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

Fact Sheet - Surface Sand Filters

The surface sand filter has been employed since the early 1980s to provide stormwater quality management. One of the forerunners in developing the surface sand filter design has been the City of Austin, Texas. As shown in Figure 14, the Austin design consists of a bypass chamber, a sedimentation chamber that provides pretreatment, a flow distribution cell, and a sand filter bed. The design illustrated shows many of the features common to surface sand filters. Typically, the filter bed has a 450 to 600 mm (18 to 24 in) deep sand layer that traps or strains pollutants before runoff is collected in an underdrain system (gravel and perforated pipe) and conveyed to a discharge point.

Figure 14. Austin sand filter with full sedimentation protection (Young et al., 1996)

A bypass chamber is used to protect the BMP from high inflows, diverting any flow in excess of the capacity of the structure. This works with the sedimentation cell(s) to prevent high loads of coarse sediment from entering the filter bed. While the design illustrated in Figure 14 consists of concrete structures/walls, earthen walls backed with geomembranes and riprap sections can be substituted in the basic design. In terms of drainage area, the Austin design has been successfully employed for drainage areas ranging from 0.4 to 40.5 ha (1 to 100 ac).

Surface sand filters are very well suited to managing the first flush volume, which typically contains the highest concentration of pollutants. However, the design is poorly suited to providing stormwater quantity management to prevent flooding because high flows can easily damage the filter bed. As a result, it is strongly recommended that the design be installed in an off-line configuration.

The Austin filter works by a combination of sedimentation, filtration, and adsorption. The sedimentation section located just upstream of the filter section serves as pretreatment, removing larger-diameter suspended solids. Partially treated stormwater then flows slowly into the filter section, where fine-grain material is strained from the stormwater as it passes through the filter medium. The sand medium filter traps up to 90 percent of the small particles in stormwater runoff (6 to 41 microns) if a 460 mm (18 in) layer of sand is used. However, the extent of adsorption by sand of some dissolved pollutants is relatively small when compared to other filter media. For example, sand medium adsorbs much less positively charged dissolved metals and hydrocarbons than either soil or peat medium primarily due to its relatively low cation exchange capacity (CEC); sand has a CEC that is 13 percent that of the soil medium and 0.002 percent of the peat medium.

Applicability

Although it has been applied within an urban setting, the Austin sand filter may require a significant commitment of land area (generally between two and seven percent of the drainage area). Consequently, many of the installations within the City of Austin are in newer, less densely developed portions of the municipality. Within an ultra-urban setting this design might be restrictive requiring a completely subsurface BMP (see underground sand filter design in the Underground Sand Filters Fact Sheet).

The applicability of surface sand filters to roadway projects has been demonstrated. For example, the Texas Department of Transportation has designed and/or installed Austin sand filters to provide stormwater management for several large highway projects. Overall, the design provides dependable performance and can be designed so it does not pose an additional safety hazard for automotive traffic.

Effectiveness

The Austin sand filter design has demonstrated good total suspended solids (TSS) removals, typically providing 85 percent treatment. Performance for nutrients is less significant, and in fact the sand filter may be a source of nitrate (NO₃) since ammonia in stormwater will undergo nitrification in the aerobic filter. However, sand filters are reported to decrease the total nitrogen (TN) load by approximately 35 percent. Total phosphorus (TP) removals range up to 55 percent, and there is a wide variation in metal removal rates (ranging between 35 and 90 percent). Removal of oil and grease by sand filters has been reported to average between 55 and 84 percent (Horner and Horner, 1995). Reduction in fecal coliform bacteria ranges between 40 and 80 percent.

The bulk of Austin sand filter designs have been in a warmer climate (central Texas) and reported removal rates probably reflect this influence (see Table 13). The filter performance would probably decrease if exposed to prolonged cold periods, which freeze the filter media. However, in a recent application of a sand filter in Alexandria, Virginia, it was reported that the filter operated effectively immediately after an arctic freeze even with several inches of frozen runoff in the settling area (Bell et al., 1995).

With the integration of a sedimentation chamber, the design provides pretreatment for the filter. However, where high loadings of oil or grease are encountered, additional pretreatment measures, such as grassed swales or vegetated filter strips are advisable.

Siting and Design Considerations

Various design approaches can be taken in designing surface sand filters, including those developed in Austin. Design differences tend to be found in the size of the sedimentation area, the duration of sedimentation, and the loading rate of the filter media. For practicality, most designs limit the maximum water depth in the facility to less than 2.4 m (8 ft) and drain the system by gravity.

There are two basic designs for the Austin surface sand filter that manage the first 12.7 mm (0.5 in) of runoff, a partial sedimentation design and a full sedimentation design. The designs differ in terms of the volume of the sedimentation chamber and the size of the filter area. A partial sedimentation design creates a smaller footprint than a full sedimentation design but typically requires more maintenance. The partial sedimentation design is intended for areas that are relatively flat sloped and requires sufficient sedimentation area to store 20 percent of the water quality volume. The partial sedimentation design requires 16.7 m² (180 ft²) of filter area per impervious acre. The full sedimentation design provides sufficient sedimentation area to store the entire water quality volume (100 percent), a volume that is subsequently released to the filter bed over a 24-hour period. The full sedimentation design requires 9.3 m² (100 ft²) of area per impervious acre (assuming a permeability of the sand medium of 1 m/day [3.5 ft/day]). More extensive information regarding the design process used for the Austin sand filter should be acquired directly from the City of Austin's Environmental Criteria Manual (City of Austin, 1991).

There are also other approaches to surface sand filter designs that can be considered. One general rule of thumb is the required sedimentation area in square meters should be equal to 0.020 times the water quality volume in cubic meters (0.066 for area in square feet and volume in cubic feet) for drainage areas with an imperviousness of less than 75 percent (Claytor and Schueler, 1996). For areas with imperviousness greater than 75 percent, the sedimentation area commitment is 0.0024 times the water quality volume (0.0081 for area in square feet and volume in cubic feet). These recommendations recognize that ultra-urban runoff typically contains a high percentage of large-diameter sediment particles and therefore the settling area can be decreased (Shaver, 1994). When using this design approach, the recommended length-to-width ratio of the settling chamber is 2:1 or greater to limit short-circuiting, and the minimum recommended water depth in the settling chamber is 0.92 m (3 ft). This design approach also calls for the total storage volume in the sedimentation chamber and filter chamber to be equal to 75 percent of the water quality volume. At least half of the total storage volume should be located in the sedimentation chamber. The facility storage volume calculation should include void storage in the sand medium (typical porosity between 30 and 40 percent). In sizing the filter area it is recommended that a drawdown time of 40 hours be used and that the total depth of sand medium not exceed 0.61 m (2 ft). More information regarding this design approach can be found in Design of Stormwater Filtering Systems (Claytor and Schueler, 1996).

It should be noted that for any of the surface filter designs it is possible to substitute filter media other than sand. Refer to the Organic Media Filters Fact Sheet for additional information on organic media filters (peat/sand and compost media) and their advantages and disadvantages. Although over 500 Austin sand filters are currently operating, it is not known how long the basic design will last. Given the relatively low level technology typically employed, it seems reasonable to assume an effective life between 25 and 50 years with regular maintenance.

Maintenance Considerations

In general, the recommended frequency for performance monitoring is at least once per year. Each inspection should log information on the depth and location of any ponding, the depth of discoloration in the filter bed, and the depth of accumulated material over the sand media.

Most filters exhibit diminished capacity after a few years due to surface clogging by organic matter, fine silts, and hydrocarbons. Restoration of the original filtration capacity includes manual removal of any accumulated material and the first several inches of discolored sand. New sand is placed to reestablish the design grade of the filter medium. From a review of numerous references, it appears the material (sand/silt) accumulates in most sand filters at a rate between 13 to 25 mm/yr (0.5 to 1 in/yr). Maintenance can be reduced by employing surface sand filters only in drainage areas with 100 percent imperviousness. This significantly reduces the fine-grain material reaching the filter (silt and clay) which can clog the filter bed (Schueler, 1995). In areas with high trash loading, a wide-mesh geotextile screen can be placed over portions of the filter surface to simplify removal of the debris.

Regarding specific maintenance issues for the Austin sand filter design, the partial sedimentation design requires more frequent maintenance of the filter bed because there is less settling of solids in the sedimentation chamber. This tends to lead to greater sediment loads entering the filter bed than is experienced for full sedimentation designs (Young et al., 1996). Greater sediment loads translate into higher maintenance costs because more frequent replacement of the sand media will be required.

Cost Considerations

The surface sand filter design is a moderately expensive BMP to employ (Claytor and Schueler, 1996). However, the cost of installation is strongly correlated with the nature of the construction employed. If the filter is installed within an ultra-urban setting, it is likely that relatively expensive concrete walls will be used to create the various chambers. This type of installation will be significantly more expensive than an earthen-walled design, where relatively inexpensive excavation and compaction construction techniques lower the installation cost. However, earthen-wall designs require a greater land area commitment, which can offset the reduction in construction costs.

The construction cost of surface sand filters is also related to economies of scale-the cost per impervious hectare or acre served decreases with an increase in the service area. In 1994, the construction costs for Austin sand filters were $39,500 per impervious hectare (or $16,000 per impervious acre) for facilities serving less than two acres and $8,400 per impervious hectare (or $3,400 per impervious acre) for facilities serving greater than five acres (Schueler, 1994). These construction cost estimates exclude real estate, design, and contingency costs. (Note that these unit cost values should be used for conceptual cost estimating purposes only.)

References

Bell, W., L. Stokes, L.J. Gavan, and T.N. Nguyen. 1995. Assessment of the Pollutant Removal Efficiencies of Delaware Sand Filter BMPs. City of Alexandria, Department of Transportation and Environmental Services, Alexandria, VA.

City of Austin. 1991. Water Quality Management: Environmental Criteria Manual. Environmental Resources Management Division, Environmental and Conservation Services Department, City of Austin, Austin, TX.

City of Austin. 1990. Removal Efficiencies of Stormwater Control Structures. Environmental Resources Management Division, Environmental and Conservation Services Department, City of Austin, Austin, TX.

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

Horner, R.R., and C.R. Horner. 1995. Design, Construction, and Evaluation of a Sand Filter Stormwater Treatment System. Part II. Performance Monitoring. Report to Alaska Marine Lines, Seattle, WA.

Schueler, T.R. 1994. Developments in Sand Filter Technology to Improve Stormwater Runoff Quality. Watershed Protection Techniques 1(2):47-54.

Schueler, T.R. 1995. Performance of Delaware Sand Filter Assessed. Watershed Protection Techniques 2(1)291-293.

Shaver, E. 1994. Sand Filter Design for Water Quality Treatment. Proceedings of the Urban Runoff and Receiving Systems Conference, ed. E. Herricks. ASCE, New York, NY.

Welborn, C., and J. Veenhuis. 1987. Effects of Runoff Control on the Quality and Quantity of Urban Runoff in Two Locations in Austin, TX. USGS Water Resources Investigations Report. 87-4004.

Young, G.K., S. Stein, P. Cole, T. Kammer, F. Graziano, and F. Bank. 1996. Evaluation and Management of Highway Runoff Water Quality. FHWA-PD-96-032. Federal Highway Administration, Office of Environment and Planning.