Is it possible to have a stormwater best management practice (BMP) that reduces impervious areas, recharges groundwater, improves water quality, eliminates the need for detention basins, and provides a useful purpose besides stormwater management? This seems like a lot to expect from any stormwater measure, but porous asphalt pavement on top of recharge beds has a proven track record.

 
Figure 1. Rainfall runs off traditional impervious             Figure 2. Cross-section through porous asphalt showing subsurface asphalt (center drive) but drains through porous              infiltration bed beneath
asphalt parking spaces.

First developed in the 1970s at the Franklin Institute in Philadelphia, PA, porous asphalt pavement consists of standard bituminous asphalt in which the aggregate fines (particles smaller than 600 µm, or the No. 30 sieve) have been screened and reduced, allowing water to pass through the asphalt (Figure 1). Underneath the pavement is placed a bed of uniformly graded and clean-washed aggregate with a void space of 40%. Stormwater drains through the asphalt, is held in the stone bed, and infiltrates slowly into the underlying soil mantle. A layer of geotextile filter fabric separates the stone bed from the underlying soil, preventing the movement of fines into the bed (Figure 2).

Porous pavement is especially well suited for parking-lot areas. Several dozen large, successful porous pavement installations, including some that are now 20 years old, have been developed by Cahill Associates (CA) of West Chester, PA, mainly in Mid-Atlantic states. These systems continue to work quite well as both parking lots and stormwater management systems. In fact, many of these systems have outperformed their conventionally paved counterparts in terms of both parking-lot durability and stormwater management.

Installations Old and New

One of the first large-scale porous pavement/recharge bed systems that CA designed is in a corporate office park in the suburbs of Philadelphia (East Whiteland Township, Chester County). This particular installation of about 600 parking spaces posed a challenge because of both the sloping topography and the underlying carbonate geology that was prone to sinkhole formation. The site also is immediately adjacent to Valley Creek, designated by Pennsylvania as an Exceptional Value stream where avoiding nonpoint-source pollution is of critical importance. Constructed in 1983 as part of the Shared Medical Systems (now Seimens) world headquarters, the system consists of a series of porous pavement/recharge bed parking bays terraced down the hillside connected by conventionally paved impervious roadways. Both the top and bottom of the beds are level, as shown in Figure 3, hillside notwithstanding. After 20 years, the system continues to function well and has not been repaved. Although the area is naturally prone to sinkholes, far fewer sinkholes have occurred in the porous asphalt areas than in the conventional asphalt areas, which the site manager attributes to the broad and even distribution of stormwater over the large areas under the porous pavement parking bays.


Figure 3. Porous pavement parking bays are benched down a hillside.

Other early 1980s sites, such as the SmithKline Beecham (now Quest) Laboratory in Montgomery County, PA, and the Chester County Work Release Center in Chester County, PA, also used the system of terracing the porous paved recharge beds down the hillside to overcome the issues of slope. At the DuPont Barley Mills Office complex in Delaware, the porous pavement was installed specifically to avoid the construction of a detention basin, which would have destroyed the last wooded portion of the site. More recently (1997), the porous parking lots at the Penn State Berks Campus were constructed to avoid destroying a wooded campus hillside. The Berks lots, also on carbonate bedrock, replaced an existing detention basin and have not experienced the sinkhole problems that another campus detention basin has suffered.

How It Works

The porous asphalt mix has a lower concentration of fines than traditional asphalt, as shown in Table 1, accomplished by straightforward screening. In all other manufacturing aspects, porous asphalt is the same as conventional asphalt and can be mixed at a standard asphalt batch plant. With fewer fines, the asphalt is porous and allows water to drain though the material through virtually imperceptible openings (to the untrained eye, porous pavement properly prepared is difficult to distinguish from conventional pavement). There are several variations of the mix, including gradations developed by various state transportation departments seeking a pavement that also can be used to reduce noise and skidding. For the purposes of stormwater management, we have found the best performance from the mix indicated in Table 1.

Table 1. Standard Porous Asphalt Mixes

The underlying stone recharge bed consists of a uniformly graded (i.e., screened) 1.5- to 2.5-in. clean-washed stone mix, such as an AASHTO No. 2. Depending on local aggregate availability, both larger and smaller size stones have been used. The important requirement is that the stone be uniformly graded (to maximize void space) and clean washed. The void space between the stones provides the critical storage volume for the stormwater. Stones that are dusty or dirty might clog the infiltration bed and must be avoided. The stone bed is usually between 18 and 36 in. deep, depending on stormwater storage requirements, frost depth considerations, and site grading. This depth provides a significant structural base for the pavement. As a result, porous asphalt exhibits very few of the cracking and pothole formation problems encountered in conventional pavement.

The bottom of the recharge bed is excavated to a level surface and is not compacted. This allows water to distribute and infiltrate evenly over the entire bed bottom area. Compaction of the soils will prevent infiltration, so it is important that care be taken during excavation to prevent this. The bottom of the bed cannot be placed on fill material unless that fill material is stone. A layer of nonwoven geotextile at the bottom of the bed allows the water to drain into the soil while preventing the soil particles from moving into the stone bed.

Figure 4. Roof leaders can be connected directly to the subsurface infiltration bed (corresponds to legend below).

A - Precipitation is carried from roof by roof drains to storage beds.
B - Stormwater runoff from impervious areas and lawn areas is carried to storage beds.
C - Precipitation that falls on pervious paving enters storage bed directly.
D - Stone beds with 40% void space store stormwater. Perforated pipes distribute stormwater from impervious surfaces evenly throughout the beds.
E - Stormwater exfiltrates from storage beds into soil and recharges the groundwater.

Very often the underlying stone bed will provide stormwater management for adjacent impervious areas such as roofs and roads. To achieve this, we convey the stormwater directly into the stone bed and then use perforated pipes in the stone bed to distribute the water evenly (Figures 4 and 5).
             
          Figure 5. Perforated pipes distribute stormwater                          Figure 6. Unpaved stone edge in the event 
                  evenly throughout the infiltration bed.                                               that the surface is repaved.

Design Considerations

In the late 1970s and early 1980s, as we designed our first systems, we were uncertain how well the porous asphalt would hold up over time and use. In these first systems, such as the one at the Morris Arboretum in Philadelphia (1982), we designed the parking spaces with porous pavement but constructed the aisles and connector roadways with conventional asphalt. However, we extended the stone stormwater storage/infiltration bed under the entire parking area, including the areas with impervious paving.

Over time, we have found that the porous asphalt material has held up as well as, or better than, the conventional asphalt, largely because of the solid sub-base provided by the stone storage/infiltration bed. In subsequent designs we have paved the entire surface in the porous asphalt. We have found that sufficient asphalt content is essential to pavement durability (5.75‚6.0% bituminous asphalt by weight). In sites where a lower asphalt content was used, some surface scuffing can be observed on the pavement surface. In different situations, we have tried various commercial additives intended to improve strength or performance in cold weather, but in general we have avoided any proprietary mixes or additives. Porous pavement is not a product but a design technique.

We have also taken the "belt and suspenders" approach to all of our systems. If the pavement were to be paved over, forgotten, or clogged, stormwater still must reach the stone bed below the pavement. Often we have used an unpaved stone edge, as shown in Figure 6, for this purpose. We have also used catch basins that discharge to perforated pipes in the bed.

Additionally, in case the bed bottom clogs (which has not happened yet), we have always designed the underlying bed systems with a "positive overflow." During a storm event, as the water in the underlying stone bed rises, it must never be allowed to saturate the pavement. We have used a catch basin with a higher outlet than inlet to provide positive release. In this way the bed also serves as an "underground detention basin," eliminating the need for a separate basin.

As a design rule, if the stone bed can provide a storage volume equal to the volume of increased runoff during a local two-year storm event (that is, the difference in the volume of runoff before development and after development), this will provide sufficient storage to mitigate the peak rate of runoff during larger storm events (25- to 100-year). Most local ordinances are concerned with rate of runoff. Essentially the bed acts as an underground detention basin in extreme storm events, albeit one that also reduces volume. A storm can be "routed" through the bed using the same calculation methods employed to route detention basins to confirm peak-rate mitigation.

As a final design consideration, infiltration systems also work best when the water is allowed to infiltrate over a large area. As a rule of thumb, we usually design to a ratio of 5:1 impervious area to infiltration area. That is, the runoff from 5 ac. of impervious area would require a 1-ac. infiltration bed. Because parking tends to consume so much of our landscape relative to other impervious surfaces, meeting this ratio is rarely a problem.

Soil and Subsurface Conditions

Suitable soil conditions are required for infiltration. The designer must evaluate a number of factors, including soil type, infiltration rate, depth to bedrock, and depth to water table. Some of the guidelines we have used in design are shown in Table 2. The most important factor is that the location of the porous pavement infiltration system be considered early in the design process. Traditionally engineers have designed stormwater systems that collect and convey runoff to the lowest point. By the time you have done this, you are likely to be at the wettest location on the site, next to the stream or wetlands or in poor soils. Infiltration systems perform best on upland soils. Some of our more recent designs integrate a mixture of large and small infiltration systems throughout the site, including porous pavement, to avoid conveying stormwater any distance.

Table 2. Design Guidelines for Subsurface Infiltration
- Avoid piping water long distances. Look for infiltration opportunities within the immediate project area.
- Consider past uses of site and appropriateness of infiltration design and porous pavement.
- Consider the source of runoff. Incorporate sediment reduction techniques as appropriate.
- Perform site tests to determine depth to seasonal high water table, depth to bedrock, and soil conditions, including infiltration capabilities. Design accordingly. Maintain 3 ft. above high water table and 2 ft. above bedrock.
- Avoid excessive earthwork (cut and fill). Design with the contours of the site. Maintain a sufficient soil buffer above bedrock.
- Do not infiltrate on compacted fill.
- Avoid compacting soils during construction.
- Maintain erosion and sediment control measures until site is stabilized. Sedimentation during construction can cause the failure of infiltration systems.
- Spread the infiltration over the largest area feasible. Avoid concentrating too much runoff in one area. A good rule of thumb is 5:1 impervious area to infiltration area (i.e., 5 ac. of impervious area to 1 ac. of infiltration area). A smaller ratio is appropriate in carbonate bedrock areas.
- The bottom of the infiltration area should be level to allow even distribution.
- The surface of the porous pavement should not exceed 5%. Use conventional pavement in steep areas that receive vehicular traffic.
- Provide thorough construction oversight.

Before any infiltration system is designed, soil investigation must be done. This consists of two steps. First, a simple "deep hole" 6-8 ft. in depth is excavated with a backhoe and the soil conditions are observed. While some designers prefer an auger, we believe that there is no substitute for physically observing and describing the soil horizons. Next, infiltration measurements are performed at the approximate bed bottom location. We have used both simple percolation tests, which are not very scientific, as well as infiltrometer readings. We do not consider infiltration rates between 0.1 and 0.5 in./hr. too slow; rather, this means that infiltration will occur slowly over a two- to three-day period, which is ideal for water-quality improvement.

Underlying geology must also be considered in areas such as those underlain by carbonate formations. In that situation, more detailed site investigation may include borings and ground-penetrating radar. Contrary to popular belief, properly designed infiltration systems do not create sinkholes. A number of our systems, including older systems, are located in carbonate areas. In several situations we have successfully installed porous pavement infiltration systems adjacent to areas where detention basins created sinkholes.

When Infiltration Is Limited

Despite the need for infiltration, not all sites and soils are suitable. In those situations, we have designed porous pavement systems to reduce impervious surfaces or as part of a water-quality improvement program. The porous pavement parking lots recently constructed at the John Heinz National Wildlife Refuge near the Philadelphia Airport are located in a wet, low-lying site that has been subject to fill over the years. The soils are not well drained. In this situation, a trench was excavated to a lower gravel layer to facilitate infiltration, but the parking lots primarily serve to avoid the creation of new impervious surfaces at this valuable wildlife refuge.

At the Ford Motor Company Rouge Facility in Dearborn, MI, the use of porous pavement is an important part of Ford''s commitment to sustainability. The original manufacturing plant was constructed in a low-lying wet area and has been subject to a century of industrial use. In 1999, Ford constructed a porous parking lot designed to slowly drain to a series of planted wetland swales (Figures 7 and 8). The stormwater stored in the beds beneath the porous pavement supports the vegetated swales by discharging slowly to the planted areas. The system is specifically designed to improve water quality. Referred to as the "Mustang Lot" (because new Ford Mustangs are parked there after assembly), the lot has worked well, and current Ford plans include the construction of 62 ac. of porous pavement areas that will drain to constructed stormwater wetlands.
 
Figures 7 & 8. This lot at Ford Motor Company in Detroit is made of porous asphalt that drains stormwater to vegetated bioswales.

Water Quality

There have been limited sampling data on the porous pavement systems, although the available data indicate a very high removal rate for total suspended solids, metals, and oil and grease (Table 3). More recently, Brian Dempsey, Ph.D., and his research assistant, David Swisher, have conducted research at the Pennsylvania State University. Dempsey has been studying a porous pavement system constructed at the Centre County/Pennsylvania State Visitor Center in 1999, comparing the water quality in the infiltration beds to observed runoff from a nearby impervious parking lot. He has monitored the infiltration rates of this system and found that the system has maintained a consistent infiltration rate. During a 25-year precipitation event, there was no surface discharge from the stone beds.

Table 3 Pollutant Removal Efficiencies for Infiltration BMPs (with porous paving highlighted)
    

Cost

Porous pavement does not cost more than conventional pavement. On a yard-by-yard basis, the asphalt cost is approximately the same as the cost of conventional asphalt. The underlying stone bed is usually more expensive than a conventional compacted sub-base, but this cost difference is generally offset by the significant reduction in stormwater pipes and inlets. Additionally, because porous pavement is designed to "fit into" the topography of a site, there is generally less earthwork and no deep excavations. When the cost savings provided by eliminating the detention basin are considered, porous pavement is always an economically sound choice. On those jobs where unit costs have been compared, the porous pavement always has been the less expensive option. Current jobs are averaging between $2,000 and $2,500 per parking space for parking, aisles, and stormwater management.

Construction

Invariably, when an infiltration BMP fails, it is because of difficulties and mistakes in the design and construction process. This is true for porous pavement and all other infiltration BMPs. Carelessness in compacting the subgrade soils, poor erosion control, and poor-quality materials are all causes of failure. For that reason, we provide detailed specifications on site protection, soil protection, and system installation. On every project, we meet with the contractor before construction and discuss such things as the need to prevent heavy equipment from compacting soils, the need to prevent sediment-laden waters from washing onto the pavement, and the need for clean stone. We verbally review the installation process with the project foreman. During construction, we routinely stop by the site or provide construction advice. Successful installation of any infiltration BMP is a hands-on process that requires an active role for the designer. Although we have prevented failures with this approach, most of the problems we have seen at other infiltration BMPs are a result of construction problems. Often the failure does not lie with the contractor or with poor soils but instead is due to a lack of specific guidance for construction procedures.

Because construction sites are inherently messy places, we often find it best to install the porous pavement and other infiltration BMPs toward the end of the construction period. By doing this, there is less risk of creating problems. On many projects, we will excavate the stone bed area to within 6 in. of the final grade and use the empty bed area as a temporary sediment basin and stormwater structure. Care must be taken to prevent heavy equipment from compacting the soils, but sediment can accumulate. In the later stages of the project, the sediment is removed, the bed is excavated to final grade, and the porous pavement system is installed. This also avoids the need for a separate sediment basin during construction.

Maintenance

We recommend that all porous pavement surfaces be vacuum swept twice per year with an industrial vacuum sweeper. Unfortunately, like many stormwater maintenance requirements, this advice often is overlooked or forgotten. Fortunately, even without regular maintenance, the systems continue to function (we routinely send graduate students and recent hires out in hurricanes to confirm this).

When runoff is conveyed from adjoining areas or roof surfaces into the bed, we often use a drop inlet box or other structure to reduce the amount of detritus and sediment that is conveyed to the bed. This structure also requires regular removal of sediment and debris.

Deicing and Freezing Issues

One of the most common questions relates to concerns about freezing conditions. Freezing has not been an issue, even in very cold climates. We were quite surprised when the owners of early installations first told us that there was less need to snowplow on the porous pavement surfaces. The underlying stone bed tends to absorb and retain heat so that freezing rain and snow melt faster on the porous pavement. The water drains through the pavement and into the bed below with sufficient void space to prevent any heaving or damage, and the formation of "black ice" is rarely observed. The porous surfaces tend to provide better traction for both pedestrians and vehicles than does conventional pavement. Not a single system has suffered freezing problems.

Obviously the use of sand or gravel for deicing would be detrimental to the porous surface. However, salt may be used, and the surface may be plowed if needed. Most sites have found that light plowing eliminates the need for salt since the remaining snow quickly drains through the asphalt. This has the added benefit of reducing groundwater and soil contamination from deicing salts.

Where It Doesn''t Work

Because porous asphalt has reduced fines, it has less shear-strength capability and therefore is not recommended for such situations as airport taxiways or slopes greater than 6%. We have not used the material for roadways, although it has been applied more extensively in Europe. There are also locations where the threat of spills and groundwater contamination is quite real. In those situations (such as truck stops and heavy industrial areas), we have applied systems to treat for water quality (such as filters and wetlands) before any infiltration occurs. The ability to contain spills must also be considered and built into the system. Finally, we have avoided the use of porous asphalt in areas where the pavement is likely to be coated or paved over because of a lack of awareness, such as individual home driveways. In those situations, a system that is not altered easily by the property owner is more appropriate (i.e., an infiltration system under a conventional driveway). "Preventative design" is critical.

Variations on the Theme: Porous Walkways and Playgrounds, Porous Concrete

More recently, we have applied the asphalt to situations such as walkways and playgrounds, including paths at Swarthmore College in Philadelphia (Figure 9), and an urban playground at the Penn New School in Philadelphia (Figure 10). At Swarthmore College, the paths are not part of an infiltration bed but are merely intended to reduce impervious cover. The Penn New School project works to reduce the volume of stormwater discharging to the Philadelphia combined sewer overflow. Both of these applications are "retrofits" in urban areas that were previously paved. Howard Neukrug, head of the Office of Watersheds for the Philadelphia Water Department, describes the Penn New School playground as part of "a vision of moving up and down the sewer [water]shed with urban projects which use the rainwater as an asset to the community through sustainable, aesthetic improvements and environmental education. These efforts enhance water-quality and -quantity issues in our rivers and streams and lead toward the environmental outcome we are all striving forófishable, swimmable, drinkable, and accessible waters."
            
                      Figure 9. Swarthmore College uses                                      Figure 10. Porous asphalt playground
                           porous asphalt walking paths.                                           at Penn New School in Philadelphia.

Additionally, we have been applying the use of porous concrete for both sidewalks and parking areas. Similar to the asphalt, the concrete has performed well in cold-weather climates. Because asphalt is less expensive, it remains our first choice for parking lots, but the porous concrete provides a good alternative where asphalt is not appropriate. At the University of North Carolina in Chapel Hill, two large commuter parking lots have recently been installed using a combination of porous asphalt and porous pavement. The university was specifically interested in comparing the installation and performance of both materials in a southern climate but was also driven by a university commitment to manage stormwater for both volume and quality.

Summary

In many new development projects, two-thirds of the new impervious surfaces are related to the automobile. Lost recharge, depleted groundwater levels, low stream baseflows, eroded streambanks, and degraded water quality all are effects of this extensive paving program. Flood and drought are both worsened by a development program of "sealing the earth''s surface." We can put parking lots to work for better stormwater management. There is nothing very exciting about a parking lot, but a parking lot designed to maintain the hydrologic balance that existed before development is worth notice. We believe that porous pavement and other infiltration BMPs are critical to successful watershed programs. It is our hope that others will also find this technology worthwhile.