NEBA is intended to provide the opportunity to explore unique regional features, identify sensitive species and habitats in order to rapidly develop an oil spill response strategy that will produce the least overall environmental impact should the occasion arise. The dynamic nature of the Arctic environment, including both physical and ecological features, taken together with known uncertainties, poses unique challenges to applying traditional NEBA concepts to oil spill contingency planning in the Arctic.  Oil exploration and production are currently increasing in the Arctic Circle (AMAP 2010) and as a result, the probability of oil spills from a variety of sources is also increasing.  Potential spill sources include exploration and production operations, vessels traffic, pipelines, and oil storage equipment. To date, there have been no large marine oil spills in the Arctic.  While natural oil seeps remain one of the larger sources of oil in the marine environment, most spills in the Arctic have been small and many are associated with fishing activities (AMAP 2007).  If NEBAs are to be conducted for oil spill response in the Arctic, they should address a range of potential oil spill scenarios, from small vessel refueling or operational releases to those scenarios considered to be “worst case” in terms of oil volumes, seasonal environmental sensitivities and weather conditions.

The fact that large regions of land and maritime waters in the Arctic are relatively untouched by anthropogenic influences has led to an emphasis on ecosystem consequence analysis (ECA) and response readiness based on seasonal windows as part of the development of Arctic-specific NEBA support framework.  The Arctic ecosystems are complex and also exhibit seasonal variability. Biological indices such as species/population migratory patterns and forecasted oil encounter rates, resilience, and recovery as well as relative environmental compartment health indices are key components in an Arctic NEBA.  Many birds and aquatic mammals are migratory, and their densities will vary widely geographically as well as seasonally.  Openings in ice may be recurring, in the same general areas, such as “polynyas”, or may occur at the edges of pack ice, but all are important to such species.  Thus their potential exposure during various oil spill scenarios will also vary widely. There are many uncertainties regarding the distribution, behavior, toxicological sensitivity and ecological vulnerability of many of these species that must be addressed as part of a comprehensive ECA undertaking.  Most terrestrial NEBA evaluations include consideration of potential restoration efforts as compensation for injuries determined during a damage assessment.   In contrast, it is likely that restoration efforts will be considerably more limited with respect to Arctic wildlife and aquatic resources, and therefore, environmental protection is paramount.  There are also uncertainties in our technical understandings of the influence of ice and extreme cold on the behavior of oil spills.  In addition, the extreme weather conditions and remoteness of locations in the Arctic pose challenges to deployment of emergency response resources.  Sea ice is perhaps the single most complicating factor in oil spill response because of its seasonal nature.  At various times, the same area may contain open water, broken ice, pack ice, etc.  Sea ice, depending on its nature, can trap oil under ice, funnel oil into fissures between ice packs, or retard the spreading and movement of oil slicks during slushy conditions.  Ironically, the winter conditions that produce enough ice to potentially preclude on-water response efforts, could also improve conditions for mobilizing and transporting spill response resources to a site.  During warmer months, tundra thaws partially, making overland travel difficult.  During winter months, it is possible to build ice roads that could be used to transport response resources to staging areas.  All of these factors have the ability to influence the effectiveness and acceptability of oil spill response strategies that are used successfully elsewhere around the world.

Assessment of the availability and effectiveness of spill response resources and strategies is a key component of NEBA.  The logistical challenges of deploying and supporting response resources in the Arctic will affect their utility during a response, and must be factored into the NEBA process.  For example, in Barrow, Alaska on the US Arctic Coast the temperature is below freezing approximately 320 days per year (wind chill makes this worse).  In Barrow, winds average 12 mph on an annual basis, but maximum daily average wind speed ranges from 44 mph to 35 mph in October and December respectively.  Extended periods of darkness could render water-based spill response activities challenging.  The sun does not rise above the horizon for six weeks out of the year.   There is little infrastructure available to support teams that would be assigned to large scale response actions (i.e. food, water, lodging).  In addition, there are no deep sea ports to host large equipment vessels, and only one paved road to provide access to the entire Alaskan Arctic coast.  There are several small coastal cities that have airstrips, but no paved roads connecting them.  Additionally, estimates of the effectiveness of available response resources will need to be adjusted for the seasonally variable weather.  For example, the effectiveness of skimmers in open water will be reduced by the presence of slush or pancake ice, and skimming is likely to be completely ineffective in heavier ice conditions whereas in-situ burning may be more effective in areas where ice prevents surface spreading.

Use of NEBA implicitly requires access to information on the composition and sensitivity of the potentially affected environment, information and case histories of the use of relevant response methods, and information on the likely impacts of using such response methods.  Because of the divergent seasonal conditions in the Arctic, use of NEBA for oil spill response assessments could require an analysis of each of the information categories described above, for each of the vastly different seasonal conditions.  Much of the existing practice of NEBA in the Arctic regions is based on use of sophisticated computer models that are used to understand the environmental transport and fate of spilled oil.  Governments and industry rely on advanced 3D-spreading model tools to simulate oil spill scenarios and to estimate (quantitatively) the exposure to natural resources both in the water column and on the sea surface and oil coming on shore. The results from analysis using such tools have shown to be very valuable input for the NEBA process.  There are other tools and approaches that could be used for evaluation of response options and equipment staging for response time optimization, but these have had limited review among spill response teams to date. 

Because NEBA remains a complex process, assimilating numerous parameters and extensive data, some countries such as France have developed a more expeditious alternative for one OSR, dispersant use.  The assessment outcome for dispersant considerations, while based on much of the information discussed above, is simplified into scenario based alternatives that define geographical limits where dispersant can be used without major threat for the environment. However, in coastal waters, where depth is limited, the rate of dilution of a dispersed plume is constrained.  These limits are based usually on depth limits for dispersant zones and distance to the shore (minimum depth and minimum distance to the shore) and are designed to ensure that the chemical dispersion can be undertaken far enough away from the sensitive coastal area to provide adequate dilution of dispersed oil and avoid potential environmental impacts. According to the different regulations, limits can be also defined according to spill size (e.g. France, three limits are defined for spills of 50, 500 and 5000 t. of oil to be dispersed).  The shallower the coastal zone is an area where many sensitive species and life stages are concentrated during some parts of the year so in some cases, the limits can be valid at certain times of the year to take into account migrations.  It is commonly accepted that in offshore situations, use of dispersants does not pose undue risk of harm to the environment due to the fact that, in open water there is a quick “dilution” of the dispersed plume.  Thus France and other countries have regulatory controls on dispersant use that strictly regulate the use of dispersants along their coastline.  Such limits, which can be considered as a simplified scenario based regulations are the outcome of ECA-like considerations, and at the time of a spill incident can provide direction to responders for a large number of possible scenarios.  For instance, scenario-based regulations a priori can help inform decisions on the applicability and benefits of dispersant use according to the net environmental benefits.

The geographic scope of this review is land and water masses north of the Arctic Circle.  This spans across eight nations, with differing cultures, languages, and regulatory approaches to spill prevention and response.  There are many indigenous peoples, with differing resource rights and needs, cultural value systems, and religious beliefs.  These socio-economic factors provide context for the application of consequence analyses and the outcomes regarding spill response strategies for contingency planning and emergency response decision making processes.  Once an oil spill has occurred, some level of negative environmental impact will result.  Emergency managers are responsible for selecting response strategies that will mitigate the negative environmental impacts to the maximum extent possible.  Strategies considered include the no response option, or natural attenuation.  The goal of the NEBA process in the Arctic is to select the response strategy that will minimize any long-term environmental impacts, thus producing the maximum net environmental benefit.  One of the first tasks that must be accomplished when a multi-stakeholder NEBA workgroup is formed is to determine the extent to which additional factors will be included in the analysis.  In the Arctic, several of these factors may play unusually significant roles in response strategy selection processes.  For example, the potential number of jurisdictions and stakeholders, both regulatory and indigenous, with interest in the Arctic may complicate the political, regulatory, and social dimensions.  Extreme and highly variable weather conditions may also affect the technological feasibility of certain response options that are widely used elsewhere.  In NEBA, the process of assigning the importance of each of these factors relative to ecological risk is a semi-quantitative process, and final outcomes are sensitive to the make-up of the NEBA workgroup.  The relative values assigned to different factors are likely to vary with geographical location.  For this reason, it is very important to make sure that all potential local stakeholders are included in the workgroup, or provided access to workgroup members.

Oil spill scenarios should be considered that span the range from worst case to most probable case.  Potential spill sources include exploration, development, production and transportation activities.  In the United States, spill size categories of minor, medium, and major are established in the National Contingency Plan, with threshold volumes of less than 40 m³ (10,000 gallons), 40 to 400 m³  (510,000 to 100,000 gallons)  and greater than 400 m³ (100,000 gallons), respectively.  Other organizations, such as International Petroleum Environmental Conservation Association (IPIECA) and the American Petroleum Institute use a three tiered, non-quantified classification system, where Tier 1 is the least environmentally significant or complicated oil spill event, and Tier 3 is the most significant.  Tier 1 spills can usually be responded to using local resources provided by the responsible party, whereas Tier 3 requires external response resources from sources such as national stockpiles or international cooperatives and significant involvement by local and national authorities. 

Various national regulations require plans that can be good sources of information on spill scenarios and associated response strategies and equipment.  Examples are the Facility Response Plans (FRP), Vessel Response Plans (VRP), and Oil Spill Response Plans (OSRP) required in the U.S.  When utilizing scenarios from plans required by regulations, it is important to ensure that the NEBA workgroup agrees that the scenarios are realistic, and address worst case spill situations in order to maintain the credibility of overall NEBA assessment.  Historical records can also be used to identify applicable spill scenarios.  In the U.S., Oil Spill Reports can be accessed through the National Response Center at www.nrc.uscg.mil.  Additional accidental release data sources are the DNV Worldwide Offshore Accident Database (Woad 2008) and the SINTEF Blowout and Well Release Characteristics and Frequencies database (SINTEF 2010).  In areas such as the Arctic, where oil exploration and production is increasing, and there have been few historical spills, the professional judgment of oil industry and regulatory emergency response professionals may need to be relied on to identify applicable scenarios.  However, across all nations, the historical record of spills from military vessels in the Arctic is likely to be insufficient and full reporting of releases from smaller fishing or shipping vessels may not be reliable. 

Estimating the fate and effect of oil from various spill scenarios can be accomplished by use of spill trajectory models, and applicable case histories.  Oil spill models commonly used include GNOME (NOAA), OILMAP and SIMAP (ASA), and OSCAR (SINTEF).  Use of these models requires historical weather data, ocean current data, and identification of likely oil types and their physical and chemical characteristics.  In the Arctic, trajectory modeling can be severely affected by sea ice. Use of computer models to predict oil spill trajectories under ice conditions is difficult, but there are coupled ice-ocean models that may be used to provide trajectory input for oil spill models.  Coupling sea ice forecasting models with oil spill models is an emerging area of science, and rapid progress is expected in the next 5-10 years. In many cases, combinations of best modeling practices, local knowledge and professional judgment may provide the best available information.

Once project-relevant planning scenarios have been developed, potentially applicable OSR strategies must be identified.  The OSRs that are typically evaluated include but may not be limited to:  1) natural attenuation (monitoring only), 2) mechanical containment and/or recovery using booms, skimmers, and sorbents; 3) in-situ burning using fire-proof containment boom, natural booming structures such as ice structures, or chemical herders; 4) oil dispersion using chemical dispersant formulations or application of oil-mineral aggregates, and 5) manual clean-up of oiled shorelines and debris.  Response strategies must be evaluated based not only on their anticipated effectiveness, collateral impacts, and availability as well as regulatory approval and technical feasibility.  Sea ice is likely to be an extremely limiting factor on the feasibility of some response options.  Mechanical recovery using booms and skimmers may be effective under open sea conditions, but its effectiveness decreases rapidly with increasing ice coverage, however the applicable response time may be longer than in open water conditions.  In-situ burning may be difficult in open water conditions due to wind and wave heights, but may be very effective in open areas between ice packs, or on solid ice surfaces.  Any response method that involves movement of large equipment overland may be difficult during summer months due to the lack of hard surfaced roads.  Manpower intensive operations may be hampered at any time due to lack of infrastructure to support large numbers of personnel (i.e. lodging, transportation, food, etc.).  Safety of response personnel must be considered in all scenarios, and waterborne operations will likely be restricted at times due to heavy weather and/or visibility constraints.  Aerial applications of dispersant may have significant limitations under certain icing conditions, but may in fact be the only response option that can be safely employed under other response condition.  Estimating the efficiency of any response method in ice conditions is complicated by the icing extent, seasonal factors, and geographic location; however, Evers et al. (2006) produced a chart of operational limits for various response techniques that may be useful for NEBA purposes (Table 9-1). This basic and widely distributed information has been updated based on findings from the SINTEF Oil in Ice JIP report regarding the potential for use of dispersants in ice (Lewis and Daling 2007, Sørstrøm et al. 2010, Faksness et al. 2011).  These studies evaluated new methodologies and strategies for dispersion of oil in high ice coverage (80-90%) which included maneuverable dispersant spray units and use of energy and turbulence to complete the dispersion process (see Figure 9-2; also discussed in Daling et al. 2010).

Table 9-1. Indication of expected operational limits of different response methods as a function of ice coverage.

Figure 9-2. Advanced technology for application of chemical dispersants in ice extends operational window.

Selection of response methods and resources for environmental conditions under which they  are unlikely to be available or not technically feasible will do little to aid in the development of effective or credible response plans, and could create unrealistic expectations for response capabilities.  On the other hand, if methods and resources are identified that could produce desired environmental and socioeconomic outcomes capturing such information could help guide future response capacity improvement efforts. Assessment of available and potentially effective response resources could be facilitated by the development of a database that would capture resource information and their capabilities under a range of arctic conditions from existing industry, local, regional, and national response plans.

Ecological resources that may be threatened by a spill incident must be identified for each scenario evaluated.  Factors that should be considered include the relative importance of resources from an ecological perspective [i.e. importance to food web or habitat structure including VECs, species of special concern (rare, threatened, endangered species), and importance to indigenous human populations (refer to Section 2 of the report for further information on Arctic ecosystems and VECs)].  Estimation of the sensitivity of various ecological resources to oil spills requires access to representative laboratory toxicity data, or information from field experiments or observations from previous spill events.  Some information is available through databases such as Naturbase, and Marine Resource Database - MRDB® (OBIS www.tos.org/oceanography/archive/ 13-3_grassle.pdf) and the Access database linked to Section 6 of the report.   It is also important to assess the rates and extent to which resources may recover naturally from spill impacts or may be amenable to or non-responsive to recovery processes.

The Coastal Response Research Center (CRRC) conducted a Natural Resource Damage Assessment (NRDA) workshop in 2010 to assess the current state of knowledge that could be used to support NRDA in the event of future oil spills, and identify knowledge gaps (CRRC 2010).  The workshop identified six different environmental communities and ecotypes for evaluation.  Those included birds, marine mammals, fish and invertebrates, ice and under-ice habitats, lagoon and nearshore shallow water habitats, and freshwater and coastal tundra.  The conceptual basis of Arctic NEBA under development is also focused on important functional attributes of environmental compartments (ECs), such as the surface microlayer, pelagic open water, convergence zones, riverine discharge areas, polynyas, etc.  Oil spills have the potential to cause acute and long term affects in all these areas, and also extent of oiling and its persistence may affect long term recovery rates.  Assessment of the magnitude of those dynamics, however, may be complicated by underlying changes driven by climate change, which also has the potential to cause long term environmental changes. 

The size and distributions of bird populations in the Arctic are very dynamic, and are influenced by characteristics such as coastal geomorphology, oceanic current patterns, seasonal migrations, and life stages.  An impact of a major spill would be damage to bird food supplies, and cause disruption of breeding grounds.  Information on the presence and movements of various species is available in literature, but less is known about the off-shore feeding habits of birds in the Arctic.  The CRRC workshop participants considered this a critical information gap, and also identified a need to better understand the impacts of oil spills on food sources, trends in health, productivity, and reproduction over time (CRRC 2010).

The abundance and distributions of marine mammals is also subject to seasonal variations.  Many species leave the Arctic during the winter, but some remain year-round.  Whales and seals typically breed and calf in the spring, and summer and fall are important times for feeding.  Many terrestrial mammals migrate to the Bering Sea during winter months, and return to the Arctic in the spring.  Those that overwinter typically spend the winter in dens or lairs.  Arctic mammals (marine and terrestrial) tend to be at the top of the food chain, and are valued highly by human populations for subsistence, tourism, and culture.  Like birds, the impacts of an oil spill will be highly dependent on the season, and location in which it occurs.  The CRRC workshop identified high quality population studies as an area of critical need, and suggested that an extensive synthesis of available literature should be conducted for marine mammals (CRRC 2010).

Fish and invertebrate populations are less variable although they can be highly patchy in their distributions, and may be less prone to drastic seasonal variations than birds and mammals.  The habitats likely to be the most sensitive to oil spills are coastal areas, under ice, and river deltas/lagoons.  Certain fish and invertebrate species have high ecological and commercial value in the Arctic, and are important for human subsistence and to the arctic food web.  As with biota in other parts of the marine environment, larval and juvenile stages of most arctic species are likely to be more sensitive to oil spills, and nearshore environments are likely to be particularly vulnerable.  The dynamics and distributions of fish and invertebrate populations are influenced by tide and ocean currents, depth, and sea floor and nearshore benthic conditions.  Environmental Sensitivity Index (ESI) atlases are available for some regions and can be an important source of information.

Ice and under ice habitats include multi-year ice, first year ice, land fast ice, bottom fast ice, ridges, level ice, melt ponds, below ice, snow, fall freeze ice , columnar ice, spring break ice, summer melt, break out ice, brine channels, and pack ice.  Each of these are important to differing degrees to a host of organisms ranging from those at the bottom of the food chain such as algae and amphipods, to those at the top such as seals, walruses, and polar bears. Openings in ice are particularly important to many species, and could also potentially serve as collecting areas for oil spills.   Many seabirds congregate to feed at ice edges, polynyas and open leads, where their prey species congregate (AMAP 2010). During spring migration, the entire population of bowhead whales in the Bering, Chukchi and Beaufort Seas (BCB population) travels north through the U.S. Arctic Ocean ice lead and polynya system, and their path is relatively constrained (AMAP 2010). This migration path is also the primary calving area for this population of whales.  Oil spills onto, or in these habitats could cause acute and chronic toxicity among exposed individuals, leading to shifts in population structures.  Ice related human uses such as hunting and subsistence fishing, and travel could also be affected by an oil spill.  Factors that could influence the magnitude of these impacts include use of dispersants and in-situ burning, solubility of oil, amount of mixing between dispersants and oil, movement of oil within brine channels, melting rates of the ice, and the amount of affected ice that does not melt each year.  Factors affecting oil on ice surfaces include evaporation, biodegradation, and the melting rates of ice. 

Lagoons and nearshore shallow water habitats are likely to be of seasonal importance, with relatively high biological productivity in the summer months, and low productivity in the winter.  Lagoons can provide key ecological services such as shelter, nursery habitat, and feeding grounds for many species of mammals, birds, fish and invertebrates, and are also important to human populations for subsistence hunting and fishing.  Sources of information on the distribution and complexity of these habitats in the U.S. include the Alaska Resource Library Information System (ARLIS) and Alaska Environmental Sensitivity Index (ESI) maps.

Freshwater and coastal tundra environments are characterized by periods of seasonal freezing and thawing.  The peat layer is biologically active, with high productivity in summer months, and overlays a mineral layer above the permafrost.  Tundra may also be interlaced with freshwater ponds.  Most Arctic rivers are shallow, freeze during the winter months, and then experience high flow rates for short durations during the spring breakup.  Tundra is very sensitive to physical disturbance due to the short growing season, and recovery times can exceed 20 years.  Tundra is a high value bird habitat, and tundra and freshwater areas support polar bears, grizzly bears, caribou, fish, and birds.  Caribou and many of the fish species present are important subsistence resources for human populations.  Because a number of birds and mammals may use both coastal and inland freshwater habitats for feeding, oil contamination in coastal areas may get transported to inland areas (or some sort of text that links how oil in the coast can get to these more inland areas)

Industry has collected a large amount of data on the environmental characteristics and ecology in this region and synthesizing it is a crucial next step. A website (www.northslope.org) has been created that contains information on topics including: migratory birds, marine mammals, and their prey; increasing marine activity; permafrost; coastal and riverine erosion; contaminants; fire; vegetation change; caribou; tundra rehabilitation; and Arctic fish. The Audubon Society and Oceana's Arctic Marine Synthesis compiles public datasets. These data are spotty and not always comparable from site to site. A framework is needed to facilitate the sharing of quality data between industry and the stakeholders.

Large oil spills have the potential to significantly affect social, psychological, health, and economic parameters in the impacted region.  Subsistence hunting and fishing is extremely important to indigenous peoples in the Arctic, and in addition, many of the species present are assigned high cultural or spiritual value.  The Arctic area may be particularly sensitive to oil spill impacts due to relatively low density of human populations (which limits the ability to deploy large numbers of responders immediately), compressed growing, hunting, and fishing seasons, and potentially slower ecosystem recovery rates.  Even when populations of species that are important for subsistence do recover, indigenous peoples may be reluctant to use them due to actual or perceived “tainting” of those resources.  There have been no large oil spills in the Arctic that have significantly affected human populations, so for NEBA purposes, potential effects on social issues must be estimated based on local knowledge and expertise, or extrapolated from case histories of large spills in other regions.

To overcome some of these difficulties, for example, the methodology proposed in the revised version of the International Maritime Organization (IMO) guidelines on dispersants involves 4 steps for completing NEBA:

1. Determination of the drift of the dispersed and undispersed oil (using the meteorological-oceanic data, wind and current);
2. Identification of the sensitive resources of concern using drift calculations and the previously compiled National Oil Spill Contingency Plan (NOSCP) sensitive resources inventory;
3. Among these, identification of the highly vulnerable resources using the NOSCP data on vulnerability;
4. Among the identified vulnerable resources, determination of the resources which should be preserved as a priority, using the priority list previously defined in the NOSCP. The decision of whether or not to use chemical dispersion is made depending on which is the appropriate to these priorities.

This approach allows concentrating the difficult discussions between the stakeholders during the final step, whereas the first 3 steps can be completed with technical objectivity (as they are based on scientific data).  Implementing NEBA in step 4 requires considerations that include subjective best professional judgement that takes into account not only environmental considerations but also political, cultural, etc.

The decision tree used to decide dispersants use is shown in Figure 9-3 (based on the revised IMO guidelines on dispersants, 2013). The recommended approach first identifies geographical limits for dispersant use based on internationally accepted guidance (e.g., France), and considers completing a more detailed or site-specific NEBA if there are no geographical limits or if the spill is inside the geographical limits.  At this point, IMO recommends completing the NEBA according to the 4 steps outlined above, using step 4 to assign local priority.

The NEBA rationale has been widely used to assess spill response strategy effectiveness, identify potentially impacted areas, and estimate associated environmental and socioeconomic impacts during many phases of environmental and emergency management. These include permitting and regulatory actions, contingency planning, and emergency response.  Lunel and Baker (IOSC 1999) illustrated varying uses of NEBA to support strategic, tactical, and operational decision making.   SINTEF has demonstrated phased approaches to NEBA use, during planning and response, noting that “NEBA… is a continuous evaluation process that is to be repeated during an incident in the light of new information concerning the behaviour of spilt oil, the overall environmental impact, and/or the effectiveness of the activated response technique ” (Schallier et al. 2004).  In Russia, NEBA is required by regulation as a condition for dispersant use.  Norway requires use of the similar NEDRA process for oil spill response planning. While not a formalized legal requirement in the U.S., NEBA has been used in Area Contingency Planning, and in some cases has resulted in pre-authorization of dispersant use by Federal On-Scene Coordinators for certain spill types and locations (Addassi et al. 2005).  The following description of NEBA concepts, historical uses and case studies illustrate some of the potential benefits of using NEBA for contingency planning and emergency response to oil spills in the Arctic environment.

Figure 9- 3. Decision tree illustrating the NEBA process [Source: Merlin (CEDRE) and Lee (COOGER)]

A number of relatively objective analytical tools have been developed to support the process of estimating the effectiveness of potential response strategies and the fate and effect of spilled oil.  Reviews of field trials and/or case histories are also available, which help to serve as a basis for decision making, and for verifying model outputs.  

Lunel and Baker (IOSC 1999) reviewed case histories of well-studied oil spills, including the Braer (1993), Exxon Valdez (1989) and Sea Empress (1996) and developed tables that could be used to produce quantitative estimates of oil fate and effect for NEBA analyses (including ecological and socioeconomic sensitivity tables).  They demonstrated the potential use of historical data for three levels of NEBA – strategic, tactical, and operational, which corresponded roughly to tier 3, 2, and 1 spill respectively. 

French and Shuttenberg (1999) demonstrated the ability to use the Spill Impact Model Analysis Package (SIMAP) to predict impacts of a scenario based on the well-studied North Cape Oil Spill.  In the North Cape spill, oil dispersion occurred as a result of high wave energy, not chemical dispersants, but the model was capable of predicting either.  The SIMAP model includes an oil physical fates model; a biological effects model; input tools for oil physical, chemical and toxicological data; input tools for environmental, geographical, and biological data; a response module to analyze effects of response strategies, and export and graphical visualization tools. 

SINTEF uses the Oil Spill Contingency and Response (OSCAR) mathematical model to support NEBA activities.  OSCAR consists of a three-dimensional numerical model of the physical and chemical behavior and fate of spilled oil, and an oil spill response simulator for various mechanical recovery and dispersant application systems.  Prediction of oil spill fate and effect is particularly difficult in ice infested waters, but the OSCAR model showed promise during field trials conducted in the Barents Sea in 2009. Further modeling research at SINTEF is underway to develop coupled ice-ocean-oil models.  With regards to modeling ecological impact, by implementing models of biological resources (statistical distribution in time and space) into the OSCAR model system, it is also possible to perform dynamic modelling of the oil exposure to relevant marine organisms in the water column (e.g. fish eggs and larvae) and sea birds on the surface.  This has become an important tool for predicting realistic effects (acute toxicity) and losses of populations of analysing various oil spill scenarios/response options. 

The use of any of these tools to support use of NEBA in contingency planning requires that realistic scenarios are utilized and that response methods evaluated are actually available and feasible for the area being studied.

Assessing the potential ecological and socioeconomic impacts and recovery rates is inherently more subjective.  Natural resource distributions are particularly sensitive to temporal and geographic influences.  Socioeconomic impacts result from damages to natural resources, and may include loss of subsistence hunting and fishing, reduction in tourism, and reduced “sense of well-being” due to perceived environmental tainting.  The recovery rates from these types of impacts are dependent upon many factors that are highly location dependent and difficult to quantify.  As a result, it is often necessary to rely on local knowledge and expertise to predict the relative magnitude of these variables during NEBA studies.  In doing so, it is extremely important that all potentially affected stakeholders are identified and provided opportunities for involvement.  For NEBA results to be successfully used in contingency planning, stakeholders must reach consensus on the magnitude and relative importance of potential ecological and socioeconomic impacts for the range of spill scenarios considered.  This can be extremely difficult since the value assigned to environmental resources is likely to vary widely, depending on the extent of use or value of those resources by different stakeholder groups.  Even if individuals representing varying stakeholder groups reach consensus during the NEBA process, other members of the community may not understand and accept the decisions reached and support spill response decisions made during an incident.  Broad acceptability requires an effective outreach and communication strategy and frequent re-evaluations in order to determine any changes over time in response capabilities and technologies or resource dynamics and valuations.

Despite the semi-quantitative nature of resource valuations, some tools have been developed to assist with more objective analysis.  Some of the previously discussed models, such as SIMAP and OSCAR do contain biological assessment algorithms.  Aps et al. (2007, 2009) demonstrated that Bayesian inference networks, which capture uncertainties in terms of probabilities, can be very useful in supporting ecosystem consequence analyses by providing decision makers with more objective numerical estimates to weigh alternatives against each other. Bayesian networks were shown to be useful in integrating surveillance data, mathematical simulation results, and ecological sensitivity GIS maps during ECA analyses (Aps et al. 2007, 2009).

NEBA can be an effective means of generating valuable discussions around the potentially disparate views of industry, academia, government regulators, and local stakeholders, even if consensus is not reached.  This was illustrated in 2011 during a Workshop on Dispersant Use in the Canadian Beaufort Sea that was conducted as part of a multi-year Beaufort Regional Environmental Assessment (BREA).  The workshop included simplified NEBAs addressing the potential benefits of dispersant use in three different scenarios.  Over 50 persons from stakeholders including Inuvialuit communities, government agencies, and the oil industry were involved.  The participants did not necessarily reach consensus on the desirability of dispersant use, but each of the major stakeholder groups did present valuable perspectives on the path forward for future planning activities.

The literature contains many examples of accidental and intentional research related spills in non-Arctic areas that have produced information that is relative to NEBA assessments.  Spills conducted for research purposes have allowed for observation of response strategy impacts and long term recovery rates in controlled environments.  In some instances, NEBA has been used to guide response actions during actual spills, and its benefits have been well documented.  Examples are provided below.

The Baffin Island Oil Spill (BIOS) field-based experiment examined the acute and long-term effects of crude oil and dispersed oil released to separate embayments of Baffin Island in 1980.   The untreated oil dispersion simulated an oil slick moving onshore from a blowout situation (producing oil at 1000 m³/d or approximately 15 m³); chemically dispersed oil concentrations were based on typical oil concentrations beneath dispersed oil slicks (averaging 10 µg/g after dispersant application).  Faunal groups assessed included infauna, epibenthos, macroalgae, and under-ice meiofauna in the vicinity of nearshore, landfast ice; endpoints included mortality, growth and community metrics (e.g. species abundance and diversity).  An additional bay was observed for the same endpoints to serve as a reference site.  Inter-annual patterns were established for the reference bay and the authors concluded that only minimal effects could be attributed to oil.  Some short-term behavioral patterns and a reduction in urchin densities (due to emigration) was observed as a consequence of dispersed oil exposure; oil had a possible minor effect on the reproductive viability of Spio sp. of polychaetes.  No adverse effects to biomass, number of species and reproductive condition of macroalgae were noted.  Observations of the under-ice communities noted some evidence of a stimulatory effect of oil on harpacticoid adults and copepodites, and on cyclopoid nauplii, whereas a dispersed oil exposure maintained for 5 h indicated that harpacticoid adult copepods showed greater immediate response than did copepodites.  The community structure at the reference site was composed of: copepods = 92.3%; nematodes= 4.0%; and polychaetes = 3.2%.  Natural weathering of the oil was observed periodically for twenty years.  By 1989, there was approximately an 80% reduction in oiled area; further reductions occurred by 2001 and lingering oil was less than 5%.  Much of that reduction was attributed to photo-oxidation and biodegradation (Cross WE 1987, Cross and Thomson 1987, Cross et al. 1987a, Cross et al. 1987b).

The Tropical Investigations in Coastal Systems (TROPICS) study began in 1983/84 near Bocas del Toro, Panama.  The study involved dosing of two sites, containing intertidal mangroves and subtidal seagrass-coral zones, with both dispersed and non-dispersed oil.  The treated areas were monitored periodically for 25 years.  The study found that plots treated with dispersed oil recovered to near reference level conditions within 10 years, whereas sites treated with undispersed oil exhibited long term disruption after 25 years (DeMicco et al. 2011).  The results provided useful scientific data that can support future NEBA projects in tropical, nearshore environments.

In 1993, during heavy weather conditions (Beaufort force 10 and 11 winds, 55 – 74mph), the tanker Braer ran aground in the Shetland Islands, and released 84,700 tons of crude oil.  During the early and late stages of the spill response, chemical dispersants were applied from DC-3 aircraft but the winds were prohibitively strong during most of the response.  All of the oil eventually dispersed but it was estimated that chemical dispersion accounted for only 2 – 3% of the total spill volume, with natural dispersion accounting for the rest due to the high sea state and the specific oil type (Gullfaks crude).  A comprehensive and prolonged monitoring program following the incident found that dispersed oil concentrations were high (50 ppm) in the vicinity of the wreck, and persisted for several days.  Ten days after the incident, dispersed oil concentrations were 5 ppm.  Background levels were found after 60 – 70 days.  Fisheries were very important to the local economy, and while some impacts did occur, they were localized, and far less than expected.  The incident provided a comparison of the relative expected toxicity of dispersed vs. non-dispersed oil.

The Sea Empress ran aground in 1996 in the entrance to Milford Haven, in Pembrokeshire, U.K., spilling 72,000 tons of light crude oil and 480 tons of heavy fuel oil.  The response action included spraying 446 tons of dispersant from 6 DC-3 aircraft.  Approximately one half to two thirds of the 37,000 tons of oil that was estimated to have dispersed was attributed to the addition of chemical dispersants.  While environmental impacts were significant, they were less than was expected based on the size of the spill.  Tradeoffs of environmental impacts associated with the use of chemical dispersants were identified and taken into account for the nearshore use of dispersants.  As a result, the overall environmental impact of the spill was thought to have been reduced.

The Montara oil spill (2009) resulted from a well blowout located approximately 180 km west of Australia, in the Timor Sea.  Estimates of the spill release rate ranged from 400 to 2000 bbl per day, and the spill lasted a total of 74 days.  Based on computer simulation, using the OILMAP model, it was shown that the spill had the potential to impact sensitive areas such as Ashmore Reef, Cartier Island, and Western Australian wetlands.  A NEBA was conducted to evaluate the potential impact of response strategies including mechanical recovery and use of chemical dispersants.  The NEBA considered the demonstrated benefits of use of chemical dispersants as a first response option for the Sea Empress Spill.  The results of the NEBA concluded that the potential for surface oil impacts would be significantly reduced by the use of chemical dispersants.  Consequently, provisions for spraying dispersants from a vessel onto fresh oil (less than two days old) were incorporated into the Incident Action Plan (IAP).  The NEBA further recommended use of mechanical recovery and natural recovery for certain situations where dispersant use would not result in a net environmental benefit.

The Arctic environment poses significant challenges to the application of a standard NEBA assessment.  Chief among the unique challenges is assured protection of essentially pristine ecoregions and the understanding of seasonal and inter-annual changes that determine residence and reproductive patterns of the faunal groups, such as extreme weather conditions, seasonal variation in the amount of ice cover, extended periods of darkness, and seasonal variations in the presence and sensitivity of fish, birds, and marine mammals.

Extreme weather conditions complicate development of an optimal OSR strategy because they influence the ability to deploy and sustain emergency response strategies.  During extreme weather conditions, or long periods of darkness, some response strategies may be less effective, or implementation of some methods may be hampered due to logistical and safety considerations.  Weather conditions greatly influence the ability to move and deploy response equipment.  In winter months, with heavy ice cover, it is difficult to deploy ship-based strategies such as skimmers, but it could facilitate land based resource deployment on “ice roads”.  In spring and summer months, land based transportation could be difficult because of thawing.  The presence of ice can be a positive influence in retarding oil spill movement, and can improve the effectiveness of in-situ burn.  If oil spills become trapped in ice openings such as polynyas or leads, in-situ burning can be very effective, provided that ignition systems can be deployed to the spill site and the oil is still ignitable. Because weathering processes are slower in ice, the “operational time window" for dispersant use in ice covered water may be extended compared to open water conditions. Some oils spilled in ice remain dispersible for several days after release (Brandvik et al. 2010)

Many organisms that have high ecological, socioeconomic, and cultural value, such as marine mammals are migratory and their abundance in a given area varies widely with the seasons.  Marine mammals such as whales depend on ice openings for air.  A spill that persisted in ice openings during migration periods could expose numerous individuals in the area over a period of time that could pose concerns for population altering effects.  Shoreline habitats, particularly during spring thaws are important for spawning and migration of some bird, fish and invertebrate species, and also provide access to indigenous peoples for hunting and fishing. 

The high levels of variability for all of these factors complicate the number of scenarios and considerations that must be taken into account for NEBA for contingency planning.  Use during response actions could be simpler because conditions that exist at the time of the spill can be readily observed, and it may not be necessary to consider the full range of seasonally variable conditions.  However, attempts to ensure all resources and exposure factors have been taken into account and all stakeholder issues have been addressed when based on a priori contingency planning has been based on a fully developed NEBA.

A re-examination of the key processes used in NEBA, as identified by IPIECA is useful for illustrating some of the challenges to its application in the Arctic (IPIECA 2000):

• “Collect information on physical characteristics, ecology and human use of environmental and other resources of the area of interest.”

  Such work often focuses on physical characteristics, ecology, and human uses of many areas exhibit a high degree of seasonal variability.  Background information is central to understanding local conditions and values, particularly unusual information such as known natural oil seeps or background tarball concentrations. This information is generally known, and sets the stage for the discussion of "how clean is clean" during a response, as to when an area can be considered back to normal.

• “Review previous spill case histories and experimental results which are relevant to the area and to response methods which could possibly be used.”

  Spill case histories in the Arctic are dominated by small spills, mostly related to spills during transportation and fishing vessel groundings.  There are no case histories that would be representative of the worst case scenarios.

• “On the basis of previous experience, predict the likely environmental outcomes if the proposed response is used, and if the area is left for natural clean-up.”

  Again, most previous experience has been with small spills and their environmental impacts are seasonally influenced.  There is little available data on the likely environmental outcomes of response techniques such as in-situ burning and dispersant use, and what is available has come from smaller scaled, controlled release studies conducted during less ecologically sensitive seasons or from non-Arctic environments.  More data are available for assessing impacts on areas left for natural cleanup.  Mathematical models may be helpful, but the accuracy of most models is affected by the presence of sea ice, which can be highly variable.

• “Compare and weigh the advantages and disadvantages of possible responses with those of natural clean-up”. 

  In many cases weather conditions, sea state, and ice coverage will limit response options, or at least postpone them until more favorable conditions exist.  Arctic NEBAs will have to take into account a wider range of response effectiveness and uncertainty in outcomes.

These potential challenges do not preclude uses of NEBA in the Arctic.  The fact that planning and response decisions must be made, regardless of the complexity and uncertainty involved suggests that NEBA should be used, particularly in a contingency planning basis to help guide response decision-making and to identify priority areas where uncertainties should be reduced, and to ensure spill readiness.   The methods used in implementing NEBA, however, will need to be robust and each phase of the analysis will need to account for high levels of variability and uncertainty.