (2) an overview of recent changes to the International Stormwater BMP Database to better accommodate LID studies, (3) a summary of LID studies currently included in the database, and (4) a proposed approach for evaluating performance of LID studies with regard to reducing surface runoff volumes.

Geysers, which involve the explosive release of water through vertical shafts connected to a nearly horizontal pipeline, have been attributed to either pipeline surge or the release of air. Laboratory experiments involving the release of a large entrapped pocket of air through a surcharged vertical riser indicate that the air can force water upward in the shaft but that a jet such as seen in video records of prototype systems does not form. This difference is attributed to processes that cannot be scaled down to the laboratory experiments. Data from a storm-water tunnel in Minneapolis that experienced a series of observed geyser events were analyzed. Measurements included pressures and velocity within the tunnel that can be correlated with observations on a videotape of the geysers. The pressure records do not indicate surge pressures sufficient to lift the water to the ground surface. Features of the pressure records can be interpreted to indicate the release of large air pockets through the manhole shaft similar to the laboratory experiments. These results suggest that the entrapment of large air pockets is an important component to the geysering process and that tunnel design procedures need to properly account for air effects.