NOTE TO READER: The text below was developed by the CIBRT, with a few minor updates, and reproduces information readily available in other reports. Additional details regarding the State of Alaska’s fisheries management practices and fisheries harvest information can be found in ADF&G publications, such as annual commercial fisheries management reports (e.g., Shields and Dupuis 2016). In Sections II.B.10 and III.A.6 of this document, we provided information sufficient to justify recovery criteria and actions addressing CI beluga prey. Additional CI beluga prey information follows.
Prey Abundance and Distribution
Eulachon is a primary prey item of CI belugas from May to early June. They enter glacial rivers to spawn shortly after the river ice has melted and the water flows freely. Eulachon have high oil-content (17–21% of the wet weight; Payne et al. 1999) and migrate in dense schools. Large eulachon runs in Cook Inlet occur in the Susitna River and at Twenty Mile River in Turnagain Arm, with smaller runs in other glacial rivers entering Cook Inlet (Figure F1). Eulachon biomasses in these rivers are unknown. The NMFS biennial bottom trawl survey estimates of eulachon biomass in the central Gulf of Alaska are highly variable (5,255 short tons in 1984, 104,709 tons in 2003, and 54,246 tons in 2011) (Ormseth 2011). In the Susitna River and Twenty Mile River, the eulachon spawning migration peaks in late May and is largely completed by mid-June (Barrett et al. 1984; Spangler et al. 2003). Commercial fishing for eulachon/smelt (eulachon are not distinguished from other smelt in ADF&G harvest reporting) occurs annually in saltwater between the mouths of the Chuitna and Susitna rivers (Figure F1). Harvests have ranged from 41–97 metric tons (45–107 short tons) since 2006 (Table F1) (Shields and Dupuis 2016). Commercial harvest of eulachon has increased substantially in recent years (Table F1).
Source: Shields and Dupuis 2016.
Personal use harvests in Cook Inlet are summarized by ADF&G Division of Sport Fish reporting areas (Figure F2). Although fishing effort for personal use harvests of smelt responds to socioeconomic variables (e.g., gasoline prices), recreational effort likely reflects population abundance of spawning smelt. Thus, strong spawning returns likely generate increased fishing effort such that recreational harvests index the relative magnitude of the spawning populations.
Recreational harvests for Cook Inlet during 1996 to 2011 showed high interannual variability within and among harvest reporting areas (Figure F3). Although the late 1990s and mid-2000s exhibited generally higher smelt harvests, the correlation of annual harvests among reporting areas was relatively low (the maximum correlation was 0.50 between log transformed values for the Susitna River drainage and the Kenai Peninsula freshwater). In general, the largest personal use harvests occurred in the Anchorage area, mainly represented by Twenty Mile River in Turnagain Arm. Harvests in most areas increased in recent years, particularly for the Anchorage area.
From June to September, salmon are the primary beluga prey in Cook Inlet. Quakenbush et al. (2015) found primarily coho, chum, and Chinook salmon in analyses of salmon remains in stomach contents, indicating that some salmon species may be of greater importance (Table 2). During this period, belugas are often found from Tyonek to the Little Susitna River and in river mouths of Knik and Turnagain arms. The largest salmon runs in Cook Inlet enter the Kenai, Kasilof, and Susitna rivers. Chinook salmon runs peak in the Susitna and Little Susitna rivers in mid-June, in the Kenai River in mid-July, and in the Kasilof River in late June to early July (Figure F4). Sockeye salmon runs typically peak in mid-July, pink salmon and chum salmon runs peak in late July or early August, and coho salmon runs peak in August (Figure F4).
However, run timing differs among species, streams, and years.
Sockeye salmon are the dominant species in the Kenai and Kasilof rivers with significant numbers of Chinook, coho, and pink salmon also spawning in the Kenai River. The Chuitna, Beluga, Theodore, and Lewis rivers support relatively small runs of Chinook salmon and somewhat larger runs of coho salmon (Figure F5). The Susitna River drains the largest watershed entering Cook Inlet and supports substantial runs of all five salmon species (Figure F5). The Little Susitna River supports moderately sized runs of pink, chum, and coho salmon (Figure F5). Numerous small streams along Knik and Turnagain arms support relatively small runs of all five salmon species.
Indices for upper Cook Inlet since the early 1970s show general increases in sockeye and coho salmon return abundances, an odd/even year cycle in pink salmon abundances, and a decline in chum salmon abundances (Figure F6). Sockeye salmon run sizes, indexed as catches and escapements into major river systems, increased primarily due to larger returns to the Kenai and Kasilof rivers. Pink, coho, and chum salmon indices, derived from test fishery catches, provide temporal trends, but give only an order of magnitude indication of abundances. Mark- recapture abundance estimates for coho and chum salmon are more accurate, but are only available for 2002. Although commercial drift gillnet catch per unit effort is based on harvests by several hundred boats and test fishery estimates are based on catches of a single boat, these indices show similar trends (Figure F6).
Commercial salmon catches in northern Cook Inlet (above the Forelands), where belugas have concentrated in recent years, were relatively low in the late 1960s and early 1970s, relatively high in the 1980s, and have subsequently declined (Figure F7). This catch decline is partly attributed to fisheries management constraints on fishing effort in order to increase escapements of primarily Chinook, sockeye, and coho salmon. Although salmon returns to the major river systems of northern Cook Inlet have exhibited broad swings in return abundance, many stocks and systems have shown declines in recent years. Sonar estimates of total salmon entering the Yentna River (a Susitna River tributary) ranged from about 0.4 to 1.6 million fish, with no clear temporal trend during 1982 to 2009 (Figure F8). However, the contribution of most species to fish wheel catches in the Yentna River declined as the run was increasingly comprised of pink salmon after 2005 (Figure F8). Chinook and coho salmon weir counts on the Deshka River (a major tributary of Susitna River) and coho salmon weir counts on Little Susitna River peaked in 2004 and have since declined (Figure F9). Sockeye salmon weir counts on Fish Creek (Knik Arm) have been weak in some recent years, but the 2010 weir count was the highest since 1985 before declining dramatically in 2011 and 2012. Coho salmon entering Jim Creek (Knik Arm) increased from the late 1990s to 2006, but have decreased since 2008 (S. Ivey, ADF&G, pers. comm.; Figure F9).
An important concern is that salmon are an essential feature of CI beluga critical habitat, and some species of salmon, most notably Chinook, have had reductions in run strength in Cook Inlet and throughout Alaska. Responding to a request from Alaska Governor Sean Parnell, Acting U.S. Secretary of Commerce Rebecca Blank determined that commercial fishery failures due to fishery resource disasters had occurred for Chinook salmon stocks in the Yukon (2010, 2011, 2012), Kuskokwim (2011, 2012), and Cook Inlet (2012) regions. 39 The declaration acknowledged hardships for commercial, sport, and subsistence users as a result of the Chinook fishery failures. To identify key knowledge gaps and discuss how best to address those gaps, ADF&G sponsored a Chinook salmon symposium, “Understanding the Abundance and Productivity Trends of Chinook Salmon in Alaska,” in Anchorage during October 22–23, 2012. 40 Subsequently, ADF&G worked collaboratively with federal agencies and academic partners to develop a stock assessment and research plan with recommended studies to address critical knowledge gaps (ADF&G Salmon Research Team 2013).
Northern pike were not found in any Cook Inlet streams until being illegally introduced in the 1960s. The spatial distribution of pike has since expanded to include many northern Cook Inlet streams and lakes. In the Susitna watershed, invasive northern pike have impacted many salmonid populations (e.g., Alexander Creek, Shell, and Hewitt lakes) and have largely eliminated salmon from some lakes (e.g., Trapper, Red Shirt, Sucker, and Caswell). The capture of northern pike by commercial salmon fishermen in upper Cook Inlet waters also indicates a potential expansion to other watersheds. Although we do not know to what extent salmon production in Cook Inlet has been impacted by northern pike, pike have clearly reduced salmon production in some areas.
Prior to 1990, belugas were often found in central and lower Cook Inlet, but it is not known what prey were consumed in these areas. In the 1970s, Kamishak Bay supported large commercial catches of Tanner and red king crabs, and summer concentrations of Pacific halibut were found north of Augustine Island (NOAA 1977; Bechtol et al. 2002). While commercial fisheries have not occurred since the early 1980s for red king crab and the early 1990s for Tanner crab, Pacific halibut still support fisheries extending north into central Cook Inlet (Meyer et al. 2008). In spring, Pacific herring aggregate in shallow, nearshore areas of Kamishak Bay to spawn. Peak biomass reached 35,513 short tons in 1983 (Figure F10), declined to 2,906 tons in 2004, and has subsequently ranged from 3,100 to 4,100 tons (Otis and Hammarstrom 2004; Hammarstrom and Ford 2011; Hollowell et al. 2012). Due to low spawning biomass, the commercial herring fishery in lower Cook Inlet has remained closed since 1999. Although herring resources in upper Cook Inlet are not formally assessed, low-level commercial fisheries occur, with annual harvests generally totaling less than 20 tons over the past 15 years (P. Shields, ADF&G, pers. comm.). At Chisik Island, large shallow schools of eulachon, herring, and crangonid and pandalid shrimps were found in May 1997 and 1998, while lower density schools of herring, eulachon, and longfin smelt were found deeper in this area during summer (Fechhelm et al. 1999). Piatt (2002) found cold, nutrient-rich Gulf of Alaska waters upwelling at the entrance to lower Cook Inlet supported high densities of juvenile pollock, sandlance, and capelin. Demersal fish resources in this area were dominated by walleye pollock, Pacific cod, butter sole, and Pacific halibut (Blackburn et al. 1980).
For commercially fished species, the availability of potential beluga prey in upper Cook Inlet during spring and summer can be somewhat inferred from the timing and location of fishery harvests and upriver spawning migrations (also referred to here as “escapements”). However, actual quantitative data on the spatial and temporal distribution of these beluga prey in upper Cook Inlet are limited. For example, long-term salmon escapement estimates are available for the three large middle Inlet rivers, the Kenai, Kasilof, and Crescent river systems, and for the Yentna River, a tributary of the Susitna River, with less frequent estimates available for some other Cook Inlet tributaries (Westerman and Willette 2011). Because sockeye salmon returns to the Kenai and Kasilof rivers comprise the largest component of upper Cook Inlet salmon returns, the bulk of fishing pressure by humans occurs south of these two river systems and, thus, “downstream” of the current primary beluga summer habitat. While more salmon are available in the central Cook Inlet areas, few belugas venture into the central Cook Inlet area in most years. Belugas in northern Cook Inlet likely benefit from the tendency of anadromous prey species to be concentrated by shallow water and the time required to transition from salt water to fresh as they enter the stream mouths, which presumably makes these prey easier to capture.
Management of anadromous fish populations in Alaska attempts to constrain harvests to be no greater than the level of surplus production, defined as returning adult salmon in excess of the spawning production needed to maintain productive salmon populations (Quinn and Deriso 1990). In addition to reproductive needs, harvest considerations must include upstream consumptive uses such as recreational and subsistence fisheries (Shields 2010), as well as allowances for natural mortality, which includes predation by beluga whales, bears, and other species. Stock productivity and the level of surplus production are notoriously difficult to predict and estimate accurately due to high annual variation in factors such as freshwater and marine survival. To account for this uncertainty, for targeted species, fisheries are managed with in- season reductions or closures if those fish stocks appear to be weak. However, the potential for overfishing exists annually, and it is unlikely that escapement goals will be met in all tributaries across all years. While corrective management measures are typically implemented in any year following an under-escapement, prediction of future fish returns and managing for optimal harvest of those returns remains uncertain. Thus, while fishery management, on average, should provide sufficient total numbers of prey for belugas, the timing of prey concentration or densities in the river mouths may not be adequate for efficient feeding by belugas. In addition, a fishery would not be reduced or closed if escapement goals are met. But if the escapement goal arrived in a shorter time period (e.g., 30 days instead of 90 days), the benefit of optimal returns to CI beluga energetics may be very different.
A contrasting management situation for beluga prey exists with eulachon, which also return to freshwater to spawn. Although eulachon spawning stocks can be found in numerous central Cook Inlet rivers, human fishing effort occurs primarily in tributaries in Knik and Turnagain arms. Because fishing tends to occur near the river mouths or upriver, this fishing effort often occurs “upstream” of beluga foraging, such that population level effects of overfishing would be reflected by poor spawning escapement and reduced prey availability in subsequent years. Eulachon populations are not assessed or monitored, but ADF&G uses the Statewide Harvest Survey to derive recreational harvest estimates post-season. These estimates are presumed to be somewhat related to eulachon population abundance. If a decline in annual harvests occurs and is suspected of indicating a substantive decline in eulachon abundance, ADF&G may implement more restrictive fishing measures in subsequent years. There had been a sporadic commercial fishery for eulachon since 1978 (taking from 300–100,000 pounds in 1978, 1980, 1998 and 1999; Shields 2005). Based on a concern that a reduction in the availability of eulachon could be detrimental to belugas, NMFS recommended to the Alaska Board of Fisheries that this fishery be discontinued effective beginning in 2000, in part due to the lack of data on the eulachon runs into the Susitna River, and due to the absence of any evaluation of the effect of this fishery on belugas in terms of disturbance/harassment or competition for these fish. Additionally, it was noted: belugas may be heavily dependent on the oil-rich eulachon early in the spring (preceding salmon migrations), the runs are very short in duration, and large eulachon runs may occur in only a few upper Inlet streams. The commercial fishery for eulachon was closed in 2000, but reopened in 2005, under restrictions to hand-operated dip nets in saltwater between the Chuitna River and the Little Susitna River, with a total harvest of 100 tons or less (Shields 2005, Shields and Dupuis 2012; Shields and Dupuis 2016; P. Shields, ADF&G, pers. comm.).
Beluga prey resources, such as salmon and eulachon, typically represent a mixture of spawning stocks that are also harvested in mixed-stock fisheries (Shields 2010; Westerman and Willette 2011; Shields and Dupuis 2016). Effects of overfishing by humans on beluga foraging success are not well known, yet likely include spatial and temporal components for any specific prey resource that is overfished. Stock composition is dynamic and varies annually in both the run strength and run timing of individual contributing stocks. For major stocks or indicator stocks, harvest managers have tried to determine the relationship between annual escapements and returns in subsequent years. These relationships often have an optimal range such that escapement larger or smaller than this range are presumed to generate reduced adult salmon returns in future years. Harvest managers attempt to regulate fishing effort, typically in mixed- stock fisheries, to ensure that spawning escapement goals are achieved for each monitored salmon stock. However, it is not always possible to ensure that all target stocks are under fished, without exceeding the upper bound (over fishing) on some stocks.
Competition for CI Beluga Prey Resources
Over time, selective fishing pressure, or other factors, can alter reproductive migration timing of some prey species. For instance, intensive fishing during the early part of a salmon run can reduce the portion of the stock that returns early in the run and slightly shift future run timing, but the extent of that shift is limited as survival decreases outside of an optimal migration timing (Smoker et al. 1998). Thus, the timing of prey concentration or densities in the river mouths may not be adequate for efficient feeding by belugas. Chronic and persistent overharvesting of one or more unique salmon stocks or stocks from a specific spatial and/or temporal component (e.g., repeated overharvesting of upper Cook Inlet, early season runs) also has the potential to restructure the ecosystem. Such a pattern could cause a shift in beluga foraging toward less- nutritious prey items or a geographic displacement from the optimal foraging habitat, ultimately with reduced survival and reproductive success. However, the time frame over which such shifts could occur is unknown, and no baseline data currently exist to detect such shifts.
Although there is no definitive analysis of competition between CI belugas and other marine mammals that consume the same prey, the possibility of competitive overlap in prey exists. For example, Chinook and coho salmon were found to be prey items for CI belugas (Quakenbush et al. 2015), so that any predator (including humans) that takes these species from stocks used by belugas are potential competitors. Resident (fish-eating) killer whales along the north Gulf Coast of Alaska are known to focus on salmonids, particularly Chinook, chum, and coho salmon (Matkin et al. 2010). These fish-eating resident killer whales are common in lower Cook Inlet and may intercept salmon destined for rivers and streams in the upper Inlet that are potential beluga prey; however, resident killer whales are not known to range into the upper Inlet where they might compete directly with CI belugas for prey. Harbor seals and Steller sea lions are also known salmonid predators that occur within the range of CI belugas and could compete with belugas and each other for these prey. Harbor seals, Steller sea lions, killer whales, humpback whales, gray whales, minke whales, harbor porpoises, sea birds, sea otters, and humans may also have competition effects on belugas through their consumption of eulachon.
The estimated annual rate of increase in sea otters in Kachemak Bay between 2002 and 2008 was 26% per year, exceeding the estimated maximum productivity rate for this species and is presumably due in part to immigration from other areas (Gill et al. 2008). Sea otters have been found as far north as Ninilchik (V. Gill, USFWS, pers. comm.). Systematic surveys have not been done for several years and trends are unknown for Cook Inlet/Shelikof stocks of harbor seals, the Gulf of Alaska stock of harbor porpoise, the Alaska stock of Dall’s porpoise, or the Alaska stock of minke whales (Allen and Angliss 2012). The Eastern North Pacific gray whale stock and both the Western and Central North Pacific stocks of humpback whales have been increasing based on recent abundance estimates (Allen and Angliss 2012). None of these potential competitive effects have been quantified.
Resident killer whales likely do not directly compete for prey resources within the range of CI belugas, given limited to no overlap in their distribution with CI belugas (Lammers et al. 2013). Similarly, sea otters and Steller sea lions are likely not effective competitors with CI belugas, as they overlap with belugas in only a small portion of their range in lower Cook Inlet. While likely not in direct competition for adult salmon, the introduction of northern pike, an invasive species found in freshwaters of northern Cook Inlet, has likely reduced local salmon stocks, particularly Chinook, through predation on juveniles (Oslund and Ivey 2010).
Ecosystem Shifts and CI Beluga Prey
Both the relative and total abundances of any beluga prey item are not constant and can be expected to change over both space and time. Productivity of many marine species, including, but not limited to, potential beluga prey, may have responded to decadal-scale climate shifts in the North Pacific (Hollowed and Wooster 1992; Beamish and Bouillon 1993; Hare and Mantua 2000). Recognized climate regime shifts that occurred around 1976 and 1989 (Anderson and Piatt 1999; Zheng and Kruse 2000; Hare and Mantua 2000; Kruse 2007; Mueter et al. 2007) may have affected the productivity of marine species in the North Pacific, although response to ecological changes can vary temporally by species, with some responding sooner than others, or in different trends, or greater magnitudes (Rodinov and Overland 2005). For example, the northern Gulf of Alaska changed from an ecosystem dominated largely by invertebrate (crabs and shrimps) biomass in the 1960s to 1970s to dominance by gadids and flatfishes. Robards et al. (1999) found a 1,000-fold increase in gadid abundance in lower Cook Inlet between the 1970s and 1990s, and a lesser increase in abundances of pleuronectids and salmonids. Small-mesh trawl surveys in Kachemak Bay documented a decline in pandalid shrimps and an increase in demersal fishes since the 1970s (Figure F11). Walleye pollock, flathead sole, and starry flounder became the dominant demersal fishes, comprising over 40% of the survey catch in 2004 to 2006 (Goldman et al. 2007).
A similar change was observed in small-mesh surveys from Kodiak Island to Pavlof Bay (Anderson and Piatt 1999), with ongoing surveys indicating continued low levels of stock biomass for many potential forage species including shrimp, juvenile pollock, and herring (Figure F12; D. Urban, NMFS, pers. comm.). Eulachon exhibited a resurgence in the 2000s, but declined in 2010, coincident with an increase in commercial harvest. The climate regime shift in the North Pacific during the late 1970s was associated with aspects such as increased ocean temperatures and increased abundances of predatory fishes, such as Pacific cod. A study of the decline in the Kachemak Bay stock of northern shrimp found that a strong increasing trend in natural mortality followed the 1976 to 1977 regime shift, paralleling trends in increased Pacific cod abundance (Fu and Quinn 2000; Fu et al. 2000). A study of red king crab around Kodiak Island attributed the initial population crash to overfishing, but suggested that, despite a fishery closure since 1983, the stock has failed to recover due to increased juvenile mortality associated with higher ocean temperatures and greater abundance of predatory fishes, such as Pacific cod (Bechtol and Kruse 2010). Pacific cod and walleye pollock, while not historically “rare” in Cook Inlet, occurred at much lower levels of biomass and abundance prior to the late 1980s, when recent commercial fisheries developed (Bechtol 1995).
Surveys show biomass of Pacific cod and walleye pollock remained relatively high through the 1990s (Figure F13; R. Gustafson, ADF&G, pers. comm.). Meanwhile, Tanner crab data from lower Cook Inlet indicate dramatic declines in abundance of harvestable crabs after the mid-1970s (Figure F14; Bechtol et al. 2002; R. Gustafson, ADF&G, pers. comm.); these crabs are seasonally important to belugas in upper Cook Inlet.
While ecosystem response to environmental forcing is likely nonlinear (Hare and Mantua 2000), evidence exists for climate-driven changes in the physical environment affecting other fish populations in the Gulf of Alaska and eastern Bering Sea. For example, strong pollock recruitment in the eastern Bering Sea appears connected to above normal air and bottom temperatures and reduced sea ice cover, factors that promote zooplankton production (Quinn and Niebauer 1995). Solid sea ice is not a factor in the northern Gulf of Alaska, but the pre-1976 regime was associated with low sea surface temperature and low biomasses of predatory fishes, such as flatfishes and Pacific cod. During and following the 1976 regime shift, high sea surface temperatures enhanced zooplankton production in the Gulf of Alaska, supporting strong pollock recruitment amid low demersal fish predation (Bailey 2000; Ciannelli et al. 2005). However, high zooplankton populations may have been detrimental to phytoplankton needed for first- feeding larvae of many species. Sea surface temperatures declined somewhat following the regime shift, but ecosystem “maturation” in the subsequent decade resulted in increased biomass of predatory fishes, particularly Pacific halibut, arrowtooth flounder, flathead sole, and Pacific cod (Bailey 2000). The North Pacific ecosystem has been generally characterized by moderate sea surface temperature in recent decades, but relatively high demersal fish biomass (Hare and Mantua 2000; Mueter and Norcross 2002; Ciannelli et al. 2005). As a result, a compromised feeding environment for many larval forage species was coupled with intensified groundfish predation.
A cautionary note is warranted regarding interpretation of the role of long-term environmental effects as drivers of potential ecological change. Ecological systems are complex, and trends in abundance and biomass are typically the result of a variety of factors. A first step in understanding ecosystem change is to have a sufficiently long time series of indices for both potential ecosystem drivers and the species of interest. Unfortunately, these indices are often discontinuous over time or of an inappropriate spatial coverage. Surveys of potential CI beluga prey in marine or estuarine areas of upper Cook Inlet have been infrequent and short-term, typically implemented to address ad hoc environmental assessment needs for resource development. Use of commercial harvests to represent potential CI beluga prey is likely biased because harvests typically occur “downstream” of feeding CI beluga. Use of salmon escapements to represent CI beluga prey is also biased because escapements occur “upstream” of CI beluga foraging areas. In addition, many escapement indices are discontinuous over time as monitoring techniques or tributaries change in response to management priorities and agency budget limitations. The small-mesh trawl survey in Kachemak Bay dates to 1977 and provides a basis for long-term ecosystem changes, but was reduced in frequency and then discontinued after 2006 due to financial priorities. A multi-species trawl survey, focused on Tanner crab, but also providing population estimates of species like Pacific cod, walleye pollock, and arrowtooth flounder in lower Cook Inlet, dates to 1990 but has also been reduced in frequency due to budget priorities.