Stormwater | Constructing an Infiltration Trench Retrofit BMP

A demonstration site shows promise for BMPs in space-limited urban settings.

By Clay Emerson and Robert Traver

In the spring of 2004, an infiltration trench was built on the grounds of Villanova University near Philadelphia, PA. Funded by the Pennsylvania Department of Environmental Protection’s Growing Greener Program and the EPA’s Section 319 Nonpoint Source Pollution Management Program, the trench is the most recent addition to Villanova’s Stormwater BMP Research and Demonstration Park. Other best management practices (BMPs) include a stormwater wetland, a bioinfiltration traffic island, and a porous concrete site (see the article “Lessons in Porous Concrete” in this issue). All of the BMPs, including the infiltration trench, are retrofits. Each site is instrumented and serves as a research tool. The BMPs are currently under study through the EPA’s Section 319 National Monitoring Program.

Figure 1. Site before construction showing grading problems and consequent erosion

Project Background
The construction of the infiltration trench began in the spring of 2004, and monitoring was initiated in July. The BMP was constructed in a small area between an academic building and a parking garage. The area was considered unsightly by the university because of poor grading (Figure 1).

The design for the infiltration trench BMP included three main components. First, the site’s basic stormwater functions were to collect and infiltrate runoff from the upper deck of the adjacent bi-level parking garage. Second, the site was designed to provide sufficient research and demonstration opportunities to study the BMP’s hydrologic and water-quality performance. Third, the infiltration trench was designed to improve the site’s aesthetic appeal and function for the university.

Initial Site Investigations
Originally the site was chosen because of its proximity to the parking garage. The garage is used only by university staff and does not receive any large truck or delivery traffic, so there is little chance of a major hazardous spill. More investigation was required, first to ensure that the general area was appropriate for such a BMP and later to determine the specific location and orientation of the trench. University plans were collected in an effort to determine the various existing utilities and infrastructure in the area. This was a particularly arduous task, because although plans existed for both the garage and the academic building, there were no plans specific to the common area between them. Onsite investigation proved to be the only reliable method. The grassed area contained three large electrical conduits, which were encased in concrete; one single telephone line; two stormwater conduits; and an existing stormwater inlet—a surprisingly dense network of utilities for a grassed area only 45 feet wide. Ultimately it was the locations of these utilities that determined the exact size, location, and orientation of the trench.

During the early stages of the design process, a site investigation and feasibility study was performed. The results of this process showed that the site was suitable for the installation of an infiltration trench. First a test pit was dug to determine the depth to bedrock or water table and to classify the local soil properties and their suitability for stormwater infiltration. The depth to bedrock in the test pit was approximately 6 feet. Several authors recommend that 4 feet of soil are desirable between the bottom of an infiltration BMP and bedrock. Therefore, the 6-foot depth to bedrock in the test pit was a concern. Based on the location and elevation of a nearby first-order stream, the depth to groundwater was estimated to be about 15 feet. The test pit showed that there were 18 inches of disturbed topsoil followed by an equally thick layer of undisturbed and heavily weathered schist. The bottom 3 feet of the soil profile were dominated by a light tan sand layer that appeared to have excellent potential for infiltration. Soil samples were also taken from this layer and later analyzed in a laboratory setting. The particle size distribution of the soil sample classified the material as a loamy sand according to the Soil Conservation Service Soil Texture Triangle. An undisturbed 2-foot-deep step was left in the sand layer in which to perform some basic infiltration tests.

A constant-head infiltrometer was used in the test pit to estimate the soil’s infiltration capacity. The apparatus consists of a 6-inch-diameter metal ring, which is hammered into the soil to a depth of 3 inches. A long, graduated water supply tube stands on top of the ring and maintains a constant 6 inches of head on the soil surface. The flow rate is then calculated directly using the graduations and a stopwatch. The percolation test showed that the soil was able to absorb water a rate of 8.1 inches per hour. The three-dimensional flow characteristics associated with this type of infiltration test seem to be more appropriate for the infiltration trench than they would be for some of the other infiltration BMPs at Villanova. The trench is relatively small and deep, and the near-vertical walls of the trench likely are responsible for a large portion of the infiltration. Other infiltration BMPs that have a larger and flatter soil-water interface seem to exhibit infiltration rates closer to the soil’s hydraulic conductivity; this is often much lower than a standard “percolation test” would suggest.

Figure 2. Excavated trench showing monitoring wells and geotextile fabric liner

After the initial site investigations, the limiting factor appeared to be the somewhat shallow depth to bedrock observed in the test pit. However, due to the unexpected discovery of some of the utilities, it was decided that the final location for the trench would be approximately 10 feet away from the test pit. This new location was directly downhill along what was likely the original slope of the site. Therefore it was hypothesized that the depth to the fractured bedrock would be significantly larger. The final depth of the completed trench is 6 feet. No bedrock was encountered by hand augering to an additional 4-foot depth within the base of the trench (10-foot total depth).

Although not directly part of the infiltration trench project and funding, a retaining wall was first constructed to alleviate the erosion of the steep slope adjacent to the future location of the trench. The construction of the wall was done prior to the excavation of the infiltration trench to prevent potential compaction and the migration of sediment into the newly constructed BMP.

Design and Construction
As previously stated, the first design objective of the infiltration trench was its stormwater function. However, because of the project’s research and demonstration goals, there were many measures taken that were specific to the site’s monitoring and demonstration purposes. These features include the installation of two monitoring wells in the trench (Figure 2). Both wells contain a pair of groundwater samplers, one at 2 feet and one at 4 feet beneath the bottom of the trench in the undisturbed subsoil. One well houses a pressure transducer, which enables the depth in the bed to be monitored. Another monitoring-specific feature is the bench located against the wall of the parking garage as shown in Figure 3. This bench houses a small grit and trash separator, a baffle, and a V-notch weir. The weir, in conjunction with a second pressure transducer, allows for continuous inflow monitoring. The top of the bench is locked and mounted on hinges for maintenance and demonstration purposes. A tipping bucket rain gauge was also installed on the upper level of the parking garage. All the instrumentation is wired to a datalogger located in a secure area of the garage. Finally, information signage has been installed at the site.

Figure 3. Completed infiltration trench with monitoring/pretreatment bench and picnic table.

Figure 4. Original drainage configuration consisting of PVC drains and small concrete channels conveying runoff onto street.

The third objective of the design was to improve the site’s aesthetic appearance and functionality as a campus common area. Therefore, the infiltration trench was overlain with EP Henry ECO Pavers, which provide a strong, durable, and porous surface. Picnic tables were placed on the patio created by the trench. The bench used to house the grit and trash separator and the weir also provide additional seating for the students and faculty that frequent the newly renovated common area.

Like the other BMPs in the Stormwater BMP Research and Demonstration Park, the infiltration trench was a retrofit. The garage’s existing stormwater system consisted of PVC-piped gutters that were directly connected to the street by small concrete channels as shown in Figure 4. Street inlets in the area along with the overflow of the infiltration trench drain to the stormwater wetland just downstream of the site. The original PVC gutters were disconnected and rerouted to the infiltration trench. There a crushed stone bed provides temporary storage while the runoff is allowed to infiltrate the surrounding soil. The rerouted gutters capture roughly a third of the parking garage runoff, resulting in a 100% impervious drainage area of approximately 17,000 square feet.

Figure 5. Porous paver surface of infiltration trench

The trench is approximately 13 feet long, 10 feet wide at the top, and 6 feet deep. The trench is filled with large clean-washed stone aggregate, approximately 3 to 6 inches in diameter. The crushed stone bed provides approximately 40% void space, which results in a total capacity of approximately 200 cubic feet of storage in the trench, or about 0.15 inch over the drainage area. This relatively small capture depth was acceptable because the contributing drainage area can be easily decreased if it is determined to be too large by simply reconnecting the original PVC gutters at the desired downspouts. To date, the drainage area has not been decreased, and overflow generally does not occur until approximately 0.5 inch of rainfall is observed. This is because of the rapid infiltration rate observed in the observation wells at the site. The stone bed is completely wrapped in a Class C geotextile fabric. The fabric prevents the surrounding soil from migrating into the stone bed and decreasing its effective storage. At the top of the bed above the geotextile is a 2-inch layer of choker stone overlain with ECO Pavers. Decorative 6- by 6-inch timbers were used to outline the ECO Pavers. The pavers have nubs that evenly space them and provide 17.4% open space according to the manufacturer’s specification sheet. The ECO Pavers were donated to the project by EP Henry. The open space between the pavers was filled with small choker stone to complete the installation. See Figure 5.

A perforated four-way PVC fitting at the top of the bed just beneath the pavers connects to a 6-inch PVC overflow pipe, which carries flows in excess of the bed’s storage capacity to a nearby existing storm drain. However, the flow in the 6-inch overflow pipe is somewhat impaired by the crushed stone in the bed. Therefore, in periods of intense rainfall when the storage bed is full and the capacity of the overflow pipe is exceeded, the ECO Pavers act as an additional overflow. Excess runoff is allowed to flow up from between the pavers and over 2 feet of grass and into the storm drain. This design feature has worked perfectly and results in evenly distributed sheet flow at fairly low velocity that has not caused any erosion of the small grass strip between the trench and the inlet.

Conclusion and Recommendations
One lesson learned during the construction arose when the crushed stone arrived at the site for installation. The stone was specified to be “clean washed.” However, the stone clearly had not been washed and contained fine sediment that could decrease the life expectancy of the BMP. Fortunately, this was noticed before placement of the stone. The dump truck was then driven to a contained area where the bed was inclined and the stone was washed by a worker using a hose. Although this process was time consuming, it was absolutely imperative that the stone not contain significant amounts of fine-particle-size sediment. This experience illustrates the importance of both clear specifications and, most importantly, onsite supervision at similar BMP construction sites. A comprehensive site investigation is very important, especially if the project is a retrofit and there is a possibility of encountering existing utilities. In all, very few problems were encountered during the construction of the infiltration trench. This demonstrates the importance of proper siting, a thorough site investigation, and sound design.

The infiltration trench is reducing erosive storm flows and nonpoint-source pollution to the headwaters of Mill Creek. Additionally, the BMP is replenishing local groundwater supplies and helping maintain baseflow in the local first-order tributaries of Mill Creek. The long-term monitoring being conducted at the site is providing detailed insight into the performance of the BMP. This experience and performance data will be used to better understand and implement similar stormwater BMPs in the future. The extreme design limitations imposed by the relatively small area with numerous utilities show that the infiltration trench BMP is capable of being successfully retrofitted into some of the tightest existing urban conditions. The trench, still only a year old, has been the subject of numerous tours attended by students, professors, regulators, watershed groups, and design engineers.

Acknowledgements
Funding for the project was provided through the Pennsylvania Department of Environmental Protection’s Growing Greener Program with continued support from the EPA’s Section 319 Nonpoint Source Pollution Management Program. This support does not imply the endorsement of this project by the EPA or the Pennsylvania Department of Environmental Protection. The design and construction of the infiltration trench was overseen by the Villanova Facilities Management Department’s Executive Director Robert Morro. Special thanks go to Project Manager Nick Grosso at Facilities Management, and Erika Dean for her continued hard work and dedication to the project. The authors also wish to thank EP Henry for donating the porous pavers.

Robert Traver is a professor in the Department of Civil and Environmental Engineering at Villanova University and directs the Villanova Urban Stormwater Partnership. Clay Emerson is a VUSP research associate.

SW July/August 2005

Home + About + Subscribe + News + Calendar + Glossary
Talk + Images + Advertise + Contact Us + Search + Register + Services