Emergency response to stormwater contamination: A framework for containment and treatment

Abstract

This paper presents a Stormwater Emergency Response Framework (SERF) for use in the containment and treatment of stormwater runoff following a hazardous material release. The framework consists of four high level process steps and a decision tree. These resources are intended to assist stormwater managers in fulfilling their emergency response responsibilities within the United States’ National Incident Management System. Robust hydraulic and watershed modeling may take weeks to months to develop for a contaminated site, whereas decisions made in the initial hours can have a significant impact on limiting contamination spread. Many web resources are publicly available to assist responders in visualizing stormwater runoff flow paths. A case study provided in this paper also demonstrates how simple calculations may be utilized to estimate peak flows and storage volumes necessary to respond to precipitation events immediately. These calculations are useful for decision makers’ allocation of containment and treatment resources within the impacted area. This includes where to deploy available resources to minimize contamination risks to downstream communities and where supplemental resources from outside partners are urgently needed.

1. Introduction

1.1. Background

Management of contaminated water is often an after-thought when first responding to a natural or human caused disaster (e.g., a bomb, an industrial explosion, a cross-connection, a train derailment, or a tanker truck wreck). Treating the sick and injured, fighting fires, and preventing explosions are the immediate priority; however, decisions during the initial hours following an incident can have tremendous impacts on public and aquatic health long after the event. Following such a hazardous material release, contaminated water can be generated in a variety of ways including:

• Direct contamination of drinking water or waterways,

• Decontamination washdown activities of indoor/outdoor areas and personal protective equipment,

• Runoff from precipitation events/groundwater intrusion.

There is no tool currently available to stormwater managers that can help them in deciding what containment, treatment, or diversion resources should be placed in which locations to optimally reduce the risk to human and aquatic health other than a person’s best judgment. The goal of this paper is to introduce the reader to a framework and tools for making real-time decisions for managing contaminated water prior to development of a more detailed model.

The quantity, storage, and remobilization of contaminated water after a wide-area disaster may be staggering. For example, during the simulated remediation of a suburban area following a release of radiological contamination, researchers estimated millions of gallons of contaminated water would be generated by decontamination efforts (Boe et al., 2019). After the Fukushima nuclear disaster, reactors experienced groundwater intrusion and exfiltration through cracks in their foundations that resulted in 264 million gallons of contaminated water being stored in onsite tanks (Beiser, 2018; Gallardo and Marui, 2016; Takenaka, 2019). Enormous amounts of water are sometimes necessary for fire suppression; for example, it was estimated that half a million gallons of water were necessary to suppress a 30 ton ammonium nitrate explosion at a Texas fertilizer facility (Laboureur et al., 2016). Further, adverse weather conditions, particularly rainfall, are known to play a role in the spread of hazardous contamination (Ruckart et al., 2004). After both the Fukushima and Chernobyl nuclear disasters, radiological contamination spread to far-reaching areas of the watershed due to rainfall-runoff processes (Konoplev et al., 2016; Laceby et al., 2016).

Contaminated water poses a threat to the public through direct physical contact, ingestion, or inhalation. This may be through drinking water (if the treatment plant was not designed to remove the contaminant), showering, basement backups of sewage, exposure at combined sewer overflows, or direct contact with contaminated waterways. Utility workers may be particularly vulnerable due to exposure to contaminated source waters and sludge handling at drinking water and waste-water treatment plants (U.S. EPA, 2016). Preventing the spread of contamination with sound early decision-making can limit the extent of chemical, biological, or radiological (CBR) agents spatially, minimize acute and chronic exposure to much larger population centers, and reduce ecosystem destruction. Additionally, treating, containing, and preventing further water contamination can result in less costly remediation and in much less hazardous waste produced, transported, and disposed of in highly regulated secure landfills.

1.2. Need

In the United States, there are differences in approach and structure of stormwater utilities at the local level. There are communities with dedicated stormwater districts with complicated stormwater networks and full-time staff responsible for operation and maintenance. In contrast, there are also many small and/or rural communities with a county engineer with more multidisciplinary responsibilities. In both cases, responding to routine system issues leaves little time for developing tools and approaches for emergency only use. At the national level, an overarching framework exists to manage disasters. The Federal Emergency Management Agency (FEMA) maintains the National Incident Management System (NIMS) which provides a common, nationwide approach to engage communities in managing threats and hazards and contains the National Response Framework (“National Incident Management System,” 2020). The NIMS applies to any type of incident regardless of cause, size, location or complexity and is structured around an Incident Command System (ICS). This allows a single unified response that avoids duplicative efforts by multiple agencies (Broder and Tucker, 2012). The ICS can be as simple as involving several agencies from a single jurisdiction such as the local fire, police, and water departments to coordinate response to a small hazardous spill to a complex response involving multiple agencies across jurisdictions by using the Multi-Agency Coordination System (MACS) for large-scale emergencies (Hunter, 2009, 2017). At the international level, the United Nations Office for the Coordination of Humanitarian Affairs and the United Nations Environment Programme have a Joint Program that coordinate guidelines for environmental emergencies and outline a number of global frameworks for emergency response and recovery (UN Environment & OCHA, 2017). As useful as these overarching frameworks are to enable coordinated emergency responses, there are knowledge gaps that exist as one works through the frameworks, especially down to the local level where resources need to be allocated and hazardous material needs to be treated or disposed of.

1.3. Framework goals

For many utilities, containment and treatment actions within the initial hours require quick decisions without the benefit of highly precise models. Conceptual frameworks have proven useful tools in organizing technical capabilities related to flood forecasting (Maidment, 2017). The proposed Stormwater Emergency Response Framework (SERF) being described in this paper utilizes an approach that couples local knowledge of drainage systems with simple calculations to optimize resources and minimize risks to downstream waterways and neighborhoods. The merits of an easily scalable approach to water-related management (volume, quality, etc.) are well established (Montgomery et al., 1995; U.S. EPA, 2008). It is also critical to consider the long-term recovery of an affected area. Clean-up efforts may take weeks to years depending on the severity of the incident. Within that timeframe, there will be additional precipitation events that must be accounted for, but there is also the extra time to acquire additional treatment and containment resources that can then be optimized to ensure the most complete and safe recovery.

The SERF described herein is adaptable and ready for rapid execution during the first few hours of a disaster as well as for the long-term recovery. It utilizes simple but informative calculations to help managers optimize their diversion, containment, and treatment strategies to minimize risks to downstream waters to the greatest extent available given the local inventory of Best Management Practices (BMPs), equipment, and other resources. It also aids decision-makers in identifying which processes and tools should be invested in to develop more sophisticated modeling and analyses in the years to come. The SERF raises awareness to existing decontamination and containment strategies currently in practice and provides a step-by-step decision tree with example scenarios and pragmatic resources adaptable to real events. Advanced training on the tool could also assist utilities in identifying on-hand supplies and equipment (e.g. pumps, excavators, etc.) that could be useful during an emergency response and where it may be advantageous to expand their inventories.

2. Framework

2.1. Process steps

The SERF aims to provide a methodical process for local stormwater managers and engineers to efficiently participate in the emergency response and recovery process. The main components presented here were developed based on the authors’ experience in the stormwater modeling and emergency response sectors and informational interviews with two stormwater utilities with different emergency response protocols. It consists of the following four high-level process steps:

Step 1 Determine the extent and type of contamination;

Step 2 Locate stormwater infrastructure and delineate drainage paths;

Step 3 Diversion, containment, and treatment; and.

Step 4 Long-term and extreme weather planning.

Additional granularity of these steps is provided in the form of a decision tree where diamonds represent system questions, rectangles represent intermediary actions, and ovals represent initial and final actions (Fig. 1). The decision tree walks the stormwater practitioner through critical questions that require technical calculations or resource allocation decisions tailored to their system. The decision tree primarily focusses on identification of existing stormwater Best Management Practices (BMPs) in the contaminated area and quantification of peak stormwater flows. As such, the SERF provides a structured method to bound the contamination issue and determine the best allocation of resources. An example of how to use the SERF is provided in the case study section of this paper. The remainder of this section provides further context and identification of resources for the SERF’s main steps.

Fig. 1.

Schematic depicting the SERF’s decision tree for responding to wide-area stormwater contamination event. BMPs in the flow chart refer to management practices and equipment.

Step 1 Determine the Extent and Type of Contamination.

The first step in the SERF is to identify the type and extent of the contaminant (e.g, surface area, depth, and concentration). Determining this will help identify vulnerable areas of the watershed, important infrastructure, and appropriate BMPs. The scope of the SERF is not limited by contaminant type; however, it is recognized that CBR agents require drastically different analytical capabilities and are susceptible to different environmental fate and transport processes (Lemieux et al., 2018; U.S. EPA, 2015). For example, enumeration of biological contamination may require a longer turn-around time due to incubation than gamma emitting radionuclides which is often measurable in the field. A discussion of stormwater contaminants of concern related to a wide-area terrorism incident is available in (Mikelonis et al., 2018). In addition to tanker truck spills of chemical contaminants, long half-life radionuclides such as ¹³⁷Cs and ⁹⁰Sr and environmentally persistent bioterrorism agents such as Bacillus anthracis spores are particularly concerning and could have long-term impacts on stormwater systems. The extent of the contamination informs the demarcation between diversion strategies and containment/treatment controls. For the earliest stages of response, visual inspection of the incident must be relied on to determine the extent of the contamination from the incident or early response actions, followed by targeted grab sampling and finally a thorough, overarching sampling plan during the recovery phase. This sampling plan will need to be tailored to hydroclimatic setting and season, with different times of year requiring more expediency than others (e.g., the rainy season). Many resources are available to assist in sample collection, including, but not limited to, sample analytical procedures (see summary of methods in (Campisano et al., 2017)) and tools for selecting sampling locations (Kincaid et al., 2015; Matzke et al., 2014) (Table 1).

Table 1.

Supplemental tools^a.

Sampling Analytical Methods

•U.S. EPA Environmental Sampling and Analytical Methods (ESAM) Program Sampling Plans • spsurvey (R software statistical package) • Visual Sample Plan (VSP) software Watershed Mapping/Parameters • Model My Watershed webpage • StreamStats webpage • U.S. EPA’s How’s My Waterway webpage • U.S. EPA’s WATERS GeoViewer webpage Flow Accumulation Pathways • ArcGIS software’s hydrology toolsetb • PCSWMM software’s watershed delineation toolb • U.S. EPA EnviroAtlas webpage’s raindrop tool • GRASS GIS software’s r.watershed module Precipitation • NOAA ATLAS 14 Precipitation Frequency Data Server Satellite Imagery • Google Earth software

^aNon-exhaustive list of tools useful during initial response. These tools are not intended to replace a calibrated watershed/hydraulic model if maintained by the utility.

^bRequires a license fee.

Step 2 Locate Stormwater Infrastructure and Delineate Drainage Paths.

Understanding what waterbody(ies) the contaminated area drains to and its flow paths are critical to determine effective mitigation strategies. In the United States, several governmental agencies, nonprofits, and commercial vendors offer resources that allow the user to search for a particular watershed and visualize characteristics such as boundaries and flow statistics based on geographic coordinates (Aufdenkampe et al., 2017; Ries III et al., 2017; U.S. EPA, 2020b, 2020c). These tools are often based on the United States’ National Hydrography Datasets (Moore et al., 2019) (Table 1). Subscription and open source services also have tools that allow the user to select a single point to calculate the total contributing watershed area, trace the furthest downhill path of a water droplet, and/or display flow accumulation lines (CHI, 2020; ESRI, 2020; Neteler et al., 2012; U.S. EPA, 2020a). Input requirements for these tools are typically a digital elevation model file, which are available from the USGS or from local survey initiatives. The accuracy and/or utility of the results from these tools are highly dependent on the resolution of the elevation survey and should be ground-truthed by field investigations. In the absence of an area covered by these or equivalent local tools, watershed delineation is still possible using a topographic map. This involves identifying a waterbody outlet, highlighting nearby watercourse flow directions, marking ridge lines and saddles, and hand drawing arrows indicating the direction of flow. Locally maintained databases of stormwater infrastructure, as-built drawings, satellite imagery (Google, 2020), and/or walking the area are necessary for rapidly identifying stormwater infrastructure in the contaminated and adjacent zones (e.g., look in a catch basin/manhole and determine which pipe is lower/larger to establish flow direction if personal protective equipment provides adequate protection). These actions should be taken within the initial hours of the contamination to bound the problem. At this point, the responders should also assess the impacts that infiltration of contaminated water will have on the water supply. In many locations more time is required for the assessment of groundwater intrusion, however areas where there is potential for contamination to short-circuit directly to the water table (e.g, sand and loamy sand areas with high hydraulic conductivity) should be identified and actions such as laying plastic sheeting, where possible, to prevent this discussed.

Step 3 Diversion, Containment, and Treatment.

This step in the SERF is related to identifying locally available diversion, containment, and treatment technologies that can be applied to impacted sites both as short-term and long-term mitigation and recovery technologies. This includes existing BMPs and techniques tailored to the contaminant such as lining basins to prevent seepage of contaminated water into the groundwater and absorption, diversion, or collection technologies for the overland flow (Biwer et al., 2008). Examples of equipment range from simple hand tools (e.g., shovels, brooms, squeegees, skimmers, and mops), to granular adsorbent media, to basin filter inserts, berms, and inlet seals, to large excavators (e.g. to dig diversion ditches or expand available storage volumes for containment) (Fig. 2). Multiple strategies may need to be attempted during this stage of the framework. For example, during a train derailment in Miamisburg, Ohio, several thousand gallons of white phosphorus and animal tallow were released into a nearby creek that fed into the Great Miami River. Responders initially deployed booms which were washed out by flows within a week. There were also several attempts to build dams before a successful diversion channel was finally constructed to prevent contamination from the creek entering the Great Miami River (Scoville et al., 1989). After a warehouse explosion in Tianjin, China, surface water outlets, ditches and rain drainage pipes were closed (using cement) and cofferdams of sand and earth were constructed to isolate an approximately 1 million square foot area from releasing cyanide-tainted rainfall runoff (Zhang et al., 2017). The cofferdams were designed to prevent infiltration of the contamination into the groundwater.

Fig. 2.

Example diversion, containment and treatment devices: A) granular absorbant media, B) inflatable dam, C) oleophilic skimmer, D) containment booms.

Step 4 Long-term and Extreme Weather Planning.

The final step in the SERF involves estimating the capacity of the chosen diversion, containment, and treatment strategies from step 3 and understanding the risk that a rain event will exceed this capacity or infiltrate into the groundwater. This must be estimated based on historical and projected rainfall/runoff information and the expected duration of the recovery process (Eldardiry et al., 2015; Koutsoyiannis et al., 1998; Kunkel et al., 2007). This assessment is necessary to plan for additional controls to ensure adequate detention, diversion, and treatment over the duration of the cleanup. For example, there is an 87.5% chance the site will experience a 2-yr storm (or larger) within a three-year recovery period. Such analyses may highlight flow paths in need of secondary or tertiary controls to manage runoff to meet acceptable levels of risk.

2.2. Case study

In the subsections below, we present a step-by-step example that applies the decision tree to a neighborhood scenario involving a hypothetical contamination event. The event covers portions of a commercial property, an adjacent residential subdivision, and two roadside ditches. The parameters of the scenario are summarized in Table 2.

Table 2.

Neighborhood case study scenario.

Characteristic	Quantity or Type
Watershed area	75 acres
Size of initial contamination plume	25 acres
Contaminate	Unidentified chemical spill
Land use	commercial, residential, and farmland
Stormwater infrastructure/BMPs	• ditches
	• stormwater pipes (separate)
	• detention basins
Readily available containment equipment	• absorbant pads (50)
	• sand bags (20)
	• pumps (3)
	• inflatable sewer plugs

2.3. Determine type and extent of contamination

In this hypothetical suburban scenario, chemical contamination covers several roof tops, a commercial parking lot, a portion of a residential street with stormwater inlets and pipes, and a portion of an old country road with conventional ditches along both sides (Fig. 3). The contamination source is unknown and therefore will require further laboratory testing to specify its physical and chemical properties however its’ residue is visible on surfaces. Although there are numerous factors to consider regarding the characteristics of the hazard, physical characteristics that are particularly relevant to water-related planning are solubility and specific gravity, namely does it dissolve and will it float or sink when encountering water. For example, this can inform whether one might use “underflow” or “overflow” bypass devices during events that might overwhelm the storage/treatment capacity within the contaminated area. For unknown or new hazardous materials full site characterization may take a long time and the extent of contamination may need to be revised as data becomes available. However, the identification of water flow paths and availability of BMP resources may continue in parallel.

Fig. 3.

Hypothetical contamination event in a suburban neighborhood.

2.4. Identify water flow entry and exit locations

After determining the extents and type of contamination, the first question in the SERF’s decision tree is: “Are all water flow paths entering and exiting the contaminated area known?” (Fig. 1). This underscores the importance of the second major step in this framework—locating stormwater infrastructure and determining flow directions. In this case study, there are two entrances and four exits (Fig. 4a). The entrances are the conventional ditches that drain from the top right of the figure towards the top left. The exits include those same two ditches where they leave the contaminated area, along with a stormwater pipe in the neighborhood flowing toward the bottom right of the figure and a stormwater network draining the commercial property to the detention basin in the center left of the figure.

Fig. 4.

A) Identification of locations where water could enter or exit the contaminated area to inform where diversion vs. containment/treatment measures may be needed. Enter 1 & Enter 2: conventional ditches entering contaminated area; Exit 3: storm sewer exiting contaminated area; Exit 4: detention basin exiting contaminated area; Exit 5 &6: conventional ditch exiting contaminated area.

2.5. Delineate drainage areas to entry and exit points

Once the entry and exit points are identified, delineating drainage areas to each of those locations is necessary to better understand the volumes of water that have the potential to drain to those points during a given precipitation event (Fig. 4b). In this scenario, the largest drainage areas are attributable to the commercial lot (area #4 = 12 acres) and the residential neighborhood (area #3 = 8 acres), whereas the conventional ditches along the country road have smaller drainage areas.

2.6. Drainage area delineation

Identify BMPs and equipment available to divert water from entering the contaminated area.

After delineating drainage areas to all entrances and exits, the decision tree advances to: “Item 1: Identify readily available BMPs and equipment to divert water from entering the contaminated area” (Fig. 1). In this case study, the utility has available 50 absorbent pads, 20 sand bags, three pumps rated at 10 ft³/s, and several inflatable sewer plugs (Table 2). The two entrance points into the contamination zone are the conventional roadside ditches in the top right of Fig. 3. BMPs that could intercept and divert the off-site stormwater around the contaminated area would help to minimize the volume of water that needs to be contained and treated.

In the immediate response period, any diversion that can be installed ahead of the next precipitation event is useful. That said, employing some quick calculations with a standard design storm (such as the 1-year event) can help provide context to the level of service a diversion can provide. Depending on the season, weather forecast, and other factors, there is only about a 2% risk of experiencing a 1-yr event in any given week of the year. That risk increases to ~8% in any 30-day period, 15% in any 60-day period, and 99% within a one-year period (National Weather Service, 2020). Diversions that provide a level of service equivalent to the 1-year storm, therefore, have relatively little chance of being overwhelmed in the first few weeks following the event but will almost certainly be exceeded within a year, underscoring the importance of long-term asset allocation and extreme event planning.

The success of a diversion depends on its ability to divert the maximum (peak) discharge for any given event. The Rational Method (Q_p = CIA) is a straightforward method to estimate the peak discharge for recurrence interval events in small drainage areas such as the ditches in this example. “Q_p” is the peak discharge (ft³/s), “C” is a runoff coefficient ranging 0–1 depending on land cover (e.g. lawns = 0.05–0.10, impervious areas = 0.75–0.95), “I” is the rainfall intensity (inches/hr) that can be retrieved for any location in the US from the National Oceanic and Atmospheric Administration’s (NOAA) Precipitation Frequency Data Server as a function of the time of concentration (often less than 10 min in small/urban drainages such as these), and “A” is the contributing drainage area (acres) (NOAA, 2017).

In the example herein, the estimated peak discharge is 9.4 ft³/s for ditch #1 and 11.3 ft³/s for ditch #2 (Table 3). With three available pumps rated at 10 ft³/s, one pump could be placed in ditch #1 and two pumps could be placed in parallel in ditch #2 to exceed a 1-yr level of service for both ditches (Fig. 5). Sand bags could be used to complement the pumping operations by inducing pooling and blocking the uncontaminated stormwater that would otherwise flow into the contamination zone.

Table 3.

Drainage area, runoff, and rainfall characteristics for using the Rational Method to estimate the 1-yr peak discharge for several entrances and exits to the contamination zone.

Location	Drainage Area (A)	Runoff Coefficient (C)a	Time of Concentration (Tc)b	Rainfall Intensity Ic	Peak Discharge (Qp)
1 – ditch	6 acres	0.38	6 min	4.1 in/hr	9.4 ft3/s
2 – ditch	4 acres	0.75	7 min	3.75 in/hr	11.3 ft3/s
3 – sewer	8 acres	0.55	8 min	3.7 in/hr	16.4 ft3/s
6 – ditch	1 acre	0.80	2 min	5.0 in/hr	4.0 ft3/s

^aRunoff coefficients for a given drainage area are weighted by their proportions of runoff generating surfaces. For example, for ditch #1, if 2.2 acres were paved (C = 0.95) and 3.8 acres were lawn (C = 0.05), the area-weighted C would equal (2.2 x 0.95 + 3.8 x 0.05)/6 = 0.38.

^bTime of Concentration is defined as the time required for stormwater to reach the watershed outlet from the most remote point (often < 10 min in small/urban watersheds) (NRCS, 2004).

^c(NOAA, 2017).

Fig. 5.

Control Implementation (<72 h after contamination event).

2.7. Identify available BMPs to contain/store and treat the water within the contaminated area

Concurrent with the diversion efforts to keep unsullied stormwater from entering the contamination zone, containment and treatment BMPs are necessary to keep tainted water from leaving the contamination zone including infiltration (Item 2 in the Decision Tree, Fig. 1). In this example, three areas stand out as potential containment BMPs that could be implemented quickly with relatively minor alterations. First, the existing detention basin could be plugged. Second, a small berm could be constructed on the downstream end of the conventional roadside ditch to convert it into a storage area. Third, an empty swimming pool within the contamination zone could be used for a containment/treatment area.

Next, it is important to place the volume of these storage areas into context relative to their respective drainage areas and the associated runoff volumes they might receive. That is, can the containment BMP store the total volume of stormwater runoff from its drainage area without overtopping? A conservative estimate of the maximum runoff volume for a given event is to multiply the drainage area by the precipitation depth (i.e. this approach assumes all rainfall turns into runoff with no infiltration or evapotranspiration losses). Rainfall depth for a given recurrence interval event can be looked up for any location in the US (NOAA, 2017). In this example, both the detention basin and the ditch storage areas have storage volumes large enough to store the maximum runoff volume from a single 1-yr storm event (Table 4); however, the swimming pool appears to be much too small to contain any sizable storm events. This implies that either additional storage BMPs are required or treatment BMPs are needed to treat and release the contaminated water during the rainfall events (the far-right portion of the decision tree in Fig. 1). In this example, let’s assume that a neighboring city has a weir skimmer that could be borrowed and installed in the swimming pool to treat the pool water prior to being discharged.

Table 4.

Compare the maximum runoff volume for a given event to the available storage volume of each BMP.

Location	Drainage Area (A)a	1-year Storm Rainfall Depth (D)b	Max Runoff Volume (Vw)c	Storage Volume (Vs)d	Vs > Vw?
3 – pool	8 acres	1.8 inches	52,272 ft3	1604 ft3	No
4 – detention basin	12 acres	1.8 inches	78,408 ft3	120,000 ft3	Yes
5 – ditch	1.5 acres	1.8 inches	9801 ft3	10,000 ft3	Yes

^aDrainage area in acres can be converted into ft² by multiplying by 43,560 ft²/acre.

^bRainfall depth can be looked up for any location in the US from this NOAA as a function of the recurrence interval (NOAA, 2017). This example uses the 1-yr, 24-hr storm.

^cThe max runoff volume (volume of water, V_w) in ft³ for a given event can be calculated by multiplying the drainage area (A) in ft^b by the rainfall depth (D) in ft.

^dStorage volumes for irregularly shaped pools and detention basins can be estimated by measuring the surface area at regular increments of depth to calculate incremental storage volumes and then summing them up. For example, 100 ft² x 1 ft of depth in the bottom foot of storage +120 ft² x 1 ft of depth in the second foot of storage + … + 200 ft² x 1 ft of depth in the sixth foot of storage = total storage volume. Storage volume in a uniform swale can be estimated by multiplying the cross-sectional area by the ditch length. In long and/or steep ditches, multiple berms may be necessary to account for the ditch slope.

Additionally, two portions of the contaminated area do not drain directly to the containment BMPs via gravity (areas 6 and 3 in Fig. 4b); therefore, pumps are needed to route the flows from these areas into one of the containment BMPs. Recall that the detention basin has the largest capacity relative to the runoff volume it is anticipated to receive (Table 4), so that could be the preferred pumping destination with adequate hose length. In this example, the agency has already deployed its three pumps for diversions. Let’s assume that a neighboring city has three additional pumps to lend to the hazard response effort, and enough hose length to route storm sewer #3 to the swimming pool and ditch #6 to an inlet that flows to the detention basin (Fig. 5). Note, capturing the total volume of stormwater is a conservative approach as some contaminants could be diluted and naturally attenuated for certain precipitation events. As more information is learned about the contaminant this may not be necessary for all rain events.

2.8. Identify BMPs required for long-term clean up duration

The longer the cleanup effort takes, the more likely it is to experience a large precipitation event. The risk of experiencing an event with an annual recurrence interval of “T_r” within n years is calculated as 1-(1-1/T_r)ⁿ. For example, if the clean-up period takes three years, there is a 27% chance of experiencing a 10-yr event, a 12% chance of incurring a 25-yr event, and a 3% chance for a 100-yr event. (For context, the midwestern United States’ experiences approximately 3.5 inches of rain for a 100-year recurrence interval, 1-h duration rain event (NOAA, 2017).) Although the initial diversion, containment, and treatment controls mentioned above should suffice for the first 7, 30, or even 60 days, managers should move to long-term/extreme event planning as quickly as possible, especially if the contamination occurs during the wet weather season. A more thorough assessment of the water table, local geology, and infiltration rates are also necessary to assess potential underground transport of the contamination and plan treatment solutions. In this example, pump systems are replaced by gravity flowing culverts and ditches wherever possible, mobile treatment systems, absorbent pads, and media filters are used to treat primary and secondary discharges, and an additional storage area is excavated to provide backup storage and tertiary treatment (Fig. 6).

Fig. 6.

Long term and extreme event control implementation.

3. Conclusions & future work

Contamination events require rapid decision making at the time of the incident to prevent the transport of hazardous material within the watershed and to downstream communities. Existing national emergency response frameworks provide a placeholder for stormwater utilities, but do not serve as a resource for the steps necessary to actively contribute to the response. The proposed SERF provides a structured approach to engaging local knowledge of a drainage system to optimize the use of available resources. While every incident will have unique constraints, through a coordinated response, risks to downstream waterways and populations may be minimized. Many open source software tools and databases exist to rapidly determine rough estimates of storm frequency and intensity, provide visualizations of local topography and flow paths, and assist in sample planning and analysis. This information may be coupled with simple calculations (e.g., Rational Method, storage volumes) to estimate peak runoff volumes and size containment and diversion technology. When possible, utilities should plan on having maps and digital elevation model data organized prior to an emergency so that the initial response is as effective as possible. Use of the SERF is an iterative process as more information becomes available about the contamination.

The longer the recovery phase, the higher chance the site will experience disruptive weather events that generate and/or transport contamination. If the BMPs initially installed fail to capture the full wet weather event, then an extension of the contaminated area may cause additional site remediation. Watershed and hydraulic models are indispensable for urban engineering design. The SERF can also be utilized to decide which more computationally intensive models are important to invest in to develop more robust containment and treatment strategies. These models are more frequently used for regulatory and flood control purposes than for simulating hazardous materials releases. The adaptation of watershed and hydraulic models to simulate CBR agents is currently an area of research requiring further exploration. Agent-specific parameterization and linkages to resource optimization software are necessary to accurately bound these issues. Future work also needs to focus on the development of training tools for stormwater utilities. Terrorism emergency response table-top exercises are frequent training tools for drinking water utilities, but a rarity amongst the stormwater community. This framework may serve as a resource for changing this paradigm.