Eastern Bering Sea Assessment - 2019

Ecosystem Assessment (pdf)

Elizabeth Siddon

Auke Bay Laboratories, Alaska Fisheries Science Center, NOAA Fisheries Contact: elizabeth.siddon@noaa.gov

Last updated: November 2019


The primary intent of this assessment is to summarize and synthesize climate and fishinge.ects(historical and future) on the eastern Bering Sea shelf and slope regions from an ecosystem perspective. The Ecosystem Status Reports of the Groundfish Stock Assessment and Fishery Evaluations (SAFE) provide the historical perspective of status and trends of ecosystem components and ecosystem-level attributes using an indicator approach. For the purposes of management, this information must be synthesized to provide a coherent view of the ecosystem effects to clearly recommend precautionary thresholds, if any, required to protect ecosystem integrity. The eventual goal of the synthesis is to provide succinct indicators of current ecosystem conditions and a prognosis of how fishstocks are expected to fare, given concurrent information on ecosystem status. To perform this synthesis, a blend of data analysis and modeling is required annually to assess current ecosystem status in the context of historical and future climate conditions.

Recap of the 2018 Ecosystem State

Some ecosystem indicators are updated to the current year (2019), while others can only be updated to the previous year (or earlier) due to the nature of the data collected, sample processing, or modeling efforts. Therefore, some of the "new" updates in each Ecosystem Status Report reflect information from the previous year(s). Below is a complete summary of 2018 that includes information from both previous and current indicators. The next section (Current Conditions: 2019) provides a summary of the 2019 ecosystem state based on indicators updated this year.

With rapidly changing environmental conditions and subsequent changes in sea ice extent in the Bering Sea, NOAA's Alaska Fisheries Science Center has conducted ecosystem and stock assessment surveys in the southeastern and northern Bering Sea to better understand how such changes will impact fish stocks in the region. In this Report, we describe ecosystem responses within the southeastern and northern regions separately, while recognizing that recent reduced sea ice has highlighted the connectivity between them.

The winter of 2017/2018 saw a composite of unusual weather events that resulted in the first unprecedented near-complete lack of sea ice in the northern Bering Sea. Several climatic forces occurred that resulted in the warm conditions: (i) residual heat maintained above-average water temperatures that caused delayed freeze-up (e.g., the Chukchi remained ice-free into January 2018, ice arrived late (March) and departed early (April)), (ii) a large and persistent high-pressure system from February through April over the Aleutian Islands and southern Bering Sea, which shifted the position of the Aleutian Low Pressure System (ALPS) northwest over Siberia, and (iii) highly unusual winds from the southwest that brought warm air over the Bering Sea and prevented sea ice from forming until March.

Southeastern Bering Sea

In the southeastern Bering Sea, conditions were typical of a low-ice year with above-average water temperatures and complete lack of a cold pool. Reduced strength of water column strati.cationmeant waters were easily disrupted by storms. Notably, both temperature and salinity below the mixed layer depth were above-average in 2018, which has not been observed previously. A small phytoplankton bloom began May 25 and increased surface dissolved inorganic nitrogen concentrations indicated increased summertime wind mixing events. Higher nutrient stores at the surface generally mean higher production and more energy available for transfer to higher trophic levels. In early spring, small and large copepods were in low abundance, but by late spring small copepod abundances increased to one of the highest levels recorded. Adult euphausiid densities remained low, continuing a trend since 2012. Overall, delayed production gave rise to a low quality zooplankton prey base over the entire shelf. Reduced energy transfer from the prey base to the top-level predators likely contributed to poor body condition and observed mortality events (i.e., seabirds and mammals).

Larval pollock production was high, similar to other warm years, but the poor prey base was predicted to result in reduced growth and survival of juvenile fish. The bottom trawl survey indicated Pacific cod and pollock abundances were below their long-term means. Pacific cod abundance was lower, but biomass was higher (fewer, larger fish). There was no evidence of pollock recruitment, with low age-1 numbers and abundance dominated by the 2012 year class. Age-1 pollock abundance has been low since 2014.

Unusual observations of adult pollock behavior were reported from community members as well as subsistence and commercial fishers in Bristol Bay from May-July 2018. Sightings of pollock swimming 'with their heads out of the water' and 'behaving odd' were reported. Adult pollock were reported to have washed up on shore in high numbers, and pollock were reported to have been caught in subsistence set nets during the salmon fishery(both near Pilot Point, UgashikRiver, and from the NushagakDistrict). "Traditionally, it is unusual to see pollockin the salmon season in Bristol Bay" (CatieBursch, fisher, pers. comm.).

Samples were sent to NOAA Fisheries and fish were "definitely skinny compared to average weights by year/size from NOAA's bottom trawl survey" (B. Lauth, pers. comm.). Stomach contents included Caridea shrimp, gammarid amphipods, and other fish, which "seems pretty typical for Bristol Bay" (K. Aydin, pers. comm.). Stomach and intestine samples were processed for PSP toxins (i.e., saxitoxin) at the NWFSC/WARRN-West Program. All samples contained low but detectable levels of PSP, confirming that an exposure risk was present in the food web. Therefore, PSP toxins could have played a role in the unusual behaviors and mortality events observed in Bristol Bay, although it is important to note that the levels were well below the seafood safety regulatory limit and therefore were not a risk for human consumption (K. Lefebvre, pers. comm.).

Northern fur seal pup production at St. Paul Island was ~6% less than 2016. Pup production has been declining at St. Paul Island at an approximate annual rate of 4.0% since 1998. Anecdotal reports suggest smaller or skinnier pups at some rookeries on St. Paul Island. The estimated St. George pup production is approximately 5% greater than 2016 with no unusual pup mortality, but shows no signi.canttrend since 1998. The decline of northern fur seals is also apparent to Elders, adults, and youth. Subsistence harvests represent the importance of securing seal meat to provide food security for winter (L. Divine, Aleut Community of St. Paul Island).

At the Pribilof Islands, seabird nesting was delayed, and/or unsuccessful and reproduction was poor. For example, black-legged kittiwake reproductive success failed completely, indicating prey for these surface-foraging birds was likely not su.cientfor them to successfully rear chicks. On St. George Island, common murresexperienced the latest mean hatch ever recorded and thick-billed murres experienced their 2nd latest hatch date. The Ecosystem Conservation Office(ECO) of the Aleut Community of St. Paul Island reported that St. Paul Island community members noted how 'quiet' the cliffs were in May 2018. Although birds arrived around the usual time of year, they did not engage in typical nest building activities. Residents were unable to collect subsistence murre eggs because there were none on the cliffs during the egging season (June). Least auklets have been declining since 2015; no auklets have been subsistence harvested in the last two years due to declining breeding colonies. In early July 2018, approximately 50 fresh-dead northern fulmars, shearwaters, and murres were collected on beach surveys. All necropsies revealed the birds died of starvation. Community members continue to speak of a lack of food in both summer and winter as the cause for die-offs, population declines, and reduced productivity and fledgling success (L. Divine, Aleut Community of St. Paul Island).

Northern Bering Sea

2018 marks the lowest ice year on record for the eastern Bering Sea. Bottom temperatures in the northern Bering Sea (NBS) were 1°C to 2°C rather than <-1°C, and no cold pool formed. Historically, salinity and temperature contributed equally to the vertical stratification of the water column in the NBS. In 2018, salinity in May was vertically uniform. The lack of salinity structure resulted in weaker vertical stratification, permitting greater vertical mixing.

The near-complete lack of sea ice created an absence of ice algae to 'seed' productivity; the bloom began quite late and strati.cationwas weak. A low proportion of large phytoplankton cells was observed, consistent with indications of low productivity. The abundance of small copepods was similar to 2017, but the abundance of large copepods was an order of magnitude lower and abundance of juvenile euphausiidswas near zero. The low nesting success of thick-billed murres, which eat large zooplankton, provides further evidence that large, lipid-rich zooplankton were in low supply. However, the total energy content of juvenile forage fishspecies showed positive responses to increased temperature conditions from 2017 to 2018.

Pacific cod and pollock persisted in the NBS. When sampled in July 2018, Pacific cod were 'fat' and 'healthy' and anecdotal observations were that stomachs were 'full of Opilio'. Based on bottom trawl survey results, more than 50% of the overall estimated biomass of Pacific cod was found in the NBS. With half of the biomass in the northern survey area, stock assessment models for 2019 included NBS data for the first time. In contrast, pollock estimated biomass in the NBS declined by 14% from 2017 to 2018. Pollock to the east of St. Lawrence Island were in 'poorer condition' while pollock sampled north of St. Lawrence Island were 'plump and healthy' (B. Lauth, pers. obs.). Gay Shefield(Alaska Sea Grant Marine Advisory Agent in Nome, Alaska) recounted observations from community members who noted large Pacific cod (1m) caught in crab pots, record returns of pink and silver salmon, and that halibut fishing off Savoonga was 'really great'.

Ice seal distributions were dramatically impacted by the lack of sea ice over the NBS shelf. In 2018, the nearest ice edge over the shelf was about 375km to the northeast of where it was historically found. Ice seals (particularly ribbon seals) were unusually scarce with no evidence that they moved northeast or into Norton Sound following the ice edge. Spotted seal pups weighed less than in recent years, continuing a declining trend in body condition and blubber thickness since 2014. More walrus and bearded seals were seen than typical, likely because the survey was farther north in areas where these species are more prevalent. Large numbers of dead marine mammals were found along the shorelines from north of Bering Strait, throughout Norton Sound, and on St. Lawrence Island.

Beginning in February, residents in shoresidecommunities reported walruses were harvested off St. Lawrence Island, a time when they are not typically accessible. The walruses were 'fat and in good condition'. In early June, there were reports of exceptionally high numbers (i.e., 50) of dead seals on beaches (primarily young bearded seals) along the north side of St. Lawrence Island and 48 beach-cast ice seals (bearded, ringed, and sub-adult spotted seals) near Wales. The seals appeared to have poor body condition and empty stomachs. There were also reports of seal strandings, reports of sick/dead seals out of normal range, and seals were absent or unusually scarce in the Port Clarence area (G. Shefield, pers. comm.)

A seabird die-off event, unprecedented in terms of spatial and temporal scale, occurred in 2018. Large numbers of seabirds (mainly murres) washed ashore from Bering Strait southward through Norton Sound, and along the shores of St. Lawrence Island. Starvation is the only identified cause of death. The historical diet composition of the affected seabirds suggests that large, lipid-rich zooplankton may have been in short supply in 2018. Thick-billed murresand shearwaters take euphausiids, amphipods, and forage fishwhile common murres take mostly forage fish. Reproductive failures and poor reproductive success (mainly murresand kittiwakes) were observed by community members, subsistence eggers, and scientists. Those birds that did nest, nested very late. All murrereproductive effortfailed north of St. Matthew Island, such that murresdid not produce chicks, which is exceptional, and prior to the 2016 heatwave, widespread murre reproductive failures had not been observed.

Subsistence hunters for crested and least auklet fledglings usually get ~30-40/day. In 2018, a hunter stopped after he pulled 13 dead fledglings and no live ones. An Elder on St. Lawrence Island said "when there is no dirty ice, there is less food for the krill, and consequently, no food for the birds." At Sledge Island in Norton Sound, 50-70% of murreswere missing from cli.sand those that were there were not laying eggs; local people were stunned and there was a complete lack of harvest. There were dead and emaciated murres, shearwaters, and crested auklets in Nome and on St. Lawrence Island (G. Shefield, pers. comm.).

Harmful Algal Blooms (HABs)

A 2018 coordinated research effortcentered around the Bering Strait Region made progress towards improving baseline understanding of HABs species abundance, distribution, and toxin presence. Research results found no positive toxicity or significantHAB cell counts south of the strait in the NBS (June-September), however high concentrations of HABs cells (Alexandrium spp.) were found in the Chukchi Sea in June and August.

Current Conditions: 2019

The eastern Bering Sea experienced the second year with little winter sea ice (winter 2018/2019) and reduced cold pool extent (summer 2019), prompting the expression of a 'double whammy' for ecosystem impacts. The observed e.ectsin 2019 are discussed below in terms of (i) reflecting 2018 conditions, (ii) reflecting 2019 conditions, and (iii) reflecting cumulative impacts of continued warm conditions from 2018 to 2019.

Reflective of 2018 conditions:
Ecosystem responses that may reflect 2018 conditions include the gray whale Unusual Mortality Event (UME) and the short-tailed shearwater die-off event. Lagged (delayed) impacts of poor feeding conditions experienced during 2018 are hypothesized to at least partially explain the mortality events. Both species feed in the Bering Sea during summer; gray whales are benthic feeders (e.g., amphipods) while shearwaters are planktivorous(e.g., euphausiids). Both species embark on long migrations to the southern hemisphere for breeding during the austral summer. Therefore, the 2019 mortality events may reflect 2018 feeding conditions in the Bering Sea, conditions experienced during the breeding season in the southern hemisphere, or lack of available prey to complete the migration to the Bering Sea in 2019.

Reflective of 2019 conditions:
It is important to remember that the 2018/2019 winter sea ice accumulation differed from winter 2017/2018. While residual heat delayed freeze up until mid-December (similar to 2017/2018), accumulation approached the long-term mean sea ice extent through January 2019 (whereas 2018 sea ice never approached normal levels) before southerly winds persisted during February and reduced sea ice (similar timing between years). The impact and subsequent ecosystem effects of early sea ice are not fully understood, although it is hypothesized the small, retracted cold pool (i.e., thermal barrier) may have contributed to the increased biomass of pollockover the southern shelf in 2019.

The timing of the peak spring bloom in 2019, estimated from satellite data, was earlier than the long term average (by approximately 9 days) and earlier than 2018. The spring bloom fuels secondary production of the zooplankton prey community that forms the base of energy transfer for upper trophic levels (e.g., fish, seabirds, mammals). In 2019, the zooplankton community was dominated by small copepods, which is typical of warm-year conditions over the shelf, while large copepod and euphausiidabundance was low. Warm temperatures increase copepod secondary production rates; combined with an earlier bloom, this may have resulted in sustained production and energy transfer into summer 2019. However, larger, lipid-rich copepods and euphausiidsremained low across the shelf and could affect energy storage and survival of juvenile fish within the region. Jellyfish, which are pelagic consumers of zooplankton and small fishes, showed a sharp increase in abundance in 2019 relative to 2018 and their long-term mean.

At the Pribilof Islands, reproductive success of several seabird species improved in 2019, indicating that birds were able to find suffcient food resources to support reproductive effortsduring 2019. Successful breeding events occurred for fish-eating species (i.e., murres at St. George and red-faced cormorants at St. Paul) and plankton-eating species (both species of kittiwakes at both islands). Feeding conditions and prey availability in the environment were suffcient to support reproduction, but oceanographic surveys showed the zooplankton community dominated by small copepods. Several possible explanations include: (i) seabirds may have been successful at finding lipid-rich copepods and euphausiids, even though abundances were low; (ii) competition for available prey was reduced as a result of shearwater mortalities and/or poor recruitment events for fish species; and/or (iii) colonies at the Pribilof Islands may have benefited from northward shifts in fish populations. Evidence supporting these hypotheses include the below-average coccolithophore bloom index for 2019 and the high abundance of age-0 pollock collected in the northern Bering Sea surface trawl survey.

Unprecedented warmth in the inner domain during summer 2019 likely resulted in increased metabolic demands and may have impacted fishdistributions. For example, the increased warmth may have pushed Pacific cod away from warm inshore waters and made them more available to the bottom trawl survey, contributing to the large increase in abundance of young fish in 2019. In Bristol Bay, sockeye salmon returns remained high. The large inshore run in 2019 suggests these stocks experienced positive conditions at entry into the eastern Bering Sea in the summers of 2016 and 2017, and winters of 2016/2017 and 2017/2018.

Reflectiveof cumulative impacts:
Adult groundfish condition increased in 2019 relative to 2018 and was positive for several key species sampled during the standard bottom trawl survey. Large increases were shown for adult pollock, Yellowfin sole, and Arrowtooth flounder, suggesting a possible shift to a benthic-dominated system. Arrowtooth flounder biomass continued an increasing trend since 2017. Groundfish condition reflects ecosystem productivity, especially when viewed across stocks, and population dynamics within a stock.

The 2018 pollock year class showed strong overwinter survival to age-1. However, several mechanistic relationships based on bottom-up pathways predicted below-average recruitment success for the 2018 year class (see Forecasts and Predictions below). The occurrence of strong recruitment in light of low sea ice, above-average water temperatures, and poor prey quality/quantity challenges our current understanding of recruitment processes for pollock. One hypothesis stems from the anomalous February winds from the southwest. Winds brought warm, moist air from the south over the shelf, but may have also increased on-shelf flow and upwelling conditions. Upwelling of productivity during winter may have subsidized energy transfer and contributed to increased survival of age-0 pollock. In combination with greater overwinter survival, the 2018 year class may have experienced reduced predation pressure from cannibalism because recruitment of recent year classes has been low. Pollock age-1 natural mortality peaked in 2016, but has declined in recent years and the 2019 estimate was at the long-term mean, also demonstrating reduced predation pressure for the 2018 year class. That said, the bottom trawl survey results showed a 75% increase in pollock biomass over the southern shelf from 2018 to 2019, which suggests that adult fish moved into the region (recruitment of small fish could not account for such a large increase in biomass). Therefore, the 2019 year class may experience increased predation pressure from cannibalism (from the 2018 and older age classes).

The declaration of an Unusual Mortality Event for ice seals reflects cumulative impacts of conditions in 2018 and 2019. The increased mortality of seals and the apparent decline in pup condition demonstrate immediate and delayed impacts of the loss of sea ice habitat for pupping and nursing in both years. They could also demonstrate broader ecosystem e.ects, such as competition for prey from northward shifts of fish populations. At the Pribilof Islands, community members reported unusually high numbers of male fur seals overwintering at St. Paul (i.e., they never left from the 2018 season), reflective of the continuous warmth throughout the year.

Forecasts and Predictions

9-month Ecosystem Forecast for the Eastern Bering Sea:
Since 2013, AFSC and PMEL have produced 9-month ahead forecasts of the eastern Bering Sea cold pool extent that have contributed to the ecosystem assessment. During re-forecasting experiments, and in light of the anomalous conditions of 2018 and 2019, it was determined that the previously-used method of a single ROMS model run is limited in its ability to forecast; in particular, there is a strong "predictability barrier" in January-March that limits the model's ability to forecast conditions past this window. A multi-model ensemble forecast method is currently in development to address this uncertainty for future ecosystem assessments.

Pollock Recruitment Predictions
The EBS Ecosystem Status Report includes several leading indicators of pollock recruitment that give, in some cases, contradictory results. In this section, we have summarized these predictions so that we can more easily track how they compare and how well they hold up over time. Survival and recruitment success of juvenile pollockare driven, in part, by bottom-up processes. The abundance, species composition, and quality of zooplankton prey resources are governed by large-scale oceanographic processes and vary under changing thermal conditions.

Below we track available predictions for the 2018 year class of pollock:

•The Temperature Change index (p. 118) predicts above average recruitment to age-4 from the 2018 year class. The 2018 cohort experienced a cool summer during the age-0 life stage followed by a relatively warm spring during the age-1 period.

• Surface silicate concentrations (an indicator of nutrient availability in surface waters; p. 120) were well above average in 2018, yet age-0 pollockweights were below average, suggesting below average recruitment to age-1.

• Diet energy density was low to intermediate for 2018, suggesting below average recruitment to age-1. It is worth noting, however, that euphausiids comprised over 50% of age-0 pollockdiet (p. 87). The energy density of age-0 pollock during late summer 2018 was low and would predict below average recruitment to age-1.