Environmental Review Toolkit
Water, Wetlands, and Wildlife

 Previous Table of Contents Next     

Stormwater Best Management Practices in an Ultra-Urban Setting: Selection and Monitoring

3.10    New and Innovative Practices

3.10.1    Description and Purpose

A number of new BMP designs and design concepts are of potential interest to those managing ultra-urban runoff. Although these designs have been installed and operated at relatively few locations, the field trials clearly indicate noteworthy performance.

Each of the following sections gives a brief synopsis of an innovative practice, which may be sufficient for the reader to determine its applicability. In some cases, additional (more current) information can only be obtained directly from the proprietor of the equipment used in the design.

The practices described are alum injection systems, MCTT system, biofilters (e.g., StormTreat® System), vegetated rock filters, and vertical filter systems.

3.10.2    Alum Injection Systems

Alum injection systems (AISs) have been used successfully in treating urban stormwater runoff that was significantly impairing several lakes in the state of Florida. Their small footprint, relatively dependable components, and effectiveness on a wide range of pollutants make AISs worth considering for ultra-urban applications. Furthermore, AISs have been applied to treat entire watersheds, with drainage areas between 36 and 65 ha (90 and 160 ac).

Unlike other BMPs, an AIS treats common pollutants by chemically fixing them into an inert floc, which settles from the water column. The floc binds common stormwater pollutants into a nontoxic aluminum salt that is stable as long as the pH remains between 6 and 7. Alum is an acid salt of aluminum that has been extensively used for drinking water treatment, removal of phosphorus in wastewater treatment, and lake restoration projects.

Most AIS applications to date permit the floc to settle within the receiving water body, where it augments existing natural sediment. Given the relatively low concentration of alum addition (on the order of 10 mg/L) the aggregation of aluminum salt sediment in receiving water bodies is small. Furthermore, dissolved aluminum concentrations in water bodies receiving AIS-treated stormwater have been found to remain below levels judged to be toxic by USEPA.

Alum and any other additives needed to establish the proper pH are injected into the storm drain upstream of the receiving water body. If the floc is permitted to settle within the receiving water body, there is no need for costly settling chambers or for sludge removal and disposal. AIS typically consists of an alum storage tank storing liquid alum, pumps and piping to convey the alum to injection points, and flow rate monitoring equipment for controlling the injection feed rate. As a result, the capital cost of the system is relatively independent of the size of the system. The principal cost variation between different size systems stems from the variation in the amount of alum needed annually.

In an evaluation of the improvement in water quality due to alum treatment of stormwater inputs to Lake Ella, Florida, total nitrogen decreased by 78 percent (NH3 by 95 percent and NO2+NO3 by 14 percent), total phosphorus by 89 percent, and turbidity reduced by 89 percent (Harper, 1990).

3.10.3    MCTT System

The multi-chamber treatment train (MCTT) consists of a series of treatment units that mimic those found in a conventional wastewater treatment plant (Figure 33). The first chamber aerates the stormwater as it enters the treatment train and permits preliminary settling of larger diameter sediment. Stormwater is then conveyed to an inclined tray settler, where the majority of the settleable particulates are captured. Dissolved air flotation is then provided to help lift floatables and oil to absorbent media. The last step entails passing stormwater through a sand/peat filter.

Figure 33. General schematic of MCTT (Pitt, 1996)

Catchbasin: packed column aerators; Main settling chamber: sorbent pillows, fine bubble aerators, tube settlers; Filtering chamber: sorbent filter fabric, mixed media filter layer (sand and peat), filter fabric, gravel packed underdrain

The MCTT is applicable to small and isolated paved critical source areas from about 0.1 to 1 ha (0.25 to 2.5 ac). Gas stations, high traffic areas, and car washes are examples of land uses that could warrant this practice. As a relatively expensive BMP, the MCTT is reserved for those locations equipped with electric power and where regular maintenance is feasible. A recent retrofit installation cost $95,000 to tie an MCTT into an existing storm drain system for a 1 ha (2.5 ac) drainage area (Pitt, 1996). The cost to install would be lower if the installations were in new, developing areas and if prefabricated units became available.

During 13 storms monitored at a parking lot, the MCTT was found to remove 83 percent of total suspended solids, 100 percent of lead, and 91 percent of zinc (Pitt, 1996). In addition, the MCTT was found to be effective at removing toxicants: a 96 percent reduction was found in total toxicity as measured by the Microtox® screening test. As a result of its processes, ammonia nitrogen was found to increase by several times and the water gained a color due to staining from the peat medium.

In another study, 15 storms were monitored at a municipal maintenance yard where an MCTT had been installed to measure the pollutant reduction achieved by this device. The actual quantity of water passing through the MCTT consistently was found to be approximately 87 percent of rainfall volume. High pollutant reduction efficiencies were found for all particle-associated constituents, such as total suspended solids (98 percent) and total phosphorus (88 percent), and some dissolved constituents, such as dissolved zinc (68 percent). This municipal maintenance garage and parking facility is used primarily by garbage trucks, plows, and other heavy equipment (Greb et al., 1998).

The design of the MCTT is very site-specific and depends highly on local meteorology (e.g., mean inter-event periods, local rainfall intensity/duration relationships). The design challenge is to provide sufficient equalization capacity to ensure even inflow into the filter bed. As a result, there can be a 300 percent difference in the size of the MCTT depending on the facility location. The size of components is dependent on the depth of the facility and whether the facility will drain by gravity or be pumped dry. For most applications, the commitment of surface area will probably fall between 0.5 percent and 1.5 percent of land area (Pitt, 1996).

3.10.4    Biofilters

A recent design innovation, developed in the mid-1990s, uses biofilters for stormwater treatment. An example, in Figure 34, is the StormTreat® System (STS), which consists of a circular treatment tank (2.9 m dia. by 1.2 m tall) surrounded by wetland vegetation (Allard et al., 1996). First developed in 1994, STS uses sedimentation, filtration, and biological action to manage the common stormwater pollutants. Stormwater pretreated to remove large-diameter sediment is piped into the STS tank, where the captured runoff, is treated over the course of a 5- to 10-day period. Unlike most constructed wetlands, stormwater is conveyed into the subsurface of the wetland and through the root zone (Figure 34).

Figure 34. StormTreat® System Tank (adapted from StormTreat Systems, 1996)

Click here for text explanation

Based on manufacturer's literature, four standard-size STS tanks are required to manage the 12.7 mm (0.5 in) of stormwater generated by 0.4 impervious ha (1 ac) if pretreatment by preliminary detention is provided. In the absence of preliminary detention, 10 tanks are needed to manage the same volume of stormwater. Based on a footprint area of 3 m2 per tank (includes wetland vegetation area), the total commitment of land for this BMP is approximately two percent of the drainage area.

As expected, the removal efficiency of STS is high. The STS system has demonstrated total suspended solids removals of 95 percent and removal of metals ranging from 65 to 98 percent (Allard et al., 1996). Nutrients are also significantly reduced (total phosphorus by 89 percent, orthophosphate by 32 percent, and total dissolved nitrogen by 44 percent). Finally, STS has demonstrated a removal of fecal coliform of 83 percent, which is why it has been used to protect shellfish beds closed by high coliform levels.

Based on product literature, the cost to purchase STS and install a single tank is between $3,600 and $4,000 (1996 dollars). The maintenance costs have been estimated at $100 to $150 per tank cleaning, which is typically required every two to three years. This maintenance cost does not include the cost to remove sediment from any upstream pretreatment (e.g., catch basins).

Current design information can be obtained from the manufacturers web site: http://www.stormtreat.com.

3.10.5    Vegetated Rock Filters

Another recent design innovation for stormwater is the vegetated rock filter (VRF). Although wetland treatment systems similar to the VRF have long been used to treat wastewater, only since the mid-1990s has the design concept been applied to stormwater. A number of design variants exist for VRF; the basic design concept is also found in designs called the packed bed filter, rock-reed filter, vegetated submerged bed wetland, and shallow horizontal flow wetland.

Typically, the VRF design consists of a series of connected tanks filled with several feet of aggregate that are planted with wetland species (Figure 35). Stormwater flows pretreated to remove most suspended solids are introduced below grade into the aggregate, which is maintained in a saturated condition by carefully placed standpipes.

Figure 35. Vegetated rock filter (adapted from Claytor and Schueler, 1996)

Total depth is 10 cm, inflow is through perforated pipe into anaerobic zone, outflow is through perforated standpipe. Layers from the top are: water surface elevation, aerobic zone, anaerobic zone, muck layer for innoculation, and impermeable liner. Wetland plants are through the aerobic zone and into the anaerobic zone.

Stormwater treatment is provided primarily by biological action and root uptake. Anaerobic conditions, which help with denitrification, are generated in the lower depths within the rock filter (Kadlec and Knight, 1996). Under simulated rainfall conditions removal efficiencies of VRF systems have been found to be high: total suspended solids (95 percent), metals (21 to 80 percent), total phosphorus (82 percent), orthophosphate (14 percent), nitrate-nitrogen (75 percent), and fecal coliform (78 percent) (DRMP, 1995). These removal rates do not include any pollutant removed by pretreatment sedimentation.

Although VRF systems show promise for the removal of nutrients missed by other BMPs, the major drawback of this design is its space requirements. Test designs employed an off-line storage unit to capture the first flush and provide a steady inflow into the VRF (Eagan et al., 1995). It has been estimated the area needed for VRF is between three and five percent of the drainage area (Claytor and Schueler, 1996). This commitment of area may be too high for the typical ultra-urban application. Typical loading rates for the VRF are around 0.05 to 0.19 L (0.0125 to 0.05 gal) per minute per square foot of bed (DRMP, 1995).

Although specific information is not available, it is easy to state that the cost of a VRF is high when compared to other BMPs. In fact, some designs employ expensive pump systems to control dosing of multiple VRF units. However, the additional expense of VRF systems can result in consistent removal of nutrients (principally nitrogen) that might not be sufficiently removed by other less expensive BMPs.

3.10.6    Vertical Filter Systems

Stormwater BMPs that use vertically mounted filters are being developed at this time. Typically, vertical filter systems (VFSs) consist of a single large, concrete-lined chamber that serves as a combination storage and settling area. To one side of the chamber is a permeable wall, frequently constructed of sand layered between filter fabric and then sandwiched between sets of gabions (Figure 36).

Figure 36. Typical vertical filtration system (Tenney et al., 1995)

black and white picture of system. See above description.

The VFS concept has attracted interest because the design could provide filtration of stormwater in a facility that is smaller than the typical horizontal bed sand filter. The smaller facility footprint is obtained because the vertical filter can serve as one of the vertical walls of the chamber that stores captured stormwater.

Although laboratory, pilot, and field tests of the vertical filter design have been performed (Tenney et al., 1995), as of this time the design parameters have not been fully developed. Some of the design problems encountered relate to clogging of the geotextile fabric incorporated into the filter, loss of the sand medium due to high hydraulic pressures, and piping flow at the interface of the vertical filter and adjacent walls. In addition, designs are being augmented to minimize any resuspension of settled material that is slowly moved through the settling chamber and onto the vertical filter, accelerating clogging of the filter. These problems make it difficult to provide specific pollutant reduction rates for VFSs at this time.

Some design modifications under evaluation include installation of baffles within the storage chamber to minimize sediment transport and layered multi-media filters (compost, zeolites, sand) that are resistant to clogging and effective on a wide range of pollutants.


Allard, L.A., W. Graham, W. Platz, R. Carr, J. Wheeler. 1996. The StormTreat System Used as a Storm Water Best Management Practice. In Proceedings of Watershed '96.

Claytor, R.A., and T.R. Schueler. 1996. Design of Stormwater Filtering Systems. The Center for Watershed Protection, Silver Spring, MD.

Dyer, Riddle, Mills & Precourt, Inc. (DRMP). 1995. Packed Bed Filter. Final Report, Project No. SP214. Prepared for City of Orlando Stormwater Utility Bureau. Submitted to the Florida Department of Environmental Protection.

Egan, T.J., S. Burroughs, T. Attaway. 1995. Packed Bed Filter. In Proceedings of 4th Biennial Symposium on Stormwater Quality. Southwest Florida Water Management District, Brookeville, FL.

Greb, Steven R., Roger T. Bannerman, and Steven R. Corsi. 1998. Evaluation of the Multi-Chambered Treatment Train (MCTT) in Milwaukee, Wisconsin. Prepared for the U.S. Environmental Protection Agency, Region V, Chicago, IL.

Harper, H. H. 1990. Long-term Performance Evaluation of the Alum Stormwater Treatment System at Lake Ella, Florida. Final Report for FDER Project WM339, submitted to Florida Department of Environmental Regulation, Tallahassee, FL.

Harper, H.H., and J.L. Herr. 1992. Stormwater Treatment Using Alum. Public Works.

Kadlec, R., and R. Knight. 1996. Treatment Wetlands. CRC Press/Lewis Publishers. Boca Raton, FL.

Pitt, R.E. 1996. The Control of Toxicants at Critical Source Areas. In Proceedings of Effects of Watershed Development and Management on Aquatic Ecosystems Symposium, presented by the Engineering Foundation, Snowbird, UT, August 4-9, 1996.

Schueler, T.R. 1996. Vegetated Rock Filter Treats Stormwater Pollutants in Florida. Watershed Protection Techniques 2(2): 372-374.

StormTreat Systems. 1996. Product Literature.

Tenney, S., M.E. Barrett, J.F. Malina, Jr., R.J. Charbeneau, and G.H. Ward. 1995. An Evaluation of Highway Runoff Filtration Systems. Technical Report CRWR 265. Center for Research in Water Resources, The University of Texas at Austin, Austin, TX.

Back to top

Questions and feedback should be directed to Susan Jones (Susan.Jones@dot.gov, 202-493-2139) and Marcel Tchaou (Marcel.Tchaou@dot.gov, 202-366-4196).

HEP Home Planning Environment Real Estate

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000