Using Oil Spill Modeling in Oil Spill Exercises and Drills

Oil spill response exercises have been recognized by governments, industry, and potentially impacted stakeholders as necessary and important activities to plan and prepare for potential releases of hazardous substances. The effectiveness of oil spill preparedness and response is based on emergency organization procedures, trained personnel, oil spill response equipment, and logistical support. Training, exercises, and periodic evaluation of oil spill response plans (OSRP) can therefore familiarize personnel with an emergency response, identify weaknesses and deficiencies, and greatly improve these plans to ensure that they are effective guides in the event of an actual spill. Exercising an oil spill response contingency plan can therefore be used to provide assurance that an operator's oil spill response capability is managed, organized, assessed, and improved upon as needed.

The American Petroleum Institute's Guidelines for Oil Spill Response Training and Exercise Programs Guidance for Spill Management Teams and Oil Spill Responders summarizes the need for a robust and comprehensive training and exercise program as an important element of an organization's oil spill preparedness and response capability (API, 2014). In addition, the document summarizes the U.S. Federal regulatory landscape and associated requirements and references on oil spill response training and exercises (Table 1).

Here, we focus on the use of oil spill modeling to enhance Enbridge's operational response, response support, and realism for each developed release scenario used in the exercises and drills (as defined by API, 2014).

• An exercise is a structured and supervised activity used to develop or maintain fitness or increase skill. Exercises are an opportunity to validate an organization's oil spill response capabilities through simulated response to an oil spill scenario. Exercises can be used for testing and validating policies, plans, procedures, training, equipment, and third-party agreements; clarifying and training personnel in roles and responsibilities; improving team coordination and communications; improving individual performance; identifying gaps in resources; and identifying opportunities for improvement.

• A drill is a type of exercise. A drill is a focused and repetitious exercise, very disciplined, used as a means of teaching and perfecting specific response skills or procedures.

Oil spill modeling was used to determine the trajectory (i.e. movement) and fate (i.e. behavior and weathering) of released product within the environment from assumed full bore ruptures along each pipeline. Together, results provide a site-specific and season-specific understanding of the spatial extent, timing and magnitude of potential contamination. There are several key aspects of an OSRP that could be informed by computational modeling of releases. From the operational response standpoint, computational oil spill modeling can be used to assess the discharge, enhance the containment of discharge, target regions to maximize recovery of spilled material, and identify sensitive areas that may require protection. From a response support standpoint, this modeling can be used enhance communications within the ICS structure and with the public, support field operations, and in some cases inform procurement.

Oil spill trajectory and fate modeling was performed for two Enbridge Full Scale Exercises (FSE) to evaluate the spatial extent, timing, and magnitude of hypothetical unmitigated (i.e. no response actions) releases of crude oil into aquatic environments. During the fall of 2018, the exercise was conducted on the Des Plains River, near Joliet, IL approximately 30 miles southwest of Chicago. The river is navigable and contains several locks and dams, which included waterfalls that had the potential to result in entrained oil and could result in oil settling on the sediments (i.e. sinking oil). During the fall of 2019, another exercise was conducted on the Wisconsin River near Nekoosa, WI approximately 7 miles southwest of Wisconsin Rapids. This exercise included a hypothetical release of crude oil on land that was predicted to enter a small watercourse, before moving into a rapidly moving portion of the Wisconsin River water and then entering Petenwell Lake, an open-water environment.

The RPS OILMAPLand and SIMAP models were selected for these exercises to determine the overland and in water movement and behavior of crude oil following each hypothetical release. OILMAPLand is a two-dimensional modeling system that is used to simulate the movement of released oil in the environment (Figure 2; left). It simulates the flow of oil over land as it travels over the land surface and into a surface water body. The model itself has three components, including an overland release model, a surface water transport model, and an evaporative model that describes the weathering of oil in the environment under specified conditions. SIMAP is a three-dimensional modeling system that is used to simulate the physical fates of crude oil in the water (Figure 2; right). It estimates the distribution (as mass and concentrations) of whole oil and components of oil on the water surface, on shorelines, in the water column, in sediments, and evaporated to the atmosphere. This comprehensive SIMAP modeling system allows for a more in-depth understanding of the behavior of oil in the environment, when compared to OILMAPLand. Oil fate processes included in SIMAP are oil spreading (gravitational and by shearing), evaporation, transport, randomized dispersion, emulsification, natural entrainment, dissolution, volatilization of dissolved hydrocarbons from the surface water, adherence of oil droplets to suspended sediments, adsorption of soluble and sparingly-soluble aromatics to suspended sediments, sedimentation, and degradation. Both the OILMAPLand and SIMAP models are used extensively by industry and governments (French-McCay, 2004; Horn and French-McCay, 2015; Horn, 2017; Horn and Fontenault, 2018; Horn et al., 2018). Detailed descriptions of the algorithms and assumptions in the SIMAP model are in published papers (French McCay, 2002; 2003; 2004; 2009; French McCay et al. 2015; 2018). Hindcasts and modeling studies span marine, estuarine, and freshwater environments, including river cases and orimulsion studies (French et al. 1997; French McCay 2003; 2004; French McCay et al., 2005; Horn and French McCay, 2015; Stantec et al. 2015; 2017; Horn et al. 2017; French McCay et al. 2015; 2018a,b,c).

Site-specific and season-specific geographic and environmental parameters were gathered from federal, state, and local data sources to capture the conditions within the receiving environments at the time of each simulated exercise (Table 2). Several key parameters include elevation, land cover data, shoreline type, habitats, bathymetry, temperature, wind and currents (spatially and temporally variable speed and direction), and total suspended solids concentrations within the water column. In addition, the chemical and physical properties of the crude oil and the release volume are characterized. This includes composition by chemical constituent, as well as density, viscosity, volatility, solubility, and other key factors that will affect the way in which the oil behaves and weathers, once released into the environment. Unmitigated full bore rupture scenarios were simulated at each location for a total of 3 days for each exercise.

The Des Plains River exercise considered a 6,863 bbl release of Synthetic Sweet Blend (Syncrude) directly into the watercourse. The released oil was transported downstream and allowed to continue over the Dresden Island Lock and Dam, where it experienced a roughly 3-meter drop (spillway) followed by an approximately 100-meter section of rapids. This energetic region would be sufficient to result in the entrainment of oil into the water column, with the potential to enhance the likelihood of sunken oil. The effects of winds were de-coupled within the model implying that they would enhance evaporation and entrainment under high wind speeds, however they would not physical transport oil on the surface. This was done to minimize the amount of oil from stranding on the bank based upon winds and served to maximize potential downstream movement.

The Wisconsin River exercise considered an 11,219 bbl release of Cold Lake Winter Blend (diluted bitumen) onto the land surface (a field) approximately 1 kilometer to the east of the Wisconsin River. The released oil was predicted to be transported downhill to the point where it entered Sevenmile Creek, before flowing downstream through a wetland area into the Wisconsin River and ultimately Petenwell Lake. To enhance the number of response and recovery aspects being exercised, wind, currents, and weather conditions were selected to minimize weathering and maximize transport. Temperatures and wind speeds were selected to be 1 SD below the seasonal average for each location, while river flow was selected to be the 84^th percentile to rapidly transport oil to the lake. Weather conditions including a high-pressure system were simulated to allow for variable wind speed and direction. For the Wisconsin River exercise, winds were coupled into the model to allow for transport and fate. At the start of the simulation, winds were to the south, adding to the southern movement from currents. Between 12–24 hours, winds shifted to the west and ultimately to the east after 48 hours. This allowed for oil to strand on the western shores of Petenwell Lake and then be refloated across the lake to the eastern shores between 48–72 hours following the simulated release.

Oil spill modeling was used to identify several key aspects of each hypothetical release including the spatial extent, timing, and magnitude of hydrocarbon contamination predicted from each unmitigated release through time. Although modeled at 1-minute time steps, subsampled results of predicted surface oil thickness and shoreline/sediment oiling mass throughout the modeled area over the unmitigated 3-day simulation and provided as static images at 3-hourly intervals for the first 24 hours and 6 hourly intervals from 24–72 hours. Having quantitative information like this can help responders identify the best locations to implement response strategies and equipment as well as have an understanding of encounter rate of oil at each response location. For each simulation, a high-level response narrative was provided to include downstream locations of interest (e.g. road crossings, bridges, dams, sensitive receptors, etc.) and predicted visual appearance of surface oil thickness and shoreline oil mass. GIS shapefiles were also provided at the identified timesteps to feed into the common operating picture. This allowed for enhanced communications within the ICS structure, specifically between the Planning Section and Operations Section. In addition, the spatial spill modeling results allowed for an overlay type analysis to identify specific receptors that had the potential to be impacted. Oil spill model outputs served as “injects” within each exercise with locations of contamination as well as predicted surface oil thickness, shoreline oil mass, and amount on the sediments at different points in time throughout the exercises.

Presentations were prepared and presented to federal, state, and local regulators and stakeholders during planning meetings prior to each exercise. These meetings were aimed at aligning all parties to the exact scenarios that would be a part of the exercise and aided in the communication of framing the magnitude and extent of potential contamination following a release. In these meetings, the oil spill modeling tools to be used were identified and summarized. The inputs associated with the geographic and environmental conditions as well as other release parameters were discussed to ensure that scenarios met the needs of regulators (e.g. EPA and USCG), as well as the operator. Releases captured credible “worst case” scenarios that could adequately test the OSRP.

Simulated model results were provided in video formats to better capture the movement and behavior of oil through time and space within each simulation. Several snapshots of predicted surface oil thickness and shoreline oil mass have been provided for the Wisconsin River exercise to highlight the variability in the release (Figure 3). Within the first 6 hours of the release, oil is predicted to be more continuous and observed as heavy black oil. However, once oil reached Petenwell Lake, it formed a patchy and discontinuous slick that would be observed as black oil and heavy black sheen. As wind direction shifted to the west approximately 1 day into the release, oil was predicted to strand on the western shores of Petenwell Lake. However, as the simulated high-pressure zone moved to the east and wind directions shifted to be easterly, oil was predicted to be refloated as a patchy and discontinuous slick of light and dark brown sheen, before stranding on the eastern shores of Petenwell Lake.

These types of oil spill simulations can aide in table-top exercises as well as field operations to test different response strategies to an ever-changing environment. Results can also be used to enhance communication with field operations teams and in some cases aid in informing procurement or identifying equipment that may be needed at a later point. Oil spill responders frequently note that every release is different, and oil spill trajectory and fate modeling can be used to characterize the nuances associated with each new location, environmental and geographic conditions, and release scenario.

Understanding where oil is likely to move and collect is crucial to an informed response. Identifying natural collection areas can enhance the containment of the discharged material and can target regions that may maximize recovery. Predicted extents can also be used to identify sensitive areas that may require additional protection. Modeling results are also able to identify regions where oil may behave differently. One key example of this was surface oil (portrayed as black dots) moving over a spillway at the Dresden Island Lock and Dam (Figure 4). The turbulence of this roughly 3-meter drop resulted in surface oil becoming entrained (i.e. forced underwater; portrayed as blue dots), which results in high concentrations of total hydrocarbons in the water column. This was predicted approximately 1.7 days into the release for the Des Plains exercise. Spill modeling results identified locations where entrained oil may resurface, after the oil entered more quiescent regions downstream. In addition, they provided an understanding of areas that had the potential for sinking oil, the result of turbulent mixing of oil into the water column interaction with suspended sediments. Understanding the potential for sinking oil or specific locations that may become affected following a release can also inform procurement, response strategies, or the need to implement different plans, such as a submerged oil plan.

While these two examples clearly illustrate the benefits of using computational oil spill modeling to enhancing table-top exercises, drills, and full scale exercises, there may be questions regarding their use in other aspects of projects and operations. Many operators have chosen to run numerous simulations in high risk regions (e.g. higher relative probability of occurrence and/or potential for consequence) for contingency planning purposes. With weeks or months to focus on the problem at hand, a suite of scenarios may be developed to bound the range of releases that could occur with site- and season-specific variability over the course of a year or multiple years. These scenarios may be used to better understand the potential range of movement and behavior of oil following a release. This allows for more informed planning and preparedness, as operators may be able to better define the variability in potential spatial extent following a release, allowing for the identification of additional regions requiring control points and additional resources that may be required. However, as resources (i.e. time and money) are always finite, oil spill modeling cannot be expected to be conducted for every release possible (e.g. product, volume, location, season, etc.). This leads to the question of whether oil spill modeling can be used at the time of a release.

In the event of an actual release, computational spill modeling can and has been one of many tools that are used to inform effective response and recovery efforts. While boots on the ground, drones in the air, and eyes on the released product are crucial to understand what is happening at the specific time of observation, modeling is used by many to plan for the additional movement and behavior of the release to better position existing and incoming resources at future points in time.

There are multiple tiers or phases of spill modeling that can be conducted at the time of a release. As an example, 2-dimensional modeling (e.g. OILMAPLand) is best suited for inland releases and may be able to provide high level results within tens of minutes to guide early response efforts. Results may be based upon simplifying assumptions (e.g. average monthly flow rate for the watercourse) and generic inputs (e.g. uniform shore type). But, within 1–4 hours, additional refinement may be possible as observation data becomes available and other resources, such as real-time meteorological (e.g. National Weather Service) and hydrodynamic (e.g. United States Geological Survey gage data) conditions, are incorporated into the model. The more complex 3-dimensional modeling (e.g. SIMAP) can be applied to offshore releases and inland releases. In the offshore environment, where numerous meteorological and hydrodynamic data sets are available, comprehensive 3-dimensional modeling may be available in tens of minutes to a couple hours. As was the case during the Deepwater Horizon release in the Gulf of Mexico, numerous simulations were constantly being run throughout each day using different environmental and observational inputs and release parameters (e.g. as release volume was being refined) to better guide response activities. For inland releases, 3-dimensional modeling may require a minimum of one week in inland waterways and be a less viable option in the event of a release. This is due to the lack of available detailed data required (e.g. variable river depth and water velocity over large distances, etc.) for detailed modeling in inland environments. However, simplifying and/or conservative assumptions may be made to reduce turnaround time. Additionally, if an operator had already conducted modeling at the location (e.g. during contingency planning), an additional scenario may be available in tens of minutes to a several hours.

In recent years, comprehensive spill modeling has been included in a number of inland spill response drills and exercises. As a few examples Enbridge has used OILMAPLand and SIMAP in four full scale exercises in recent years including the Straits of Mackinac on Lakes Michigan and Superior, Cass Lake in Minnesota, the Des Plains River in Illinois, and the Wisconsin River in Wisconsin. Oil spill modeling results provide a comprehensive understanding of the location, timing, and magnitude of extents for each simulated release. These are site-specific and season-specific simulations that capture the nuance of each hypothetical release location and can better inform or test OSRP's. From an operational response standpoint, computational oil spill modeling of hypothetical releases has been used to inform several key aspects of an OSRP, including assessing the discharge, enhancing the containment discharged material, targeting regions to maximize recover of spilled material and identifying sensitive areas that may require protection. From a response support standpoint, modeling can be used to enhance communications within the ICS structure and with the public, support field operations, and in some cases inform equipment use and procurement. Regulators and stakeholders are becoming more attuned and informed when it comes to oil spill modeling, and although it is not required for full scale exercises is becoming an expected norm. While this paper has focused on the use of computation oil spill modeling in drills and exercises, it can also be used at the time of a release to better inform response activities.