National Academies Press: OpenBook

The Use of Dispersants in Marine Oil Spill Response (2019)

Chapter: 5 TOOLS FOR DECISION-MAKING

« Previous: 4 HUMAN HEALTH CONSIDERATIONS
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 153
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 154
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 155
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 156
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 157
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 158
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 159
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 160
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 161
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 162
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 163
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 164
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 165
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 166
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 167
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 168
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 169
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 170
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 171
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 172
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 173
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 174
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 175
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 176
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 177
Suggested Citation:"5 TOOLS FOR DECISION-MAKING." National Academies of Sciences, Engineering, and Medicine. 2019. The Use of Dispersants in Marine Oil Spill Response. Washington, DC: The National Academies Press. doi: 10.17226/25161.
×
Page 178

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

CHAPTER 5 TOOLS FOR DECISION-MAKING INTRODUCTION This chapter examines the response tradeoff decision-making tools that were introduced briefly in Chapter 1. These decision-making tools are often used to compare and assess the benefits and drawbacks of various oil spill response options, and ultimately rely on understanding of the fate of oil in the environment (Chapter 2), the toxicity of dispersants and dispersed oil (Chapter 3), and potential human health consequences (Chapter 4). The response options considered for marine spills include surface and/or subsea dispersant, mechanical recovery, in situ burning, shoreline cleanup, protective booming, and natural attenuation. Often these response options are used in combination to effectively mount a comprehensive spill response. “As the potential use of dispersants is expanded into nearshore, estuarine, and perhaps even freshwater systems, the trade-offs become even more complex” (NRC, 2005). While this is a true statement, the primary focus of dispersant use will continue to be for offshore oil spills in the marine environment where dilution can play a large part in mitigating potential negative effects. Although this chapter focuses on offshore marine environments, work has also been done on the use of dispersants in nearshore and coastal locations and it is possible that there will be special cases where dispersant use could be valuable. As discussed in Chapter 3, the toxicity of oil to marine organisms depends on the exposure of organisms to oil and the water-soluble compounds from the oil. The extent of toxicity depends on the exposure route, concentration of the oil, and the duration of exposure. Water-soluble chemical compounds like aromatic hydrocarbons (components of crude oils) are typically more highly toxic to marine organisms (Di Toro et al., 2007) than low-solubility compounds (e.g., most alkanes and cycloalkanes). When dispersants are applied to treat floating oil, small oil droplets are dispersed into the upper water column. Surface oils will rapidly dilute if water depth allows (Bejarano et al., 2013). Marine organisms living within a few meters of the sea surface will therefore experience an increased exposure to oil (Singer et al., 1998) with the toxicity depending on the exposure duration (Sterling et al., 2003; Bejarano, 2014). In the case of sub-surface spills like a blowout or pipeline leak, dispersants may be injected at the sea floor. This will increase oil concentrations near the source but tend to decrease them further afield, especially at the surface. Marine organisms in the lower water column will be exposed to an initial increase of water-soluble oil compounds, possibly reaching a concentration of 50 ppm, which may decrease to less than 1 ppm after a few hours as the compounds dilute in the water column over time (Lee et al., 2013). Dispersant application involves a trade-off between decreasing the risk to the surface and shoreline habitat and increasing the risk beneath the surface. The optimal trade-off must account for various factors, including the type of oil spilled; the spill volume; weather and sea state; water depth; degree of turbulence; and relative abundance and life stages of organisms (NRC, 2005). PREPUBLICATION COPY 153

154 The Use of Dispersants in Marine Oil Spill Response Chemical dispersants may increase the risk of toxicity to subsurface organisms by increasing bioavailability of the oil. The reader is referred to Chapter 3 for a detailed discussion of oil and dispersed oil toxicity. However, it is important to note that at the 1:20 dispersant-to-oil ratio recommended for use during response operations, the dispersants currently approved for use are far less acutely toxic than oil. Toxicity of chemically dispersed oil is primarily due to the oil itself and its enhanced bioavailability (Lee et al., 2015). As discussed, dispersants are considered a potential response tool in many countries. Different decision-making processes are used to determine whether to proceed with dispersant application, but typically, a list of dispersants that may be applied is developed along with an approval process for their actual use. See chapter 7 for further discussion. DECISION-MAKING TOOLS As previously discussed, making the best decision possible during an oil spill incident requires a balanced consideration of the potential environmental consequences of the spill under a natural recovery scenario versus the consequences associated with each response strategy. In any spill response, the first priority is the protection of human life, and the Federal On-Scene Coordinator and Area Contingency Plans place the highest priority on decisions that may affect response worker health and safety or public health (in the case of a nearshore release). Once immediate worker and public health and safety considerations are addressed, the next priority is to develop a response based on the best combination of response strategies that most effectively reduces environmental consequences, offers the greatest resource protection, or promotes faster recovery. However, determining the preferred response approach requires a time-sensitive evaluation of multiple factors (refer to Figure 5.1). Critical to informing response decisions is the identification of resources that are at risk to adverse effects. Special consideration is given to resources within the area of potential effects that are of socio-economic, ecological, cultural or archaeological significance, and in particular, to those resources that are protected under U.S. Federal laws or comparable regulatory requirements around the world. In the U.S., for example, these resources include species listed under the Endangered Species Act, designated Essential Fish Habitat under Section 305(b) of the Magnuson-Stevens Fishery Conservation and Management Act, and archaeological sites recognized under Section 106 of the National Historic Preservation Act. For ecological resources, tradeoff decisions were originally made based on key factors such as length of recovery from potential effects, but emphasis is shifting toward the use of ecosystem services –the goods and services supplied to humans by natural resources (see Ecosystem Services section below). In the case of length of recovery, habitats or animals that are anticipated to experience slow recoveries to baseline conditions are generally given greater protection, and thus, a greater weight in tradeoff decisions. Recovery rates are resource-specific but are usually expected to be longer for resources that have slow growth rates, long life spans and low reproductive output. With an emphasis placed on recovery, protection is generally assigned to entire populations, rather than individuals, with the exception of protected species. In the case of ecosystem services, emphasis is placed on the contributions of ecological systems to humans, PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 155 Figure 5.1 The decision-making process for selecting the optimal response option requires a thorough understanding of overall response goals and priorities, knowing what response options are available and feasible, where the oil is heading, and what resources will be potentially affected by the spill or spill response activities. and results in response option selection that is guided by restoration of the most valuable resources. In 2000, IPIECA (formerly the International Petroleum Industry Environmental Conservation Association) published a report entitled, Choosing Spill Response Options to Minimize Damage: Net Environmental Benefit Analysis (IPIECA, 2000). The report highlighted the importance of “…close cooperation between industry and national administrations … to ensure the maximum degree of coordination and understanding.” Further, the report emphasized that when all involved parties work together, there is greater likelihood of mitigating potential damages to the ecosystem. This report was the original guiding document for using the Net Environmental Benefit Analysis (NEBA) concept for oil spill response decision-making, and has since been revised (IPIECA- IOGP, 2015). While the understanding of the science of dispersant use has evolved in the past decade, there are still gaps. Therefore, NEBA-based approaches should clearly identify information gaps that exist at the time the process has commenced. In the past two decades, three decision-making tools have evolved to help implement the NEBA concept (see Box 5.1): • Consensus Ecological Risk Assessment (CERA) • Spill Impact Mitigation Assessment (SIMA) • Comparative Risk Assessment (CRA) These tools can bring together elements of various regulations, policies and current state-of-the- science into an overall decision-making framework for spill response. The tools are not limited to application in a particular regulatory regime or natural system (e.g, marine, estuarine, or freshwater environments), but can be adapted to a wide range of scenarios. PREPUBLICATION COPY

156 The Use of Dispersants in Marine Oil Spill Response Figure 5.2: SIMA, CERA, and CRA bring together elements of regulations, policies and current scientific information. This figure provides some examples for illustrative purposes. SOURCE: Modified from Coelho et al. (2017b). Integrated models play an important role in each of these decision-making tools, as explained later in this chapter. A basic representation of typical inputs to these decision-making tools is presented in Figure 5.2. INTEGRATED MODELS Several possibilities exist with respect to informing the oil spill response decision-making process as it pertains to possible and actual effects. One option to quantify dispersant trade-offs is to use laboratory or field experiments, although this approach faces steep challenges in replicating the complexity and individualist nature of a real spill. A second option may be to rely on experience from past spills, but these rarely include comprehensive, high quality observations and additionally, may be different from the spill being considered in important ways. Another option is to wait until the spill happens and monitor key indicators in the field during the event. This route has many potential pitfalls, one of the most important being the difficulty in getting sufficient monitoring equipment in the right place at the right time. Furthermore, real-time monitoring cannot itself forecast future effects and does not support the exploration of “what if” scenarios. Given these limitations, integrated models are routinely used to quantify the trade-offs involved in dispersant usage especially in contingency planning. As a tool to guide decision makers evaluating trade-offs, integrated models provide a number of advantages. Models incorporate many of the processes of importance and can provide a “big picture” view of the fates and effects of a spill in many different formats. From the perspective of PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 157 “fates,” integrated models have the ability to handle different oils, gas ratios, flow rates, as well as ocean and atmospheric conditions. Some models can factor in the removal or diversion of oil by skimmers, booming, burning, etc. From the perspective of “effects,” a few of the models use their calculated 4-D (x, y, z, and t) concentration fields of oil pseudo-components coupled with toxicity thresholds and spatial distributions of important biota to calculate mortality and recovery rates. Another benefit of integrated models is that they can be used to quantify and understand the sensitivity of the results to changes and uncertainties in inputs and sub-model formulations, and the effectiveness of the various response options. Sensitivity studies are especially helpful in establishing confidence limits and focusing future research on topics that will best improve our understanding of spill effects. Finally, a model can provide a forecast that reflects changes in weather, flow rate, and response alternatives. As briefly mentioned, it is difficult to make trade-off decisions using field observations during an on-going spill. The most important reason is that observations made during actual spills are almost always limited because the vast majority of efforts expended during a spill are focused on ensuring human safety, containing the oil (including source control), and minimizing the overall environmental damage. These activities will typically have priority over monitoring and this often results in restrictions on scientists trying to gain access to key resources or critical locations (e.g., the well-head in a blowout). Additionally, there are rarely opportunities during a real spill to conduct robust sensitivity studies. For example, during a blowout being treated with subsurface dispersant injection (SSDI), it would be informative to turn SSDI on and off under comparable conditions to assess its effectiveness. However, on-scene responders, as well as other stakeholders, may be unwilling to interrupt this response option if they believe it is reducing volatile organic compounds to safer levels. There were periods during the Deepwater Horizon (DWH) spill when SSDI was reduced substantially for operational reasons, however determining the effect of this action was difficult to assess because the interruptions were too short, the observations too sparse, and the environment too dynamic. Similarly, using results from lab experiments to guide trade-off decisions poses some limitations. First and foremost is the issue of complexity. The natural world contains countless processes that are difficult to simulate in a lab setting, effectively limiting the applicability of laboratory results to predict real-world outcomes. Another issue with relying solely on experimental results is the notion of “scale.” Even the largest lab facilities are typically orders of magnitude smaller than the environment in which an actual oil spill occurs. This constraint bounds the comprehensiveness of experimental testing, which has implications in the universality and accuracy of the models that are based on them. This is especially important in the study of critical processes such as deep ocean oil and gas droplet formation and evolution as oil travels from the subsurface, high-pressure environment to the surface. Large-scale field studies, such as DeepSpill (Johansen et al., 2003), have provided important information regarding the behavior of gas and oil released at depth. Like all studies, however, its scope was limited and in the case of this experiment, did not include an oil spill treatment process such as SSDI. It is unlikely that a future deep sea oil release field experiment involving SSDI would be permitted to allow shoreline oiling, although such a study design would provide valuable insights on the fate and effects of untreated oil versus subsea dispersed oil in this environment. If such a comprehensive study was successfully executed, the results would be best used to improve models rather than to expect that the experimental results could be directly applied to the next spill. PREPUBLICATION COPY

158 The Use of Dispersants in Marine Oil Spill Response Integrated models also have limitations. Some of the complex processes resident in the environment are poorly understood and their interactions may be even less so. Mathematically describing the behavior of oil in the environment through an integrated model and subsequently validating its results is difficult largely as a result of the lack of definitive observations taken during historical spills. On a positive note, however, integrated models are composed of sub- models simulating the major processes, and these sub-models have typically been validated. Nevertheless, results from even the best integrated model should be viewed with caution and results with uncertainty bounds should always be presented to decision makers, a point reinforced by ASTM F2067-13. Unfortunately, the ASTM does not provide guidelines on how to construct such bounds, which is not surprising given the complexity and lack of research in this topic with regard to integrated models. Errors and uncertainty in modeling stem from two general sources commonly referred to as aleatory and epistemic. Aleatory uncertainty originates from variability in key model inputs such as wind or current forcing, oil flow rate, etc. Quantifying some forms of aleatory uncertainty is fairly straightforward when historical observations are available and the model is run in a “hindcast” mode, an approach that is commonly used in strategic (contingency) planning by using Monte Carlo simulations. Aleatory uncertainty due to wind could be included in tactical (real-time) forecasts by utilizing the standard ensemble wind model forecast products, but this would require running the model 30 or more times. To the Committee’s knowledge this has not been done. Epistemic error arises from uncertainty in our understanding of the underlying physical, chemical, biological, etc. processes. Epistemic error is generally less well-studied and more difficult to quantify than aleatory error. Ideally, an integrated model could estimate the epistemic error by changing sub-model formulations or, at least, running sensitivity studies to understand the impact of uncertainty in the various sub-models on key model outputs. Similarly, an integrated modeler would ideally perform sensitivity studies to understand how changes in key model inputs (aleatory errors) affect important model metrics. French-McCay et al. (2018 b) has recently studied the impact of droplet size on fates, and this work shows that uncertainty in droplet size models can have a substantial impact on calculated effects. In other words, the uncertainty in calculated fates coming from the uncertainty in just one submodel (droplet size) can be substantial and could potentially alter the decision to use SSDI in a blow-out. Finally, it should be pointed out that integrated models often provide poor forecasts when used in an actual spill, in particular as the forecast time horizon increases (e.g., a forecast at 2 days will generally have a smaller error than a forecast at 1 week). It is not uncommon to see integrated models asked to provide a tactical forecast many days in advance even though the confidence bounds in the underlying weather and current forecasts are huge. A closer look at failed forecasts often reveals that there is nothing wrong with the integrated model; rather the problem is with the input of winds and/or currents which are generally derived from numerical models. As a case in point, Cooper et al. (2016) looked at operational current forecasts in the deepwater Gulf of Mexico based on state-of-the-art real-time measurements and models and found weak correlations, i.e., r2 values of only about 0.4 were obtained for a 2-day forecast horizon. Notably, a numerical model outperformed two seasoned experts. On the other hand, this same current model can provide very accurate hindcasts. As a result of this, it may be said with some confidence that the uncertainty bounds on integrated model results for strategic (contingency) planning will be much smaller than the bounds for tactical forecasts for most sites around the PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 159 world. In other words, the inability of an integrated model to accurately forecast spill fates days in advance during an actual spill should not be taken as proof that the model cannot be trusted to develop a reasonable contingency plan. Even if the uncertainty in integrated models was thoroughly studied and quantified, there would still be a major challenge: ensuring decision makers to take adequate account of this added information. The reality is that most major decisions in actual spills are not done with quantitative methods that could explicitly account for uncertainty statistics. Instead decisions are usually made by professional judgment with some attention paid to average (expected value) model results. In summary, optimizing the strategic (planning stage) and tactical (real-time) response to an oil spill scenario requires an understanding of the trade-offs of the response alternatives. It is doubtful that optimal decisions can be based solely on the existing observational database or previous experience. Thus, a well-validated and well-understood integrated model run by a knowledgeable operator can be an essential tool for the decision maker looking to choose an appropriate response strategy. The model results will have substantial uncertainty and it is desirable to quantify those, but the present state of doing so is rudimentary at best, and most decisions make limited use of the uncertainty estimates if they are provided. It is essential for end-users of these models (e.g., spill response decision-makers) to understand the limitations and errors within these models when using this type of tool to select incident-specific response options. The next section defines what we consider to be an integrated model. Some of the more commonly used integrated models are summarized with a discussion of their origin, scope, previous real-world applications, and validation. Two of those models will be used in Chapter 6 to evaluate the tradeoffs involved in dispersant use. Overview and Comparison of Integrated Models Figure 5.3, modified from French-McCay (2017), shows the major modules that would be involved in a complete oil spill modelling system which includes the four modules in the yellow ovals: 1. Blowout module (where appropriate), which would, at a minimum, calculate the droplet size distribution and simulate the buoyant plume. It could consider formation of hydrates and dirty (see Chapter 2) bubbles. The blowout module would be bypassed in the case of a surface oil release. Output from the blowout module would consist of 4-D snapshots of hydrocarbon pseudo-component concentrations that feed into the physical fate module. 2. Physical fate module, which would track the hydrocarbons until they reach their ultimate fate. It would include physical transformations such as surface spreading, dispersion, advection, and entrainment. It would also include weathering processes such as dissolution, evaporation, photooxidation, biodegradation, etc. The ultimate product of a fates module is 4-D snapshots of concentrations of the hydrocarbon pseudo components in the ocean. 3. Exposure module, which calculates the exposure duration of important biota to the 4-D concentration fields. PREPUBLICATION COPY

160 The Use of Dispersants in Marine Oil Spill Response 4. Toxicity module, which estimates the acute and chronic effects on the biota based on thresholds, toxic units, or some other metric. Integrated models require considerable input data which are indicated by the compartments colored blue, purple and magenta in Figure 5.3. The primary products from the modules of the integrated model are shown in orange. Table 5.1 summarizes several of the more widely used integrated models as well as Texas A&M’s TAMOC model. Part A of the table covers fates while part B covers effects. Each row describes a major process while each cell briefly describes the methodology. In the case of fates many of these are fairly standard submodels so the methodology description is limited to a few key words. Standard, well documented submodels are not as typical for effects, hence these are described in greater detail. Table 5.1 is intended as a summary of the models and does not capture many subtleties and may leave out some capabilities. A comprehensive review of all oil spill models is outside the scope of this report so the interested reader is referred to Bejarano et al. (2013) for a high-level overview of several other models or to the references in the second row of the table. SIMAP/OILMAP and OSCAR are commercial models that have been used to estimate both fates and effects. The NOAA GNOME/ADIOS2 model is the most comprehensive publicly available model. Historically, GNOME has been used by NOAA, other government agencies, and private Physical fates Biological effects Figure 5.3: Schematic showing the major components of an integrated oil spill model for the ocean. Tan blocks show the major modules of a complete integrated model. The blowout module would be bypassed in the case of a surface oil release. Blue, purple, and magenta blocks show the major external databases or models which must be provided to the integrated model. The orange blocks indicate major deliverables from the modeling. SOURCE: Deborah French-McCay. PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 161 PREPUBLICATION COPY

162 The Use of Dispersants in Marine Oil Spill Response PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 163 PREPUBLICATION COPY

164 The Use of Dispersants in Marine Oil Spill Response PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 165 companies to serve as the core model for CERAs and SIMAs for surface oil spill releases. GNOME does include levels of concern derived from consensus ERAs and tracks over space andtime whole oil concentrations on which to make assessments of potential effects (Mearns et al., 2001; Mearns et al., 2003). The TAMOC model focuses on the fate of oil from a sub-surface blowout to the sea surface covering module 1 and part of module 2 of the 4 modules listed earlier in this section. However, these modules may be the most critical in determining the fate of oil from a blowout. While TAMOC is more narrowly-focused than SIMAP/OILMAP, it uses more advanced algorithms for these processes. A closer look at Table 5.1A shows that none of the models directly include the coast, sea floor, or atmosphere. A truly complete integrated model of all oil fates and effects would incorporate these spatial regions. TAMOC represents an important class of models that can be used in conjunction with the more comprehensive models like SIMAP/OILMAP. The latter models are time-consuming to run, which inhibits their use as a tool in sensitivity studies or in studying new scenarios. Alternately, models like TAMOC can be more readily run for different water depths, oil types, dispersant-to- oil ratios, etc. Thus, one way to look at a new scenario (e.g., different water depth, distance from shore, oil type, gas-to-oil ratio, etc.) is to run the TAMOC model for the new scenario and then view the results from the perspective of SIMAP/OILMAP. Socolofsky et al. (2015) compared five integrated models for 14 scenarios of a continuous 20,000 bbl/day blowout. The models included SIMAP, OSCAR, and a predecessor of TAMOC. The scenarios considered two water depths (500 and 2,000 m), two gas-to-oil ratios (500 and 2,000 std ft3/bbl), two dispersant-to-oil ratios (0 and 2%), and two horizontal current regimes (5 and 30 cm/s). Because of the importance of droplet size, all the modelers used the same droplet distribution for several of the cases. In several other cases, the modelers used their preferred droplet model. Models were compared by looking at four metrics, the most important being the mass of oil entering the intrusion layer and the downstream distance to the center of surfacing oil. Important conclusions were: 1. There is a consensus of the models that the addition of subsea dispersant moves the surfacing oil downstream by an order of magnitude and results in far less oil reaching the surface. 2. For a dispersed oil, a decrease in droplet size of ~25% can increase the volume of oil in the intrusion layer by a similar amount, but causes much larger changes in the downstream surfacing distance (5x). This suggests that the present uncertainty in droplet models (up to 2x as described in Chapter 2) will affect the assessment of SSDI effectiveness substantially. There are considerable discrepancies between the models for many of the metrics and some critics of models have used this to question the credibility of all models. While the individual models have varying histories of development and validation, the two most validated and widely- used (OSCAR and OILMAP Deep) produce fairly consistent metrics when they use the same droplet sizes. For example, Fig. 11 of Socolofsky et al. (2015) shows that the distance from the release point to the downstream center of the surfacing oil compares to within 2x for the majority of cases where a common droplet size was used. PREPUBLICATION COPY

166 The Use of Dispersants in Marine Oil Spill Response THE CERA APPROACH The integration of the NEBA concept into oil spill response planning in the U.S. increased in the mid-1990s when the U.S. Coast Guard (USCG) developed a multi-agency approach to evaluate the ecological effects from various response options. The effort was spurred from an article on the application of ecological risk analysis in dispersant use (Aurand, 1995), which outlined the essential elements of Consensus Ecological Risk Assessment (CERA). The USCG document, titled “Developing Consensus Ecological Risk Assessments: Environmental Protection in Oil Spill Response Planning: A Guidebook,” was later published in 2000, after a four-year interagency development period (Aurand et al., 2000). While this USCG CERA approach is nearly two decades old, its recommendation for using a blend of both common sense and consensus-development, as well as quantifiable scientific information, remains a valid framework for response selection and continues to be used by industry and agencies. A detailed discussion of the CERA process was presented in the last NRC dispersant report (see NRC 2005, pp 35-45). The CERA process comprises three main phases: (I) problem formulation, (II) analysis, and (III) risk characterization, and is intended to be conducted in a workshop-setting that involves members of industry, operational response experts, response decision-makers, scientists, and local resource experts. CERA participation has typically involved between 25 to 50 workshop participants, but has varied. In Phase I, problem formulation, participants formulate a scenario for analysis, determine the relevant resources of concern and associated assessment thresholds, and develop a conceptual model that directs subsequent analysis. In Phase II, the analytical phase, the participants evaluate exposure, ecological effects, and recovery by customizing standard templates and simple analytical tools like the risk square and a risk ranking matrix, for the specific spill scenario under consideration. The risk square (Figure 5.4) was incorporated in the CERA process because of its common use in other types of environmental assessments by the US EPA (MMS, 1989). It is a means to examine environmental risk by considering both the percent of a resource that is affected by a perturbation (in this case, an oil spill), and the anticipated time of recovery for a given resource. In the CERA method, a 4x4 matrix is frequently used, but during a given CERA workshop, participants customize this risk matrix by adding more columns or rows to provide greater resolution on either axis. Additionally, the length of recovery can be defined during the CERA process for a given scenario, depending on the anticipated recovery time for local resources. In one CERA (Aurand and Coelho, 2003), the slowest recovery period was defined as “> 25 years” due to coral structures in the region, that if harmed could take a substantial amount of time to recover. Once the size of the matrix is set, and the increments on each axis are customized for a given CERA, participants then color-code the risk square to qualitatively assign levels of concern. The flexibility in the CERA process allows participants to modify levels of concern, based on local expert input and stakeholder engagement, for the spill scenario under consideration. This aids the evaluation process of response option comparison later in the analysis. The customization of templates during the analytical phase is an important step in developing consensus among CERA participants and helps ensure a common understanding of which resources are considered “important” in the local area, and on how each response option will be appropriately deployed. During the workshops, the participants then use the risk square to PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 167 Figure 5.4 A typical risk square used for a recent CERA (Walker et al., 2018). This example shows a 4x4 matrix, but the CERA method encourages participants to modify the risk square by adding columns or rows, as needed for a specific spill scenario under consideration. The participants can also customize the increments on either axis to account for local anticipated resource recovery times, based on inputs from local resource experts. In some CERA’s, the recovery period on the “SLOW” end has been > 25 years. The final step of preparing the risk square is to have participants color-code the risk square to qualitatively assign no adverse effect (green), limited (yellow), moderate (orange), or high (red) levels of concern. This risk square is then used in the next step of the CERA process to assign a severity affected (A through D) as well as a recovery time (1 through 4) to every resource category for the anticipated risk for each response option. SOURCE: Aurand and Coelho (2007). PREPUBLICATION COPY

168 The Use of Dispersants in Marine Oil Spill Response Habitat Intertidal NEARSHORE - Water Surface OFFSHORE - Water Surface Non-T/E Shellfish & Other Invertebrates Non-T/E Shellfish & Other Invertebrates Non-T/E Shellfish & Other Invertebrates Non-T/E Mammals - feral cat, mongoose Crustose Calcified Algae on the bottom Coral and Live Rock on the bottom T/E Species or Rare - ANIMALS T/E Species or Rare - ANIMALS T/E Species or Rare - ANIMALS T/E Species or Rare - PLANTS Vegetation, Floating Algae Crustose Coraline Algae Sponges on the bottom Resources of Concern Coral and Live Rock Non-T/E Mammals Non-T/E Mammals Critical Habitat Critical Habitat Critical Habitat Non-T/E Birds Non-T/E Birds Non-T/E Birds Non-T/E Fish Non-T/E Fish Non-T/E Fish Vegetation Plankton Sponges 3B 2A 2B N/A N/A N/A 2A 2A 2A N/A N/A N/A Natural Attenuation and Monitoring 3B 3B 4C 2B 3B 4D 3C 3C N/A N/A 3C 3C 4C 2C 3B 2B 4C 4C N/A 4C 2B 2B 4C 4C 1B 4B 4C N/A 2B 1A 2B N/A N/A 4B N/A 2C 3C Summary Risk for Sub-habitat 4D 4D N/A N/A 4D 4D N/A N/A Mechanical Containment and Recovery N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 3C N/A N/A N/A 4D 4D 4D 4D N/A 4D 4D 4D 4D 4D 4D 4D N/A 4C 4C N/A 4D 4D 4D Summary Risk for Sub-habitat 3C 2A 2C N/A N/A N/A 2B 2A 2A 3B N/A N/A N/A Chemical Dispersion 3B 3C 4B 2A 3C 4D 3B N/A N/A 3C 3B 3B 1B 3C 2B 4B 4B N/A 4B 2C 2B 4B 4C 1B 3B 4C 4C N/A 2B 1B 2B N/A N/A 4B N/A 2C 3B Summary Risk for Sub-habitat 3B N/A 4C N/A Resource Protection 3C 3C 4B 2B 3B N/A 4C 3C N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 4B 2B N/A N/A Summary Risk for Sub-habitat Shoreline Clean Up N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Summary Risk for Sub-habitat T/E species – ANIMALS Critical Habitat Reporting order: Reporting order: • Birds • Insects CH • Marine Mammal – HI Monk Seal • Plants CH • Marine Mammal – cetaceans • HI Monk Seal CH • Reptiles – Sea Turtles • Insular False Killer Whale • Fish - Manta Ray Figure 5.5 Example of a risk ranking matrix specific for resources inhabiting surface waters (0-2 m depth) in Hawaii, including threatened and endangered species, potentially affected by response options. For this CERA nearshore and offshore surface waters are defined as being within <1 mi and >1 mi from shore, respectively. The completed CERA delivers a final summary of results which color codes each option as no adverse effect (green), and limited (yellow), moderate (orange) or high (red) level of concern. The latter does not mean to stop actions, but rather to consult with resource managers on how to minimize impacts to the resources. Not applicable (blue) is also denoted. SOURCE: Walker, et al., 2018. evaluate the resource sub-categories for each response option, by assessing what percentage of each resource will likely be impacted and how long its recovery will take. In Phase III, risk characterization, participants compare the overall environmental risks and benefits of each response option to to those associated with natural recovery (i.e., baseline). The completed risk ranking matrix (Figure 5.5) is the key to the CERA analysis as it enables comparisons between response options and within particular habitats or resource groups. Figure 5.5 depicts conceptually how various resources of concern might respond when exposed to a response option. This figure depicts decisions from a recent CERA conducted in Hawaii, where CERA discussions were focused on potential surface dispersant application, and relied on GNOME for oil trajectory and oil budgets (Walker et al., 2018). In the CERA, it is recognized that, in addition to the stress caused by the spilled oil, each response option may also be a source of ecosystem stress. The mechanisms that cause the PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 169 Figure 5.6 A simple depiction of the steps in the CERA process. SOURCE: Coelho et al. (2014). stress vary as can the magnitude of stress resulting from each option. Seven “hazards” have been identifited and represent the potential exposure pathways that connect the stressors (including natural attenuation) to the resources of concern. In summary, the hazards include: “1. Air pollution; 2. Aquatic toxicity; 3. Physical trauma (i.e., mechanical impact from people, boats, etc.); 4. Oiling or smothering; 5. Thermal (i.e., heat exposure from ISB); 6. Oil-contaminated waste materials transfer and disposal; and 7. Indirect (refers to a secondary effect such as ingestion of contaminated food)” (Coelho et al., 2015). The CERA process uses natural attenuation (a.k.a., no response option) as the baseline for the analysis. In addition, CERA assesses “levels of concern” on resource categories, habitats, and population assemblages, not on impacts to individual species. However, some protected species may drive a decision regarding how to best protect a given resource or habitat. The USCG Guidelines (2000) provide a more detailed discussion of the CERA method, and the individual steps involved in conducting a CERA. A simple depiction is provided in Figure 5.6. Dozens of CERAs have been conducted for contingency planning purposes, and case studies are presented in the next chapter. Recent CERAs have been adapted to include socio-economic and human health factors. THE SIMA APPROACH After the Deepwater Horizon spill, there was a renewed effort to further refine the process of spill response decision-making that could be applied globally, in industrialized or remote areas. Although the USCG had already developed the CERA process for contingency planning purposes, many believed that it (e.g., time, cost and logistics) could not be realistically applied at the onset of a spill response, nor tactically during an ongoing event. Simply, the time, cost, and logistics coordination to achieve a consensus-based approach for contingency planning, while appropriate for contingency planning, was believed by many to be too constraining when faced PREPUBLICATION COPY

170 The Use of Dispersants in Marine Oil Spill Response with an actual spill. Acknowledging these limitations and recognizing that CERA may not be appropriate in other regions of the world, IPIECA-IOGP (2015) released an updated document titled, Response strategy development using net environmental benefit analysis. Although both IPIECA documents (2000 and 2015 versions) stressed the importance of making trade-off analyses of response options versus consequences, neither report presented a method to consistently apply a process. In 2016, IPIECA, the International Oil and Gas Producers (IOGP), and API worked together to develop a new method for studying risk in oil spill response that could address challenges in scoring and more readily facilitate stakeholder concurrence in past CERAs. The publication, Guidelines on Implementing Spill Impact Mitigation Assessment (IPIECA-API-IOGP, 2017) provided the strategy for analyzing oil spill effects and facilitating response option selection. As part of the refinement and communication process in developing SIMA, this framework was evaluated in workshops in North America (Clean Gulf Conference Workshop on SIMA organized by API/IOGP/IPIECA – October 21, 2016), Europe, and Asia-Pacific (IPIECA workshop in Perth Australia - Nov 30-Dec 1, 2017; IPIECA workshop in Singapore – December 5, 2017; AMSA workshop in Newcastle Australia – August 6-9, 2018). In addition, two SIMA’s were completed in exploration blocks in Eastern Canada in 2017 (Coelho et al., 2017a; Slaughter et al, 2017). Like CERA, the SIMA tool uses a structured framework for evaluating response options. It involves four steps: Step 1. Compile and evaluate data NEBA analysis considers the characteristics of the spilled oil and the transformations it may be subjected to as it weathers and spreads, which may determine the magnitude of environmental, biological and socioeconomic impacts (Daling et al., 2014). Data linked directly to planning scenarios under consideration primarily include oil properties, oil spill trajectory modelling, environmental sensitivity maps, and identification of appropriate response options for that particular site. Step 2. Predict outcomes The data obtained in stage one are reviewed and assessed by the planners and responders. Figure 5.7 summarizes the tasks in stage 2 and how it interacts with Step 1. SIMA includes an evaluation of the potential effect of a baseline scenario where no response actions are taken, which covers the timescale needed for the oil to be naturally attenuated (IPIECA-IOGP, 2015). The effects of the response options are characterized and evaluated after baseline establishment. Combined interactions of multiple response technologies at this stage must be considered. PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 171 Figure 5.7 NEBA Step 2 framework. SOURCE: IPIECA-IOGP (2015). During Step 2, the user develops a table of resources versus response options (see Figure 5.8). Figure 5.8 Overview of the Step 2 and Step 3. SOURCE: API, IPIECA, and IOGP (2017). PREPUBLICATION COPY

172 The Use of Dispersants in Marine Oil Spill Response Step 3. Balance trade-offs and reach consensus This step requires a range of stakeholders to reach consensus on the relative priority of the environmentally-sensitive resources and to understand the trade-offs associated with available response technologies (ASTM, 2013; IPIECA-IOGP, 2015). Two trade-off aspects are balanced in this step (i.e., protection and response and the benefits and drawbacks of selected response options). For the former, this priority may be influenced by the ease of protection and response; recovery times; and the importance for subsistence, economic value, and seasonal changes (IPIECA-IOGP, 2015). Step 4. Select the best options Following evaluation of all data, expert opinions and identified trade-offs, the final step is focused on the selection of the optimum response strategy for both the planning scenario and the prevailing spill conditions. Prior to a spill, response strategies may be identified for various planning scenarios. During a spill, the deployment and adjustment of response capacities may be needed; and after spills, the process supports the decisions about when response end points have been satisfied by continuing monitoring of response effectiveness and evolving conditions (IPIECA-IOGP, 2015). A visual framework depicting the pathways of decision making is provided in Figure 5.9. COMPARATIVE RISK ANALYSIS A CRA seeks to compare the benefits and consequences (effects/impacts on biota) of various response options. In many ways, it can be considered an evolutionary step of NEBA, one which takes advantage of recent advances in biological modeling technology to remove some of the subjectivity out of preceding frameworks. To date, there has only been one attempt at a CRA. It focused on a DWH-like blowout in the deep water of the Gulf of Mexico (French-McCay et al., 2018a; Bock et al., 2018; Walker et al., 2018) and was extended to examine the sensitivity of the fates to changes in site location (including depth) and droplet size (French-McCay et al., 2018). As a newly-developed framework, a key rationale of the CRA is that it attempts to reduce uncertainties introduced through the use of integrated models, whose predicted results may not reflect actual occurrences in the environment, by comparing the relative risks and benefits of various response options. At the core of a CRA is an integrated model that is capable of simulating both the fates and effects of a spill. In the case of the CRA study by French-McCay et al., the SIMAP/OILMAP DEEP models were used, but of course the methodology could be employed with other models such as the OSCAR/ERA-Acute model (Libre et al., 2018). The fate component of SIMAP has been around for several decades and has evolved in terms of complexity. The basic output from the fate component is a 4-D concentration map of hydrocarbon constituents. The 4-D fields of concentrations from the fate component is used to estimate hydrocarbon effects on important biota. Given our incomplete understanding of ecosystems, modeling biological effects even for a few important species, groups or habitats, is challenging. Perhaps the most difficult and potentially controversial task is the final step which involves weighting the relative importance of the species under consideration. Bock et al. (2018) PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 173 Figure 5.9 SIMA decision framework. SOURCE: IPIECA (2017). described a method they developed, which used elements of a SIMA/CERA. In the context of an SSDI application, they concluded that SSDI provided more benefits than costs and that the tradeoff analysis was insensitive to the weighting of their important species. In other words, the positive effects of using SSDI overpowered the negative consequences regardless of how one valued the biota affected by the spill. These results will be explored more thoroughly in Chapter 6 in the case of comparing SSDI to other response options. ECOSYSTEM SERVICES As described by NRC (2013), ecosystem services are the goods and services supplied to humans by natural resources. Examples of economically important ES in the Gulf of Mexico include commercial fish and flood control (due to wetlands). Less quantifiable ES include climate regulation and water purification. The idea of an ecosystem service analysis (ESA) was introduced in the early 1980s (Ehrlich and Mooney, 1983) and it has continued to slowly evolve. By far the most ambitious ESA was started by the United Nations in 2000 to evaluate present and future conditions of major ecosystems and estimate the consequences of ecosystem change to humans (Millennium Ecosystem Assessment, 2005). PREPUBLICATION COPY

174 The Use of Dispersants in Marine Oil Spill Response BOX 5.1 Summary of the Tradeoff Decision Tools Past CERAs and SIMAs have used modelling inputs from either OILMAP, OSCAR or GNOME/ADIOS 2 (depending on spill scenario) and both use a structured approach to seek consensus among stakeholders on effective response methods that produce the lowest environmental risk. There are differences between these tools and how they are best used in spill response. Consensus Ecological Risk Assessment CERA utilizes a detailed, semi-quantitative risk ranking square to perform comparative analyses of available response methods. The risk squares assign two scores, one for extent of exposure and a second for duration of recovery. CERA originally did not consider socio-economic or commercial factors, but was modified in 2012 to add these factors, and again in 2015 to add worker health and safety. A recent industry project in the Shelburne Basin of Eastern Canada utilized the CERA tool (Coelho et al, 2015), and the CERA method is also still actively used by USCG, US EPA and NOAA, including several recently conducted in Delaware Bay for light Bakken shale oil and diluted bitumen transportation (Walker et al, 2016) and Hawaii (Walker, et al., 2018). CERA is frequently used as a risk communication method that can add value by getting stakeholders together in workshops to re-evaluate and update Area Contingency Plans and Oil Spill Response Plans. It is also a mechanism for building trust among stakeholders and resource trustees by exchanging ideas and perceptions during non-spill response conditions. One drawback of CERA is that it requires considerable time and planning to get participants to workshops. As a result, it is an appropriate tool at the Contingency Planning level, but holds limited value as a real-time decision tool during a response, as there is insufficient time to execute this level of detailed scientific literature review during a spill. The reader is referred to the NOAA ORR website17 for the latest information on CERA workshops. Spill Impact Mitigation Assessment SIMA utilizes a risk ranking process that uses a single score for extent of exposure and duration of recovery and then adds a weighting factor for resource values based on local priorities. Unlike CERA, the SIMA approach front-loads the process of obtaining stakeholder consensus on resource priorities when assigning the weighting factor via information retrieved from the Environmental Impact Statement or other biological assessments conducted during the permitting process. As such, SIMA can be done quite quickly and is therefore a useful tool for use by the Incident Command System Environmental Unit at a Strategic Level during the early hours of a spill response to document current priorities and response decisions. Another 17 https://response.restoration.noaa.gov/oil-and-chemical-spills/oil-spills/resources/ecological-risk-assessment-era- workshops.html PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 175 advantage of SIMA is that it can be very quickly re-evaluated (re-scored) on a near daily basis to support ongoing changes in spill conditions at the Tactical Level. Multiple SIMAs have been conducted in the US and abroad since the tool was introduced in 2017, in support of permitting requirements, development of Oil Spill Response Plans, table top spill response exercises, and large-scale spill drills (Coelho et al., 2017a; Slaughter et al., 2017). Comparative Risk Assessment The CRA is the latest concept in response tradeoff tools and is different from SIMA/CERA in that it includes both a trajectory fate model and an effects model. This adaptation provides an objective way to quantify the effects of the spill, rather than relying on local stakeholders and experts to qualitatively score the fates effects. It also permits the user to weight certain resources above others (e.g., protected species). A CRA is scenario dependent and results will typically take many days of computer time to produce. Hence, it is a tool better suited to contingency planning than to tactical planning during an actual spill. ------------------------------------------------------------ Spill response decision-making cannot purely be simplified into “modelling and numbers”; it ultimately comes down to trying to satisfy a complex series of trade-offs and hopefully reaching consensus among the members of the Unified Area Command. Workshops involving stakeholders with access to fate and trajectory models are a key step to optimizing trade-offs and resolving inevitable conflicts and erroneous pre-conceptions. For example, in a CERA workshop conducted in Alaska, an examination of the trajectory model output suggested that aerial dispersant would have reduced shoreline oiling. However, in this region, the local stakeholders believed that oiling the shoreline was a preferred option to putting oil into the water column, as it aligned with their value system and key subsistence biota. This level of understanding about local priorities requires tools and methods that can be flexible, and all three of these tools incorporate mechanisms for weighting importance of locally important species or habitats. In summary, all three decision-making tools have value and can each be used in uniquely different ways to evaluate dispersants at the Area Contingency Plan level, strategic planning at spill onset by the Unified Area Command, or at the tactical level within Incident Command System Environmental Unit. The tradeoff decision and communication processes continue to be studied and refined. (Bostrom et al., 2015a; Bostrom et al., 2015b; Leschine et al., 2015) Applying an ecosystem service approach requires two major components: 1) reasonable models of important physical and biological systems for the region of interest and 2) economic models (what economists refer to as “production functions”) that convert the physical and biological effects into a monetary value. In its simplest conceptual form, the physical/biological models would track the consequence of an oil spill on the local ecosystem services (e.g., fish populations) and quantify those in terms of a 4-D series of some important metric (e.g., tons of fish). This output would factor into the production function to calculate an economic effect. PREPUBLICATION COPY

176 The Use of Dispersants in Marine Oil Spill Response NRC (2013) explored conceptually the use of ESA to estimate the damage of the DWH spill as an alternative to the traditional methods used in a National Resource Damage Assessment (NRDA). They concluded that ESA could theoretically improve the fairness of financial compensation to human victims and more efficiently guide restoration of the most valuable resources. That said, the report noted the many obstacles that inhibit ESA, notably the inability to accurately model the effect of an oil spill on important ecosystems, and to quantify those effects from a financial standpoint. In this context, it is apparent that the application of ES principles to the assessment tools (i.e., CERA, SIMA, and CRA as it matures), rather than using “length of recovery” of a particular species, is a natural next step in their evolution. While ESA may not be appropriate for response option analysis during small oil spills, it is reasonable to expect that an ESA might become a valuable tool for dispersant-use decision making at larger, offshore oil spills. FINDINGS AND RECOMMENDATIONS Finding: The objective of the NEBA process is to conduct an evaluation that will allow spill responders and stakeholders to evaluate the tradeoffs involved with the various response options and choose the option(s) that will result in a reduction of potential adverse impacts and/or the best overall recovery of the ecological, socio-economic and cultural resources of concern, while satisfying the primary goal of minimizing immediate risks to response workers and public health and safety. Recommendation: Decisions should be based on a balanced evaluation of consequences not driven by specific individuals, species, or economic interests. Recommendation: Greater efforts should be taken to expand and highlight the effects on human health and safety in the decision-making tools. Finding: All three decision-making tools (CERA, SIMA, CRA) have value and can be used in support of contingency plan development, strategic planning during the initial stages of a spill response, or tactically during the active phase of a response Recommendation: Decision-makers should further evaluate surface and subsea spill scenarios using NEBA tools (i.e., CERA, SIMA or CRA) to better define the range of conditions (e.g., oil type, sea state, depth, location, resources at risk) where dispersant use may be an appropriate and/or feasible response option for reducing floating oil. Recommendation: The NEBA tools (CERA, SIMA, and CRA) should be expanded to consistently address the health of response personnel, community health, and socioeconomic considerations (e.g., beach closures). Further, these tools should be used to gain stakeholder input on local or regional priorities, expand awareness, and gain trust in the decision-making process. Finding: The complexity of the interactions among fates and effects can be best addressed using numerical models. However, expert opinions used in the CERA and SIMA processes provide valuable insight to many tradeoff decisions, and this risk communication process allows for consensus by all stakeholders. Recommendation: Response decision-making should seek to become more quantitative to improve evaluation of the ecosystem services of the whole impacted region. PREPUBLICATION COPY

CHAPTER 5: TOOLS FOR DECISION-MAKING 177 Finding: The NEBA process is best achieved by using a blend of information provided by numerical models and stakeholder input. Finding: Integrated models that calculate the fates, as well as effects, of an oil spill are now available, and most of the submodels upon which they are based have been validated. Recommendation: A controlled field experiment or spills of opportunity should be used to collect comprehensive field observations for validating the entire integrated model. Recommendation: Integrated models should be used to evaluate and optimize combinations of response options. Finding: Integrated models are routinely used in tactical and strategic oil spill planning, usually with limited insight into their uncertainty bounds. Finding: It is important for end-users of numerical models to understand that even the best models have uncertainties. Recommendation: Systematic studies of the uncertainty bounds in integrated models are needed, and methods should be developed to include these bounds as a routine model product. Tools are also needed to help decision makers quantitatively account for this uncertainty in a consistent manner. PREPUBLICATION COPY

Next: 6 COMPARING RESPONSE OPTIONS »
The Use of Dispersants in Marine Oil Spill Response Get This Book
×
Buy Paperback | $75.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Whether the result of an oil well blowout, vessel collision or grounding, leaking pipeline, or other incident at sea, each marine oil spill will present unique circumstances and challenges. The oil type and properties, location, time of year, duration of spill, water depth, environmental conditions, affected biomes, potential human community impact, and available resources may vary significantly. Also, each spill may be governed by policy guidelines, such as those set forth in the National Response Plan, Regional Response Plans, or Area Contingency Plans. To respond effectively to the specific conditions presented during an oil spill, spill responders have used a variety of response options—including mechanical recovery of oil using skimmers and booms, in situ burning of oil, monitored natural attenuation of oil, and dispersion of oil by chemical dispersants. Because each response method has advantages and disadvantages, it is important to understand specific scenarios where a net benefit may be achieved by using a particular tool or combination of tools.

This report builds on two previous National Research Council reports on dispersant use to provide a current understanding of the state of science and to inform future marine oil spill response operations. The response to the 2010 Deepwater Horizon spill included an unprecedented use of dispersants via both surface application and subsea injection. The magnitude of the spill stimulated interest and funding for research on oil spill response, and dispersant use in particular. This study assesses the effects and efficacy of dispersants as an oil spill response tool and evaluates trade-offs associated with dispersant use.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!