February 2, 2007

Influencing Factors and a Proposed Evaluation Methodology for Predicting Groundwater Contamination Potential From Stormwater Infiltration Activities

By Clark, Shirley E; Pitt, Robert


To offset the detrimental effects of urbanization on groundwater recharge, stormwater managers are focusing on infiltrating much of the runoff from a site that was generated because of development. For this to be effective, tools are required to predict the potential for contamination resulting from this infiltration for many site conditions, because infiltration should be stressed in areas where the least potential for causing groundwater contamination exists. Factors that influence contamination potential include the pollutant concentration in the runoff directed to the infiltration device and the ability of the underlying soil to remove the pollutant. The groundwater contamination potential of some pollutants, even those with high concentrations and moderate-to- high mobilities, can be reduced with proper pretreatment before infiltration. This paper presents a methodology that can be used to evaluate infiltration as an management option and introduces two different levels of models that could be used to evaluate contamination potential. Water Environ. Res., 79, 29 (2007).

KEYWORDS: groundwater contamination, stormwater, infiltration, urban runoff treatment, vadose zone.



Many studies worldwide have documented the detrimental effects of urbanization on urban water resources. The annual wet-weatherflow literature review published in Water Environment Research for the past several decades now includes several hundred references each year, with increasing numbers of citations each year. It has been well-established that, before urbanization, groundwater recharge resulted from the natural infiltration of precipitation through pervious surfaces, including grasslands and woods. With urbanization, the permeable soil surface area, through which infiltration and subsequent recharge could occur, has been reduced. This has resulted in much less groundwater recharge and greatly increased quantities of surface runoff. In addition, the waters available for recharge carry increased quantities of pollutants compared with natural conditions.

As an example, the first paragraph of Maryland's storm water management manual (Maryland Department of the Environment, 2000) emphasizes that the increased stormwater runoff associated with postdevelopment conditions should, where possible, not be allowed to enter urban receiving waters. Thus, the "excess" water must be removed by one of two primary mechanisms-infiltration to the soil profile and/or evaporation (or evapotranspiration in a device that incorporates plants to the treatment scheme). Currently, the emphasis is on infiltration, which has the additional benefit of partially restoring groundwater recharge.

In another example, in their draft Stormwater Management Manual, the Commonwealth of Pennsylvania listed accepted best management practices. Selection of a management practice is determined by the goals, which are listed as groundwater recharge, water quality, and/ or rate and volume control. Infiltration is the only option if the goal includes groundwater recharge. Infiltration also is the only practice that can meet all three goals.

Much of the information presented in this paper was developed during research performed as part of a multiyear research project sponsored by the U.S. Environmental Protection Agency (U.S. EPA) (Washington, D.C.) (Clark and Pitt, 1999; Pitt et al., 1994, 1995, 1996), including the development of a simple method for evaluating the potential of groundwater contamination resulting from infiltrating stormwater. Recent evaluations, using a computerbased vadose zone model, were performed as part of a Water Environment Research Foundation (Alexandria, Virginia) -sponsored project on developing guidance for stormwater managers considering infiltration as a management practice (Clark et al., 2006) and also are described in this paper.

Infiltration Evaluation Methodology and Model Results

The design of an appropriately functioning infiltration device, where prevention of groundwater contamination is a primary concern, requires the following three steps:

(1) Determining the concentrations and forms of the pollutants entering and leaving the device,

(2) Determining the characteristics of the soil that affect water quality, and

(3) Determining the pretreatment requirements to prevent groundwater contamination.

The focus of this paper is on how to incorporate locally derived data and the available body of research on stormwater quality and soils to an evaluation method that can be used to determine if stormwater infiltration is feasible at a site.

Step 1-Evaluate Pollutant Loadings and Chemical Forms for Pollutants of Concern. Urban surfaces are subject to the deposition of contaminants, which are then subject to washoff by rainfall or snowmelt. Typical contributors to pollutants in runoff include vehicular traffic, industry, lawn care, pets, eroded sediments, and vegetation litter (literature summarized annually in Water Environment Research). Part of addressing the appropriateness of infiltration is understanding the nature of the pollutants typically found in stormwater runoff. The major pollutant categories associated with urban stormwater runoff, and hence the ones to be assessed in initially characterizing runoff water slated for infiltration, include sediment, nutrients, toxic chemicals, microorganisms, and dissolved minerals.

An important consideration for fate and transport is whether or not the pollutant is associated with particulates or whether it is dissolved (ionic) or colloidal. Ionic/dissolved or colloidal pollutants are more likely to penetrate into the vadose zone and require a chemical interaction between the soil and the pollutant to prevent transport. Paniculate-associated pollutants are more likely to be trapped on the infiltration device's surface. For example, many metals are particulate-associated; thus, they are unlikely to penetrate deep into the soil below an infiltration structure. Table 1 shows the filterable fractions of several stormwater pollutants.

Fate and Transport of Stormwater Pollutants in Infiltration Devices. The groundwater contamination potential of urban runoff pollutants has been addressed by many researchers. This section reviews briefly the literature from Andleman et al. (1994) and Pitt et al. (1996).

Nutrients. Many researchers have concluded that the potential exists for nitrate contamination of groundwater because of stormwater runoff (several have documented actual contamination events) (Bannerman et al., 1993; German, 1989; Hampson, 1986; Schiffer, 1989) and have linked nutrient movement to soil characteristics (Butler, 1987; Crites, 1985; Gold and Groffman, 1993; Ragone, 1977; Robinson and Snyder, 1991; White and Dornbush, 1988; Wilde, 1994). However, nitrates are typically in relatively low concentrations in stormwater, reducing their contamination potential to groundwaters.

Pesticides. Groundwater contamination of pesticides from stormwater runoff has been documented by several researchers, including German (1989), Domagalski and Dubrovsky (1992), and Wilson et al. (1990). However, Gold and Groffman (1993) reported leaching losses from residential lawns to be low for dicamba and 2,4- dichlorophenoxy acetic acid (

Heavy repetitive use of mobile pesticides on irrigated and sandy soils likely contaminates groundwater. Fungicides and nematocides must be mobile to reach the target pest; hence, they generally have the highest contamination potential. Pesticide leaching depends on patterns of use, soil texture, total organic carbon content of the soil, pesticide persistence, and depth to the water table (Shirmohammadi and Knisel, 1989). Estimates of pesticide mobility can be made based on volatilization, sorption, and solubility of the compound, as shown in Table 2 (Armstrong and Llena, 1992). Jury et al. (1983) demonstrated that the potential for pesticide removal in the soil is related both to the pesticide's ability to sorb to the soil and the potential for biodegradation before reaching the groundwater.

Other Organic Compounds. Many researchers have documented groundwater contamination resulting from organics in runoff (German, 1989; Ku and Simmons, 1986; Wilde, 1994; Wilson et al., 1990). Many of these were associated with source areas, in which organic concentrations exceeded typical urban runoff concentrations. Infiltration is therefore inappropriate at critical source areas, especially without adequate runoff pretreatment. Groundwater contamination from organics, as from other pollutants, occurs readily in areas with pervious soils, such as sand and gravel, and where the water table is near the land surface (Troutman et al., 1984). Organic removal from the soil and recharge water can occur by one or more methods-volatilization, sorption, and/or degradation (Crites, 1985\). Mobility classes similar to those outlined for pesticides can be developed, because the removal mechanisms are similar to those of pesticides. These rankings are directly related to the compound's K^sub ∝^ value, where a high K^sub ∝^ indicates that the pollutant is able to sorb to the soil's organic matter (Jury et al., 1983).

Pathogens and Indicator Organisms. Microbial groundwater contamination from urban runoff, typically through the use of direct recharge devices, has been documented (see Clark et al., 2006, for a review). Removal of these organisms is dependent on the soil chemical properties that promote adsorption and retention.

Viral adsorption in soils is promoted by increasing cation concentration, decreasing pH, and decreasing soluble organics (U.S. EPA, 1992) and is controlled by both the efficiency of short-term virus retention and the long-term behavior of viruses in the soil (Crites, 1985; U.S. EPA, 1992). The downward movement and distribution of viruses are controlled by convection and sorption and hydraulic dispersion mechanisms. Because the movement of viruses through soil to groundwater occurs in the liquid phase and involves water movement and associated suspended virus particles, the distribution of viruses between the adsorbed and liquid phases determines the viral mass available for movement.

The major bacterial removal mechanisms in soil are straining at the soil surface and at intergrain contacts, sedimentation, and sorption by soil particles (Crites, 1985). Factors such as temperature, pH, metal concentration, nutrient availability, and others affect the ability of a bacterial colony to survive in the water or soil (Ku and Simmons, 1986). Bacteria survive longer in acid soils and when large amounts of organic matter are present. Bacteria and larger organisms in wastewater are typically removed during percolation through a short distance of soil (U.S. EPA, 1992). The concern for groundwater contamination is that viruses and bacteria have been shown to migrate in the soil profile. Infiltrating stormwater may collect previously deposited microorganisms and transport them to the groundwater. Bacterial transport depth appears to be related to total dose on the soil and fluid velocity through the soil (Camesano and Logan, 1998; Unice and Logan, 2000). Transport of Escherichia coli and enterococci through stormwater filters also has been documented (Clark, 2000). Microbially contaminated sediments often function as a reservoir in which microorganisms can persist (Jensen et al., 2002).

Metals. In general, studies of recharge basins receiving large stormwater metal loads show that most heavy metals are removed either in the basin sediment or in the uppermost layers of the vadose zone (Hampson, 1986; Ku and Simmons, 1986). Removal of metals by soil may be through one of several processes, including soil surface association, precipitation, occlusion with other precipitates, solid-state diffusion into soil minerals, and biologic system or residue incorporation (Crites, 1985). Most of these removal processes are pH-dependent, as is the solubility of most metals. In general, a metal's solubility increases as the solution's pH decreases (Wilde, 1994).

Most of the heavy metals in stormwater are associated with particulates and can be readily strained out through filtration as the water infiltrates to the soil (Pitt et al., 1995). Therefore, these direct physical removal mechanisms are likely more important than chemical removal mechanisms for most heavy metals. Similar to that generated for the pesticides, mobility class rankings have been generated for the filtered forms of the metals, as shown in Table 3 (Armstrong and Llena, 1992).

Salts. Soil is not very effective at removing most salts. Once contamination with salts begins, the movement of salts into the groundwater can be rapid. The salt concentration may not lessen until the source of the salts is removed. One example occurred in Maryland, where the nearby use of deicing salts and their subsequent infiltration to the groundwater shifted the major-ion chemistry of the groundwater to a chloride-dominated solution. Although deicing occurred only three to eight times a year, increasing chloride concentrations were noted in the groundwater throughout the 3-year study, indicating that groundwater systems are not easily purged of conservative contaminants, even if the groundwater flowrate is relatively high (Wilde, 1994).

Sources of Data for Pollutant Concentrations. Using locally derived data in the evaluation of the appropriateness of infiltration is preferable; however, there are often financial and time constraints that preclude local sampling. When localized stormwater quality data are absent, using regional or land-use- specific data may be the best option. Many Municipal Separate Storm Sewer Systems (MS4s) have monitored their stormwater outfalls as part of their permitting process. Outfall data from the site's watershed may prove to be a valuable source of runoff- characterization data, assuming the predominant watershed land use is similar to that of the site. Other sources include land-use- specific data cited in the literature. It has been well-documented that several land uses are more likely to generate specific pollutants of concern (i.e., hydrocarbons from highways and maintenance yards, and metals from watersheds with substantial quantities of galvanized roofing).

When regional or land-use-specific data are absent, the data collected as part of U.S. EPA's stormwater permit program and summarized in the National Stormwater Quality Database (NSQD) are available (Maestre and Pitt, 2005). The NSQD project statistically analyzed the data collected by representative MS4s throughout the United States (as described above) (Table 1 summarizes the data from all land uses; the full database, including tables showing median concentrations for different land uses [industrial, commercial, residential, freeway, open space, etc.], is located at http:// unix.eng.ua.edu/~φitt/Research/ms4/mainms4.shtml, along with several published papers describing the database features and example evaluations). The NSQD database can provide a first estimate of end-of-pipe concentrations of the pollutants of interest. These then could be used as input data to methods/models that predict the susceptibility of groundwater to contamination from stormwater infiltration.

Step 2-Evaluation of the Soil for Infiltration. Soil Chemical Characteristics. The vadose zone is characterized by unsaturated weathered rocks, mineral fragments, decayed organic matter, microorganisms, water, and air. All of these interact to affect how pollutants are transported into and through the vadose zone and potentially into the groundwater. A detailed discussion of the vadose zone can be found in Selker et al. (1999) and Winegardner (1996). The application of vadose zone theory to infiltration is reviewed in Mikula (2005) and Clark et al. (2006).

One important aspect of soils that affects pollutant movement is the soil's organic content, which ranges from less than 0.05% to greater than 80%, but is most often between 2 and 5%. In addition, the multitude of microorganisms in the vadose-zone soils fix nitrogen and degrade organic matter and various pollutants. A second important aspect is the pore volume between the grains, which may be filled with water or air. Water not only transports pollutants through the soil, but it also can dissolve minerals in the soil. The chemical reactions occurring in the soil can affect whether previously trapped pollutants are displaced. In addition, water is a reactant in hydrolysis reactions that occur in the soil.

Infiltration Rate and Capacity. The infiltration capacity of soil is the maximum rate at which infiltration can occur under specific conditions of soil moisture. The infiltration capacity of a soil depends on soil texture, water content of the soil, and its compaction (Fetter, 1994). Water flow in unsaturated soils has been found to follow Darcy's Law in situations where the hydraulic conductivity does not vary with the soil's water content. Various methods, including, but not limited to, ψ-index, Horton, and Greenand-Ampt, are used to predict the infiltration rate and capacity of the soil, given soil type, compaction, and initial soil water moisture.

Step 3-Predicting Groundwater Contamination Potential Below Infiltration Devices. The prediction of groundwater contamination potential can be very complex, depending on the concentration and form of the pollutant, characteristics of the soil, and rate at which water moves through the soil. Mobility is compound-specific and depends on the soil matrix (mostly the soil texture and organic content). Soil characteristics, such as organic content, pH, and permeability, play an important role in pollutant movement.

Two types of models have been developed to assist stormwater managers with evaluating the potential for groundwater contamination. The first is a simplified method that links, in a chart, the mobility of the pollutant, fraction in the filterable (dissolved) phase, and concentration of the pollutant. This information is combined with information on the soil type and general soil reactivity to provide a mobility class, depending on the type of infiltration device used. The second method described is a vadose-zone model, developed for predicting pollutant movement beneath a landfill. This model, unlike the simplified method, predicts concentrations in the groundwater and at various depths in the vadose zone, given input parameters of rainfall amount, soil chemistry, and pollutant concentration.

Simplified Method for Predicting Groundwater Contamination. Table 4 summarizes some of the pollutants found in stormwater that may cause groundwater contamination problems. The Groundwater Recharge Committee of the National Academy of Science (Andelma\n et al., 1994) examined risks associated with recharging groundwater with wastewaters (including stormwater). General causes of concern included the following:

* High mobility (low sorption potential) in the vadose zone,

* High abundance (high concentrations and high detection frequencies) in stormwater, and

* High soluble fractions (small fraction associated with particulates that could be removed at the soil surface by straining or by common sedimentation treatment).

It is possible to assess the need for pretreatment to reduce the runoff pollutant loadings before infiltration. Results of that analysis (the simplified method predicting groundwater contamination potential) are shown in Table 5. Sediment control is critical for long-term infiltration success, because solids will clog infiltration devices, resulting in frequent maintenance or failure. In addition, pretreatment of solids will likely provide the additional benefit of pollutant reduction. Pretreatment also decreases the likelihood that the chemical capacity of the vadose soils will be exceeded during the design life of the device.

This simplified method is based on the following assumptions. The contamination potential is the most critical rating of the influencing factors. As an example, if no pretreatment was to be used before percolation through surface soils, the mobility and abundance criteria are most important. The filterable fraction is not as important, because no pretreatment is being used. If sedimentation pretreatment is to be used before surface infiltration, then some of the pollutants will likely be removed before infiltration. In this case, all three influencing factors (mobility, abundance in stormwater, and soluble fraction) are important. If subsurface injection (with minimal pretreatment) is used, then only abundance is significant. If the pollutant is present in high concentrations, it will likely have an adverse effect on the groundwater. Attenuation through the vadose zone may be insignificant, as the water would bypass it if using direct injection.

Table 5 is only appropriate for initial estimates of contamination potential because of the simplifying assumptions made, such as the likely worst-case mobility conditions using sandy soils having low organic content. If the soil was clayey and/or had a high organic content, then most of the organic compounds would be less mobile than shown on this table. The abundance and filterable fraction information is generally applicable for warm-weather stormwater runoff at residential and commercial area outfalls. The concentrations and detection frequencies would likely be greater for critical source areas (especially vehicle service areas) and critical land uses (especially manufacturing industrial areas), with greater groundwater contamination potential. This is illustrated in the following case study at an industrial site with substantial galvanized roofing, which would be likely to contribute high zinc loadings to an infiltration device.

Case Study Illustrating the Simplified Model. This scenario considers the groundwater contamination potential associated with infiltration of roof runoff from an industrial site. Infiltration is being considered as a method of "disposing" of roof runoff and removing it from the storm sewer system. No site-specific runoff data are available and none will be collected. Is infiltration suitable? The simplified model (Tables 3,4, and 5) are used to evaluate infiltration as a management option. A site survey was conducted to identify the building roofing material. The roof is a pitched galvanized roof in excellent condition. A review of the literature indicates that zinc in runoff from this type of roof can be very high (up to several milligrams per liter).

In Table 3, zinc is shown as having a low mobility in a sandyloam soil and very low mobility in a clayey-loam soil. Using Table 4, zinc's mobility is ranked as low, its abundance in general urban runoff is low to moderate, and its filterable fraction is moderate to high. The mobility is low to very low, depending on the soil type, the filterable fraction is high, and zinc's site-specifc abundance is also rated high. Table 5 evaluates the contamination potential for zinc in infiltration, depending on how the runoff is infiltrated. For surface infiltration, it is anticipated that zinc contamination is not a concern; zinc is likely to react with the soil and be removed before reaching the groundwater, unless the groundwater table is close to the bottom of the infiltration device. Direct injection of the runoff is eliminated as an option, as the absence of soil will result in a higher contamination potential.

Computer Models to Predict Vadose Zone Contaminant Transport. Over the last 15 years, computer models for the vadose zone have become readily available. Many of these models were developed initially to predict plume migration beneath leaking landfills. They typically focused on the behavior of either organic or inorganic pollutants in the vadose zone. One model, SESOIL (Seasonal Soil compartment model), is capable of modeling both organic and inorganic pollutants (Clark et al., 2006).

The SESOIL model is a theoretically based model that calculates mass balances and assumes equilibrium partitioning between phases. It uses three submodels to simulate contaminant fate and transportthe hydrologie cycle, sediment washload cycle, and pollutant fate cycle. The hydrologie cycle models the effect of rainfall, groundwater flow, surface runoff, capillary rise, soil moisture retention, infiltration, and evapotranspiration on pollutant behavior. Next, the sediment washload cycle estimates the amount of erosion and sediment yield to predict the amount of contaminant removed from the system. Finally, the pollutant fate cycle combines natural attenuation of the compound with the values from the hydrologie and sediment washload cycles in a mass-balance approach to predict the final fate of the contaminant. The natural degradation processes considered by SESOIL include diffusion, volatilization, hydrolysis, adsorption, and biodegradation to its pollutant fate cycle (Science Applications International Corporation, 2003).

As described earlier, the specific type of soil and its properties have a profound effect on the movement of water and pollutants. Three main properties were identified by Clark et al. (2006)intrinsic permeability, organic content, and pH. Other factors, beyond the soil itself, also affect pollutant movement and groundwater contamination potential, including pollutant concentration, rainfall, and vadose zone thickness. The primary objective of the research conducted by Clark et al. (2006) and Mikula (2005) was to determine which controlling factors have the greatest influence on the movement of zinc and sodium chloride in the vadose zone beneath a typical infiltration device. This research was undertaken to determine where data collection efforts to support computer modeling should be focused. Zinc and sodium chloride were chosen as example and representative pollutants of interest because of their prevalence in stormwater, solubility, and differing migration rates. A secondary benefit was to evaluate the ease of the model for predicting the potential for groundwater contamination.

The evaluation was performed using a full factorial experiment, so that the effects of concentration, rainfall, vadose zone thickness, intrinsic permeability, organic content, and pH (plus all possible interactions) on the maximum penetration depth of zinc and sodium chloride could be evaluated (Box et al., 1978). The input variables are shown in Table 6. Actual soils that represent high and low values of intrinsic permeability, organic content, and pH were used in the model to create a more realistic evaluative environment. The time step of each simulation was set at 1 year.

Results indicated that the rainfall amount was a common factor controlling zinc, sodium, and chloride migration. Because the zinc and sodium chloride were assumed to be dissolved in infiltrating stormwater, higher rainfall amounts would naturally allow the pollutants to migrate deeper in the vadose zone. In addition, concentrations were influential in zinc(II) migration. Higher concentrations of zinc(II) reduced the zinc sorption in any soil layer, because the available sorption sites in the upper layers were filled before the zinc in the stormwater was completely removed.

Vadose zone thickness was not shown to have an effect, but this was likely an artifact of the way SESOIL calculates pollutant migration. Vadose zone thickness needs to be considered when constructing infiltration devices. pH was not a recognized significant factor in zinc migration, because non-neutral-pH hydrolysis could not be modeled. In reality, zinc should be more mobile in acidic and alkaline soils because of its higher solubility during extreme pH conditions, although zinc is more soluble and more mobile in alkaline conditions only if organic matter is present (Shuman, 1999). For the narrow pH range of typical soils, pH is not expected to be a significant factor. Organic matter also was not a significant controlling factor, even though filtration work has shown that organic matter affects zinc migration. This is likely a result of the lack of organic matter in the soils below the top layer.

The predicted migration depth of chloride was the most realistic of the three pollutants studied, based on the literature. Chloride's nonreactive nature made its fate and transport very easy to model with SESOIL. The limitation to this modeling exercise was the lack of field data for calibration, although the exercise demonstrated the power of the model in evaluating infiltration device placement in relation to soil type and groundwater depth.

Case Study Illustrating Computer Modeling. Using the same scenario described above, SESOIL was used to furth\er evaluate the use of infiltration for disposal of roof runoff at an industrial site. The data required to run the model consisted of the influent dissolved zinc concentration (data from the literature on roof runoff), soil type (obtained from local soil maps; SESOIL's database contains the soil physical and chemical parameters for each delineated soil in the United States), zinc partition coefficient to the soil (from literature; organic coefficients are built into SESOIL's database), and the monthly rainfall depth for the site (the municipality's average annual rainfall depth is typically used, although monthly rain totals can also be used). For example, if this site were located in West Palm Beach, Florida (rainfall = 154 cm/ y), the soil was sandy, and the influent zinc concentration was 2.1 mg/L, anticipated zinc migration was as great as 1.5 m per year. If the same soil type and influent concentration were evaluated for a site in Phoenix, Arizona (annual rainfall 6.7 cm/y), zinc migration depth could be up to 0.15 m/y. If the ground water was shallow (less than 2 to 3 m), there is a concern for groundwater contamination, although the contamination would be evident sooner for the higher rainfall site. Other options would need to be evaluated. Infiltration may still be suitable if a different soil was used. Adding organic matter to the sandy soil will improve its ability to remove and retain pollutants, potentially enough to ensure that the soil's zinc removal capacity was not exhausted during the life of the infiltration device.

Conclusions and Recommendations

Many types of stormwater infiltration approaches have been used in urban areas to decrease surface discharges of stormwater. The most common include the following:

* Surface infiltration devices (i.e., grass filters, grass-lined drainage swales, dry [percolating] basins, porous pavement, and grid pavers)-infiltration occurs through turf and surface soils or through pavement for grid pavers and porous pavement, providing the best opportunity for pollutant trapping in the vadose zone, assuming that the surface soil and turf layers are not removed during construction. The devices listed below discharge stormwater below organic soils, allowing increased pollutant movement to the groundwater.

* Subsurface infiltration wells, dry wells, French drains or soak- aways (source area infiltration pits and roof runoff infiltration pits), or percolating pipes; the wells are deep, relatively small- diameter holes, allowing storm water to be discharged to deep soil horizons, sometimes directly to saturated zones. French drains are relatively shallow, but still bypass the surface organic soil layer. Percolating pipes are conventional subsurface drainage pipes, but with perforations through the pipe wall or gaps between pipe segments. They are typically wrapped in geotextile fabric with coarse gravel used as a trench backfill material.

When selecting the appropriate infiltration device for a site (or for determining whether infiltration is a viable option), it is crucial to evaluate the potential for groundwater contamination below the device. As described above, the prediction of the vulnerability of groundwaters to contamination from surface water infiltration is a three-step process. The first step is determining the pollutants of concern and the concentrations and chemical forms of these pollutants. The stormwater pollutants of most concern (those that may have the greatest potential adverse effects on groundwaters) include the following:

* Nutrients-possibly nitrates in areas having high stormwater nitrate concentrations;

* Pesticides-lindane and chlordane;

* Other organics-1,3-dichlorobenzene;

* Pathogens-Enteroviruses and possibly other pathogens, including Shigella, Pseudomonas aeruginosa, and various protozoa;

* Heavy metals-nickel and zinc, followed by chromium and lead; and

* Salts-chloride in northern areas where deicing salts are used for traffic safety.

Identifying the soil characteristics for a particular site affecting pollutant migration is the second step. The information required includes the following:

* Soil texture-sand, silt, clay, loam, or a combination of these;

* Intrinsic permeability;

* Hydraulic conductivity;

* pH;

* Organic content; and

* Cation exchange capacity.

The third step requires using the information determined in the first two steps to predict the potential for groundwater contamination. Two approaches were described in this article-a simplified method that likely predicts the worst-case groundwater contamination potential and a computer model that uses actual or estimated field data to predict the depth of migration in a prespecified time period. A case study was used to illustrate the applicability of the two methods and the types of results obtainable from each method/ model.

In general, to prevent groundwater contamination below infiltration devices, surface percolation devices (i.e., grass swales and percolation ponds), which have a substantial depth of underlying organic-rich soils above the groundwater, are preferable to using subsurface infiltration devices (i.e., dry wells, trenches or French drains, and especially injection wells), unless the runoff water is known to be relatively free of pollutants (i.e., these subsurface devices may be appropriate for infiltrating the runoff from a clean, nonmetallic roof). Infiltration of stormwater from residential areas is also safer than from more contaminated areas, unless suitable pretreatment is used. Pretreatment to remove solids and paniculate-associated pollutants can be used to reduce the required maintenance frequency of the infiltration component of the treatment train.,


Funding for this study was provided from two primary sources- U.S. EPA and the Water Environment Research Foundation. The authors thank J. Bradley Mikula, formerly a M.S. Environmental Pollution Control student at Penn State Harrisburg (currently with Michael Baker Jr. in Pittsburgh, Pennsylvania), for his work on the vadose- zone computer modeling.

Submitted for publication March 30, 2006; revised manuscript submitted July 25, 2006; accepted for publication July 26,2006.

The deadline to submit Discussions of this paper is April 15, 2007.


Andelman, J.; Bauwer, H.; Charbeneau, R.; Christman, R.; Crook, J.; Fan, A.; Fort, D.; Gardner, W.; Jury, W.; Miller, D.; et al. (1994) Ground Water Recharge Using Waters of Impaired Quality; Committee on Ground Water Recharge, Water Science and Technology Board, National Research Council, National Academy Press: Washington, D.C.

Armstrong, D. E.; Llena, R. (1992) Stormwater Infiltration: Potential for Pollutant Removal. Prepared for the Wisconsin Department of Natural Resources (Madison) and the U.S. Environmental Protection Agency: Chicago, Illinois.

Bannerman, R.; Owens, D. W.; Dodds, R. B.; Hornewer, N. J. (1993) Sources of Pollutants in Wisconsin Stormwater. Water Sd. Technol., 28 (3-5), 241.

Box, G. E. P.; Hunter, W. G.; Hunter, J. S. (1978) Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building. Wiley & Sons: New York.

Butler, K. S. (1987) Urban Growth Management and Groundwater Protection: Austin, Texas. Planning for Groundwater Protection; Academic Press, Inc.: New York.

Camesano, T. A.; Logan, B. E. (1998) Influence of Fluid Velocity and Cell Concentration on the Transport of Motile and Non-Motile Bacteria in Porous Media. Environ. Sd. Technol., 32 (34), 1699,

Clark, S. E. (2000) Urban Stormwater Filtration: Optimization of Design Parameters and a Pilot-Scale Evaluation, Ph.D. Dissertation, University of Alabama at Birmingham.

Clark, S.; Pitt, R. (1999) Stormwater Treatment at Critical Areas: Evaluation of Filtration Media for Stormwater Treatment; U.S. Environmental Protection Agency: Cincinnati, Ohio.

Clark, S. E.; Baker, K. H.; Mikula, J. B.; Burkhardt, C. S.; Lalor, M. M. (2006) Infiltration vs. Surface Water Discharge: Guidance for Storm-water Managers, Project Number 03-SW-4; Water Environment Research Foundation: Alexandria, Virginia.

Crites, R. W. (1985) Micropollutant Removal in Rapid Infiltration. In Artificial Recharge of Gmundwater, Asano, T. (Ed.); Butterworth Publishers: Boston, Massachusetts, 579-608.

Domagalski, J. L.; Dubrovsky, N. M. (1992) Pesticide Residues in Oroundwater of the San Joaquin Valley, California. J Hydrol, 130 (1- 4), 299.

Fetter, C. W. (1994) Applied Hydmgeology, 3rd ed.; Prentice Hall: Upper Saddle River, New Jersey.

German, E. R. (1989) Quantity and Quality of Stormwater Runoff Recharged to the Floridan Aquifer System Through Two Drainage Wells in the Orlando, Florida Area. Water Supply Paper 2344; U.S. Geological Survey: Denver, Colorado.

Gold, A. J.; Groffman, P. M. (1993) Leaching of Agrichemicals from Suburban Areas. In Pesticides in Urban Environments-Fate and Significance, Racke, K. D., Leslie, A. R. (Eds.), ACS Symposium Series No. 522; American Chemical Society, Washington, D.C.

Hampson, P. S. (1986) Effects of Detention on Water Quality of Two Stormwater Detention Ponds Receiving Highway Surface Runoff in Jacksonville, Florida, Water-Resources Investigations Report 86- 4151; U.S. Geological Survey: Denver, Colorado.

Jensen, P.; Rifai, H.; Battenfield, T. (2002) Lessons Learned from Texas Bayous Identifying Bacteria Sources. Publ. Works., 133 (5), 308.

Jury, W. A.; Spencer, W. F.; Farme, W. J. (1983) Model for Assessing Behavior of Pesticides and Other Trace Organics Using Benchmark Properties. I. Description of Model. J. Environ. Qual.. 12 (2), 558.

Ku, H. F. H.; Simmons, D. L. (1986) Effect of Urban Stormwater Runoff on Gmundwater Beneath Recharge Basins on Long Istand, New York, Water-Resources Investigations Report 85-4088; U.S. Geological Survey: Denver, Colorado.

Maestre, A.; Pitt, R. (2005) The National Stormwater Quality Database, Version 1.1, A Compilation and Analysis of NPDES Stormwater Monitoring Information (fina\l draft report); U.S. Environmental Protection Agency, Office of Water: Washington, D.C.

Maryland Department of the Environment (2000) Maryland Stormwater Design Manual, Volumes I & II; Maryland Department of the Environment: Baltimore, http://www.mde.state.md.us/Programs/Water Programs/SedimentandStormwater/stormwater_design/index.asp.

Mikula, J. B. (2005) Modeling Zinc and Sodium Chloride Migration in Vadose Zone Soils Beneath Stormwater Infiltration Devices, M.S. Environmental Pollution Control, Penn State Harrisburg, Middletown, Pennsylvania.

Nightingale, H. I. (1987) Accumulation of As, Ni, Cu and Pb in Retention and Recharge Basins Soils from Urban Runoff. Water Resour. Bull., 23 (4), 663.

Pitt, R.; Clark, S.; Farmer, K. (1994) Protection of Groundwater from Intentional and Nonintentional Stormwater Infiltration, EPA- 600/SR-94-051, PB94-165354AS; U.S. Environmental Protection Agency, Storm and Combined Sewer Program: Cincinnati, Ohio.

Pitt, R.; Clark, S.; Partner, K.; Field, R. (1996) Groundwater Contamination from Stormwater Infiltration; Ann Arbor Press: Chelsea, Michigan.

Pitt, R. E.; Field, R.; Lalor, M.; Brown, M. (1995) Urban Stormwater Toxic Pollutants: Assessment, Sources, and Treatability. Water Environ. Res., 67 (3), 260.

Ragone, S. E. (1977) Geochemical Effects of Recharging the Magothy Aquifer, Bay Park, New York, with Tertiary-Treated Sewage, Geological Survey Professional Paper 751-D; U.S. Geological Survey: Reston, Virginia.

Robinson, J. H.; Snyder, H. S. (1991) Golf Course Development Concerns in Coastal Zone Management. Coastal Zone '9I: Proceedings of the Seventh Symposium on Coastal and Ocean Management, Long Beach, California, July 8-12; American Society of Civil Engineers: New York.

SaIo, J. E.; Harrison, D.; Archibald, E. M. (1986) Removing Contaminants by Groundwater Recharge Basins. J. Am. Water Works Assoc., 78 (9), 76.

Schiffer, D. M. (1989) Effects of Three Highway-Runoff Detention Methods on Water Quality of the Surficial Aquifer System in Central Florida, Water-Resources Investigations Report 88-4170; U.S. Geological Survey: Denver, Colorado.

Science Applications International Corporation (2003) RISKPRO. SESOIL Reference Guide and User's Guide, Version 3.1. Science Applications International Corporation: Beltsville, Maryland.

Selker, J. S.; Keller, C. K.; McCord, J. T. (1999) Vadose Zone Processes; Lewis Publishers: Boca Raton, Florida.

Shirmohammadi, A.; Knisel, W. G. (1989) Irrigated Agriculture and Water Quality in the South. J. lrrig. Drainage Eng., 115 (5), 791.

Shuman, L. M. (1999) Effect of Organic Waste Amendments on Zinc Adsorption by Two Soils. Soil Sd., 164 (3), 197.

Troutman, D. E.; Godsy, E. M.; Goerlitz, D. F.; Ehrlich, G. G. (1984) Phenolic Contamination in the Sand-and-Gravel Aquifer from a Surface Impoundment of Wood Treatment Wastewaters, Pensacola. Florida, Water-Resources Investigations Report 84-4230; U.S. Geological Survey: Denver, Colorado.

Unice, K. M.; Logan, B. E. (2000) The Insignificant Role of Hydrodynamic Dispersion on Bacterial Transport. J. Environ. Eng., 126 (6), 491.

U.S.Environmental Protection Agency (1992) Manual: Guidelines for . Water Reuse, EPA-625/R-92-004; U.S. Environmental Protection Agency: Washington, D.C.

White, E. M.; Dombush, J. N. (1988) Soil Change Caused by Municipal Waste Water Applications in Eastern South Dakota. Water Resour. Bull., IA (2), 269.

Wilde, F. D. (1994) Geochemistry and Factors Affecting Ground- Water Quality at Three Storm-Water-Management Sites in Maryland: Report of Investigations No. 59; Department of Natural Resources, Maryland Geological Survey: Baltimore, Maryland.

Wilson, L. G.; Osborn, M. D.; Olson, K. L.; Maida, S. M.; Katz, L. T. (1990) The Groundwater Recharge and Pollution Potential of Dry Wells in Pima County, Arizona. Groundwater Monit. Rev., 10 (3), 114.

Winegardner, D. L. (1996) An Introduction to Soils for Environmental Professionals. Lewis Publishers: Boca Raton, Florida.

Shirley E. Clark1*, Robert Pitt2

1* Assistant Professor of Environmental Engineering, Penn State Harrisburg, 777 W. Harrisburg Pike TL-105, Middletown, PA 17057; e- mail: [email protected]

2 Cudworth Professor of Urban Water Systems, Department of Civil, Construction, and Environmental Engineering, University of Alabama, Tuscaloosa.

Copyright Water Environment Federation Jan 2007

(c) 2007 Water Environment Research. Provided by ProQuest Information and Learning. All rights Reserved.