Recovery of environmental compartments (ECs) and valuable ecosystem components (VECs) can vary from days to decades depending on a wide range of biological, chemical and physical variables.  In order to better inform consequence assessments of response actions, evaluation of resiliency, or the ability of VECs or ECs to cleanse themselves from oil contamination and to repopulate communities that have been affected needs to be included in assessments of short-term impacts.  Differences in resiliency need to be evaluated; for example, the time for replacement of marine mammals or isolated bird or fish populations impacted from surface or stranded oil could be compared to the time for replacement of pelagic copepods or phytoplankton that are impacted by oil that is physically or chemically dispersed within the water column.  Examination of relative resilience potentials suggests that oil spill response actions minimizing impacts on interface environments, long lived or sensitive populations would be preferred.  Consequence analyses of response actions need to extend beyond the acute effects and examine chronic impacts that may put populations at risk to determine which action is most effective and results in minimizing the ultimate impact to the broader aquatic environment.

Recovery from an oil spill event is controlled by the resilience of an environmental compartment to biological, chemical and physical stressors occurring within each environmental compartment.  There is a lack of understanding of the processes that determine resilience for the Arctic as well as all aquatic environments leading to uncertainty in the assessment of recovery.  Quantification of recovery is not a standardized assessment making it difficult to document when or whether an environmental compartment has been restored.  For the purposes of this section we define recovery in terms of structural and functional components of resident populations.  The structure of the compartments is based on quantifiable characteristics of the numbers and types of species, their abundance and biomass (standing crop) or production.  The function of the compartments is based on the flow of nutrients through a compartment and the services that those organisms provide to other components.  Recovery for each of these structural and functional characteristics within different environmental compartments has different time scales.   

Following an oil spill in any environment, different oil spill response (OSR) methods may be used as countermeasures during a spill situation and/or for clean-up in the aftermath.  This section focuses on how different OSRs will affect the short and long-term fate of oil in the environment and how this will influence recovery time. Except for mechanical recovery that, under special conditions, can remove a certain percentage of the spilled oil, most OSR actions shift the quantity of oil into different environmental compartments with the objective of reducing the overall consequences of the spill and enhancing recovery.  Consequence analyses should include the immediate effects of the spill event and the longer term consequences resulting from differences in the potential for an ecological recovery of affected compartments within the broader ecosystem.  Little relevant information exists on ecosystem recovery from oil spills in Arctic environments, but we can draw conclusions and identify uncertainties based on examples from sub-Arctic environments. 

Among other factors, ecosystem recovery depends on the amount of oil in various environmental compartments.  The physical and chemical properties of the spilled oil influence the availability for microbial processes, bioavailability, and the acute and chronic toxicity of the oil mixture to local fauna.   Oil Spill Response (OSR) methods as described in Section 4 are: 1) no response, 2) in-situ burning, 3) use of oil mineral aggregates (OMA), 4) chemical dispersants or herders, and 5) natural recovery.  Application of one or several OSR methods alters the amount of oil introduced into various environmental compartments, influencing the total consequences of the spill.  Determination of the relative resiliency and the ability of impacted faunal groups to restore after oil exposure is a key component of the assessment but it is less studied and understood.   

Table 8-1 explores potential consequences of two decision scenarios (i.e. use of dispersants or non-use) and provides a description of the various attributes of environmental compartments and how they may be affected by these two potential response actions.  A question remains:  what attributes can be used as indicators of ecosystem recovery, i.e. that the environment has returned to its ‘pre-spill’ condition?  In a review by Baker (1999) a few suggestions are put forward.   Recovery was defined on the basis of hydrocarbon concentrations, i.e. that hydrocarbon concentrations do not exceed normal background levels for a particular location.  The criteria are further refined on the following basis:

1. Do not exceed any statutory limits
2. Are not lethal to specified organisms
3. Do not cause deleterious effects to specified organisms
4. Do not cause tainting of food organisms
5. Have no detectable impact on the function of an ecosystem
6. Do not impair the human use of an area
7. Are not visible to the human eye
8. Cannot be reduced by enhanced clean-up actions without causing an overall retardation of recovery or damage to the recovering ecosystem

In the context of the present section, criterion 5, "Have no detectable impact on the function of an ecosystem," is the most relevant suggestion and is discussed in more detail in the following paragraphs.

The recovery of an ecosystem following an oil spill is a function not only by the return of contaminant levels to pre-spill state, but more importantly to a dynamic equilibrium state that the ecosystem would have been in had the spill not occurred.  It must also be realized that there is no single condition of an ecosystem and that changes to the ecosystem and its environmental compartments occur through time and space.  Parker and Maki  (2003) define recovery for a biological resource as "occurring when the injured resource reaches a state it would have been if the impact had not occurred".  Temporal variations in biological resources affect the assessment of the recovery process; however, the definition is robust to temporal variation because it does not require the resource to return to a specified pre-injured state. 

Whether an ecosystem affected by human impact returns to its original condition is difficult to measure. Even if measurable parameters that describe the ecosystem structure and relevant reference sites or historical data are available, often ecosystems do not return to their original state, but rather establishes a new “baseline” over time.  At the minimum, this new state might be expected to be at a level of productivity, growth or population dynamics that would sustain the ecosystem trends at rates and states consistent with pre-spill dynamics.  With constant and/or fluctuating natural changes, how can we know what an affected area is supposed to be 20-30 years following an accidental oil spill?  Changes, human-induced or natural, lead to shifting baselines imposing dynamic trajectories for ecosystem status in reference sites as well as human-impacted sites (Duarte et al. 2009).  As such, and in an ecosystem context, ecotoxicology needs to be extended from not only treating toxicity on an individual or organism level, but also to include interactions among a variety of biotic and abiotic factors and components. 

Table 8-1.  Consequences to environmental compartments resulting from OSR decisions to employ chemical dispersants or physical recovery of surfaced oil

Parker and Maki (2003) define a biological resource in three ways: by taxa, summary statistics, and communities.  Recovery should be considered separately for individual taxa assessed, and different taxa may have different timeframes of recovery or resilience to an impact.  Summary statistics provides a means for combining taxa, and statistics can be valuable when different taxa vary in time and space.  Communities add meaning to the way taxa interact with each other, e.g. diversity indices have been used to define a resource as a community.  As such, the availability of reference or historical time series data will enhance the assessment of ecosystem recovery.  The ability of an ecosystem to recover and the speed with which it does so, depends therefore on a lot of interconnecting factors and dynamics between species and communities and the environment.  Rates and extent of recovery can be estimated through structural and functional relationships, e.g. the productivity or diversity of communities (Lotze et al. 2011).  Resilience of functions of an ecosystem is strongly correlated to biodiversity (Levin and Lubchenco 2008), the general stress condition of the systems and to the strength of impact of the perturbation and the reproductive potential of the community that was damaged (Gunderson et al. 2010).

In relation to Arctic marine ecosystems, it is important to note that recovery may be linked to generation time (Lotze et al. 2011) and specific features of the Arctic environment such as a short seasonal window of growth and presence/absence of sea ice. The latter may influence the exposure history following a spill and the possibilities of trapping or removing the oil, which is considered to be a major factor in ecosystem recovery (Chapman and Riddle 2005).  Another important note in relation to pressures on the Arctic ecosystems, is the effect of climate change, which apart from opening up possibilities of increased petroleum exploration, has increased the general stress on the system with concerns of acting with synergistic effects (Clarke and Harris 2003; Macdonald et al. 2005).

The BIOS experimental oil spill is one of the few experimental oil spills in the Arctic where changes in biota have been followed. This experimental oil release took place at Cape Hatt, northern Baffin Island in August 1981.  Approximately 15 m³ of slightly weathered (8% removal) Lagomedio crude oil was applied to the surface in one bay (bay 11), 15 m³ of dispersed oil (DOR 1:10, Corexit 9527) was released to the surface of another bay (bay 10), and finally 15 m³ was released underwater in a third bay (bay 9). A fourth bay (bay 7) was used as a reference station. Following these crude oil releases, oil concentrations were measured in 1981-1983 and in 1985 and 1987, and effects on Arctic macrobenthos were studied in the acute phase and during a period of 2 years post-spill (infauna: Cross & Thomson (1987), epibenthos: Cross et al. (1987a) and macroalgae: Cross et al. (1987b).  No apparent short-term effects of untreated oil were observed on shallow-water infauna, whereas marked acute effects were observed for the dispersed oil treated bay after 24 hours. However, no large-scale mortality of benthic infauna and no significant change in infauna community structure were observed for either treatment. During the recovery phase, the authors claim that most affected infauna organisms recovered, although some persistent changes to bivalves (reproductive effects) and polychaetes (density) were observed even two years after the spills occurred (Cross and Thomson 1987).  For epibenthic organisms, narcosis in sea urchins and starfish was observed in the acute phase; however, no major effects on the density of crustaceans were observed over time.  Some minor effects were indicated in 2 of 22 of the sites analyzed where possible delayed reproduction and depth distribution of amphipod were suggested (Cross et al.1987a). No effects were observed on macroalgal biomass, number of species and reproductive condition.  Only a slight increase in growth for one alga was observed over time, suggested to be a result of reduced herbivore activity (Cross et al. 1987b). Summary observations may best be characterized as the presence of oil with or without dispersants had relatively short-term, localized impacts for this test spill that affected a relatively small area.

Although the Exxon Valdez oil spill (EVOS; 42 million liters of Alaskan North Slope oil were spilled) was not a true Arctic oil spill, it can be used as an example of ecosystem recovery assessment in a sub-Arctic area that represents considerations of impacts over a large area.  Numerous papers have dealt with the recovery of specific biological resources in the Prince William Sound (PWS) region since the spill occurred in 1989, and it is beyond the scope of the current literature review to evaluate all information that exists from EVOS. Furthermore, there are still ongoing disagreements between scientists as to what extent the ecosystem has recovered. In a review by Peterson et al. (2003) the general conclusions were that oil persisted in some areas (limited areas compared to the miles of initially oiled shoreline) well beyond a decade in surprising amounts and toxic forms and was sufficiently bioavailable to cause chronic biological exposure and long-term population level impacts.  One of the underlying arguments for this was chronic persistence of oil in sediments causing prolonged exposure and health effects to sediment-associated organisms.  On the other hand, a review by Harwell and Gentile (2006) presented an assessment based on valued ecosystem components (VEC) and concluded that the PWS ecosystem had effectively recovered from EVOS.

"Ecosystem recovery" will occur after an oil spill but understanding when an ecosystem has recovered sufficiently has serious gaps of knowledge.  There are no papers that deal with ecosystem recovery after an oil spill in the high Arctic and only a few experimental studies in the Arctic that provide limited information to evaluate an oil spill (e.g. BIOS experiments). The studies in Prince William Sound intertidal rocky/cobble substrates suggest that there are ongoing ‘lingering’ effects of the Exxon Valdez spill that occurred more than 20 years ago while other assessment indicate the functional aspects have recovered.  These weaknesses highlight the need for a fundamental understanding of ecosystems with a focus on oil toxicity studies at the population and community/ecosystem level.  The review of ecosystem recovery in the Arctic described by the authors in this section led to suggestions of further research which can reduce remaining uncertainties.  The more generic suggestions compiled from this review are summarized below while recommendations that are important for improving Arctic NEBA are listed separately.

1. Ecosystem recovery.  New effort needs to focus on appropriate information collected throughout the Arctic and sub-arctic for decades as well as experimental studies to demonstrate rates of recovery under controlled conditions.  Expand on the conceptual basis of this section to develop an assessment of the physical and chemical attributes that influence environmental compartment resiliency and ultimately, recovery potential.
2. Utilize baseline studies.  Environmental baseline studies are conducted at exploration/production sites.  Collect, review, coordinate consistency among biological data sets, and archive the validated information on communities of organisms living in different environmental compartments.  Establish baseline conditions throughout the Arctic regions anticipated to undergo oil exploration in order to provide the natural range in those characteristics so that the state of recovery can be evaluated.
   1. Population parameters to be collated or obtained include: overall population estimates for the species abundance, variations in daily mortality rate coefficients, fecundity measures, age structure of populations, frequency of spawning, local broadcast larval distribution, and regions of special biological significance for key species.
   2. These population characteristics for each species need to be placed into a species resiliency database to provide more accurate estimates of the ability for a species to recover from a stressor. 
   3. The effects of non-dispersed oil on organisms utilizing the sea surface such as seabirds and marine mammals, as well as intertidal and shallow subtidal communities are subjects needing further study.  These investigations need to include impacts due to physical fouling and an examination of toxicity caused by various modes of action.

The recommendations presented below indicate where increased knowledge of oil transport and fate processes would result in reducing existing uncertainties in NEBA assessments.  No prioritization has been made to the list; for some of the recommendations, surrogate data may be already available.

1. Continue ecosystem-level investigations.  Conduct mesocosm experiments that focus on biological interactions of several ecosystem elements instead of single species responses, such as the BIOS experiments.
2. Compile data on resiliency metrics of VECs and ECs.  Assign semi-quantitative score to species/populations that are considered VECs to Arctic ECs. 
3. Provide a GIS mapping of seasonal and spatial VEC population use of EC and their resiliency scores.
   1. Recovery potentials
   2. Vulnerabilities
   3. Seasonal weathering/biodegradation and toxicity of OSR residuals
4. Introduce recovery metrics.  How much change is beyond natural variation?  What ecosystem-level attributes should be assessed?  Develop exemplary scenarios.  Implement full-scale ecosystem consequence analysis.

Authors Dr. Bjørn Henrik Hansen (SINTEF), Dr. Jack Q Word (ENVIRON), Drs. Gina Coelho and James Clark (HDR/EM&A), Dr. Morten Hjorth (COWI), Dr. Lionel Camus (Akvaplan NIVA), Dr. Torgeir Bakke (NIVA)