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

Monitoring Case Study-Delaware Sand Filter BMPs At Airpark, Alexandria, Virginia

This case study is based on an evaluation of Delaware sand filter BMPs by Bell et al. (1995).

Two large Delaware sand filter (DSF) BMPs were constructed to treat the runoff from AirPark, a 0.7 ha (1.7 ac) commercial parking lot near National Airport in Alexandria, Virginia. Construction was completed in the fall of 1992 with funding from a grant by the Virginia Chesapeake Bay Local Assistance Department. The performance of these BMPs were monitored for six months to establish the actual pollutant removal efficiency of a DSF treating northern Virginia's acid rainfall with locally available sands. Confirmation that efficiencies in the range experienced by other jurisdictions (namely, the city of Austin, Texas) can be attained in northern Virginia would allow economic use of a greater amount of property and substantially reduce the cost of compliance by developers with the Chesapeake Bay Preservation Act.

Study Objectives

• Define the pollutant removal efficiency of the DSF BMP in the northern Virginia area.

• Due to the variability found in the coefficients of permeability (k) for sand filters, determine the actual k value for the filter being monitored.

• Use data on iron (Fe) in input and output samples to determine the aerobic condition of DSFs for a specific storm.

• Investigate whether considerable buffering might be occurring on the asphalt surface, as evidenced by previous pH testing in the city, by collecting samples of direct rainfall from some storms. Direct rainfall samples were also used to perform limited testing on how much of the pollutant loading was occurring from wet weather atmospheric deposition.

• Determine whether adsorption to sand particles is a significant component of phosphorus removal due to substantial iron and aluminum content in the filter sand by conducting Langmuir isotherm tests.

• Due to anaerobic activity within the AirPark South Filter, development of a compound stormwater filter that enhances total nitrogen (TN) removal efficiency results was proposed.

Design and Operation

The South Filter was one of the first large stormwater intermittent sand filter BMPs constructed in northern Virginia, one of two filters treating stormwater runoff from a 0.8 ha (1.95 ac) commercial parking lot located just south of National Airport. Impervious surfaces cover 95 percent of the site. Following the original DSF design with steel grates and covers, the cost of the filters exceeded $100,000 due to the high cost of the grates and covers. By modifying the original design and using a slotted curb design, the actual cost of the two filters was just over $40,000.

The 0.8 ha (1.95 ac) paved lot is graded so that runoff sheet flow enters the West Filter; both sheet flow and flow along the curb line enter the South Filter (Figure 56). Reversed sloped curbs are provided at the lower ends of each filter to act as flow-splitters; once the storage capacity of the filter shell is exceeded, runoff backs up on the pavement until it flows around the nose of the reverse sloped curb and on into a curb inlet. The wedge of water stored on the pavement is part of the storage capacity of the overall BMP.

Figure 56. General layout of AirPark filters (adapted from Bell et al., 1995)

After construction, both filters were found to take an inordinate amount of time to drain down. Two flat plastic drain tiles were installed along the entire length of the South Filter to decrease flow-through time and decrease the time that vehicles parked along the filter were inaccessible due to pooling of water to be treated. The outflow drain of the South Filter was also found to be approximately 5 cm (2 in) above the invert of the filter box rather than at the invert. With a bottom slope of 0.5 percent, approximately 10 m (33 ft) of the bottom filter sand remains continually saturated, creating potential anaerobic conditions. Other design parameters are provided in Table 30; where design values differ from the actual system, both values are provided.

Monitoring

Monitoring problems with the West Filter required the team to shift their focus to the South Filter, where input samples could be collected from a sloping curb line uphill of the filter, and the underdrain system provided a flow rate that was much easier to monitor. Severe weather between December 1993 and April 1994, causing freezing problems in the sedimentation pools in the filters and freeze-up of the automatic monitoring equipment delayed the start of the monitoring study. Rainfall totaling 105 mm (4.13 in) fell at the airport (just across U.S. Route 1 from AirPark) immediately preceding the first sample taken for this study, so the filters had been saturated for almost two weeks immediately prior to the beginning of the monitoring effort. Active monitoring was resumed on April 4, 1994, and continued for 20 storm events through September 23, 1994.

Input and output samples for the AirPark South Filter were collected utilizing a purpose-built monitoring manhole with Palmer-Bolus flume installed in the outflow pipe for the output samples; input samples were collected from the concrete gutter that conveys the runoff from the upper part of the parking lot to the filter. Both samplers were activated by transducer-type flowmeters. Rainfall data were collected directly at the AirPark site by an American Sigma rainfall gauge. Composite sampling was utilized to obtain a relatively unbiased approximation of the mass loads of pollutants processed by the filter by multiplying the event mean concentration for each storm by the volume of runoff treated; total sampling time was adjusted to approximate the time during which the "first-flush" of the filter was processed. The storm outflows measured by the flow meter in the outflow monitoring unit were originally intended to be used to compute mass balance efficiencies of the filter. However, volumes recorded by the meter were far short of the estimates of volume treated obtained from hydrological calculations. The outflow drain was found to be installed 5 cm (2 in) above the bottom of the filter box, and the outflow pipe directly above the Palmer-Bolus flume was found to have a slope of 3.4 percent (rather than the specified 0.76 percent), which fell outside the requirements for accurate flow measurement (maximum slope 2.2 percent) for the flume and flow meter computer program. The project team decided to use the calculated treatment volumes for calculating weighted mean concentrations and mass balance removal efficiencies. Table 31 presents the results of the monitoring study.

Monitoring results for the fourteenth storm on the South Filter led the project team to reassess the work plan due to radically different outcomes than had been previously experienced. Phosphorous and orthophosphorous removal fell to negative values, while the rates of nitrogen components fluctuated widely. It was suspected that the anaerobic zone had suddenly expanded to encompass the majority of the filter; precipitation records at National Airport showed rainfall during 12 of the 18 days immediately preceding the incident. Monitoring of the South Filter was continued until all results indicated the filter had returned to a predominantly aerobic condition.

Conclusions

• The sedimentation pools in DSFs are prone to freezing during periods of low temperatures. The sand filters, however, continue to function when underdrain pipes are provided beneath the filter media. DSFs are susceptible to anaerobic conditions unless positive drainage features are provided. Anaerobic environments have a negative impact on total phosphorous removal but a positive effect on total nitrogen removal.

• The major source of pollutants in runoff at AirPark appear to be atmospheric deposition. Most of the constituents measured in the input runoff fell within the ranges of the Nationwide Urban Runoff Program (NURP) data. No detectable readings of total petroleum hydrocarbons were found.

• The AirPark South Filter was found to have a TP removal efficiency of 72.3 percent when in a predominantly aerobic state, based on mass balance calculations. Phosphorous removal efficiencies were also found to increase with higher input concentrations in the runoff being treated. TN removal efficiency based on mass balance calculations was 47.2 percent. TSS also varied with input TSS concentrations, with removal efficiencies exceeding 80 percent.

• Acid rain runoff from asphaltic concrete parking lots will likely be buffered to a neutral state before reaching stormwater BMPs.

• Coefficient of permeability for the sand filter section of the AirPark South Filter was found to be 2.6 m/day (8.5 ft/day) median, rather than the 0.6 m/day (2 ft/day) used in design assumptions. However, further study is needed before a decision is made to change this design assumption.

• Isotherm results of the filter media showed that monolayer adsorption of phosphorous to the sand particles was not a significant factor in the overall removal results.

• Placing a 33 cm (13 in) flooded gravel (anaerobic) filter beneath the sand filter may enhance nitrogen removal if sufficient organic carbon is present.

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.