The ESA (Section 3) defines "endangered species" as "any species which is in danger of extinction throughout all or a significant portion of its range." "Threatened species" is defined as "any species which is likely to become an endangered species within the foreseeable future throughout all or a significant portion of its range." NMFS considers a variety of information in evaluating the level of risk faced by a DPS, including: 1) absolute numbers of fish and their spatial and temporal distributions, 2) current abundance in relation to historical abundance and carrying capacity of the habitat, 3) trends in abundance, based on indices such catch statistics, CPUE, and spawner-recruit ratios, 4) natural and human-influenced factors that cause variability in survival and abundance, 5) possible threats to genetic integrity (e.g., selective fisheries and interactions between cultured and natural populations), and 6) recent events (e.g., climate change and changes in management) that have predictable short-term consequences for the abundance of a DPS. Additional risk factors, such as disease prevalence or changes in life-history traits, also may be considered in the evaluation of risk to a population. We briefly describe these six considerations as follows, then proceed to a detailed examination of available information for herring in the Georgia Basin DPS.
The determination of whether a species is threatened or endangered, according to the ESA, should be based on the best scientific information available, after taking into consideration conservation measures that are proposed or in place. The BRT did not evaluate likely or possible effects of conservation measures. Therefore, they did not make recommendations on whether DPSs should be listed as threatened or endangered species, because that determination requires evaluation of factors not considered by the BRT. However, the BRT did draw scientific conclusions about the risk of extinction faced by DPSs under the assumption that present conditions will continue, recognizing that natural demographical and environmental variability is an inherent feature of present conditions. Conservation measures will be taken into account by the NMFS Northwest Regional Offices in making listing recommendations. The following sections summarize the kinds of information the BRT considered in evaluating the potential effects of risk factors on the each of the DPSs identified by the BRT.
The absolute number of individuals in a population is important in assessing two aspects of extinction risk. First, a small population, even one that appears stable or increasing, may not be able to sustain itself in the face of environmental fluctuations and small-population stochasticity. This conclusion follows from the theory of minimum viable populations (MVP) (see Gilpin and Soulé 1986, Thompson 1991). Second, present abundance in a declining population is an indicator of the time expected until the population reaches critically low numbers. This follows from the idea of "driven extinction" (Caughley 1994). In addition to absolute numbers, the spatial and temporal distributions of adults are important in assessing risk to a DPS. Spatial distribution is important, both at the scale of the spawning population and the metapopulation.
Assessments of marine fish populations have focused on the biomass or numbers of adults harvested by commercial and sports fishing. Catch records, catch-per-unit effort (CPUE), and biomass estimates from research cruises constitute most of the data available to estimate abundance trends. However, the numbers of reproductive adults is the most important measure of abundance in assess the status of a population. Data on other life-history stages can be used as a supplemental indicator of abundance.
The relationship of present abundance to present carrying capacity is important for evaluating the health of a population, but a population with abundance near the carrying capacity of the habitat it occupies does not necessarily indicate that the population is healthy. Population abundances near carrying capacity imply that the effectiveness of short-term management actions is limited in increasing population abundance. The relationship between current abundance and habitat capacity to the historical relationship between these variables is an important consideration in evaluating risk. An understanding of historical conditions provides a perspective of the conditions under which present populations evolved. Estimates of historical abundances also provide the basis for establishing long-term abundance trends. Comparisons of past and present habitat capacity can also indicate long-term population trends and potential problems stemming from population fragmentation.
Short- and long-term trends in abundance are primary indicators of risk in natural populations. Trends may be calculated with a variety of quantitative data, including catch, CPUE, and survey data. Spawning-biomass estimates are available for Puget Sound populations of herring, but uninterrupted time-series of adequate length for trend analysis are only available for the six most important populations. The influence of environmental variability on population abundances also limits the interpretation of short-term trends, because the climate changes in the late-1970s and 1980s coincided with apparent declines in population abundances for the species being considered in this review.
Several natural and anthropogenic factors influence the degrees of risk facing populations of marine fish in Puget Sound. Recent changes in these factors may influence the degree of risk of a population without apparent changes in abundance, because of time lags between the events and the effects on the population. Thus, a consideration of these effects extends beyond the examination of recent trends in abundance. The BRT considered documented physical and climatic changes, but did not consider possible effects of recent or proposed conservation measures. Population variability in itself may not be an indication of risk. Habitat degradation and harvest have most likely weakened the resilience of populations in Puget Sound to climate variability. However, these effects are not easily quantified. Manmade contaminants and predation by marine mammals are additional factors that may influence herring mortality and lead to changes in abundance.
Artificial propagation of herring populations does not occur in Puget Sound so is not a risk factor for the species considered here. However, mariculture of some species is under development, and the effects of hatchery releases on natural populations may be important in the future. The interbreeding of cultured and natural fish can potentially lead to a loss in fitness of naturally-spawning populations. The genetic effects of artificially-propagated releases of species with high fecundities, as is common for many marine fishes, could be substantial. Ryman and Lairkre (1991), Waples and Do (1994), and Ryman et al. (1995) discussed possible risks associated with enhancement of marine populations, but these risks are difficult to quantify and to incorporate into risk analysis. The chief concern is that the release of propagated fish, which may be inadvertently modified by breeding practices and novel rearing environments, may lead to the erosion of genetic diversity and fitness in natural populations.
Human activities, other than population enhancement, can also influence the genetic characteristics of natural populations. These include size-selective harvest methods (Nelson and Soulé 1987); introductions of non-native species; and alterations of marine habitats by shoreline development, increased siltation in river runoff, and pollution. At the present time, empirical information documenting the genetic effects of these kinds of changes is largely lacking.
Coupled changes in climate and ocean conditions have occurred on several different time scales and have influenced the geographical distributions, and hence local abundance, of marine fishes. On time scales of hundreds of millennia, periodic cooling produced several glaciations in the Pleistocene Epoch (Imbrie et al. 1984, Bond et al. 1993). The central part of Puget Sound was covered with ice about 1 km thick during the last glacial maximum about 14,000 years ago (Thorson 1980). Since the end of this major period of cooling, several population oscillations of pelagic fishes, such as anchovies and sardines, have been noted on the west coast of North America (Baumgartner et al. 1992). These oscillations, with periods of about 100 years, have presumably occurred in response to climatic variability. On decadal time scales, climatic variability in the North Pacific and North Atlantic Oceans has influenced the abundances and distributions of widespread species, including several species of Pacific salmon (Francis et al. 1998, Mantua et al. 1997) in the North Pacific, and Atlantic herring (Alheit and Hagen 1997) and Atlantic cod (Swain 1999) in the North Atlantic. Recent declines in marine fish populations in Puget Sound may reflect recent climatic shifts (Fig. 4). However, we do not know whether these climatic shifts represent long-term changes or short-term fluctuations that may reverse in the near future. Although recent climatic conditions appear to be within the range of historical conditions, the risks associated with climatic changes may be exacerbated by human activities (Lawson 1993).
One of the greatest difficulties in the status review process is organizing the large amount of information regarding the biology of the species, genetics, and population trends over time. Often, the ability to measure or document risk factors is limited, and information is not quantitative and is very often lacking altogether. In assessing risk, it is often important to include both qualitative and quantitative information, and the method by which a BRT can do so takes several forms. In the next section, information to assist in assessing risk is presented in the format of the types of methods used in the BRT deliberations. The first is a presentation of risk factors discussed in West (1997). This is a qualitative discussion that presents and references important pieces of information, but does not attempt to make a quantitative assessment of these factors. The second method for assessing risk is that of Wainwright and Kope (1999). This method was used in the Pacific salmon BRT process and provides a method to organize and summarize the professional judgement of a panel of knowledgeable scientists. It is a risk matrix approach and includes information on abundance, population trends, productivity and variability, genetic integrity and habitat condition/capacity. Another approach recently presented by Musick et al. (2000), provides criteria by which to define risk. These criteria are based on productivity measures such as intrinsic rates of increase, age at maturity and maximum age. These criteria are similar to those examined in Wainwright and Kope, however, they are organized somewhat differently. The criteria are rarity, small range and endemics, specialized habitat requirements and population decline. Decline thresholds are based on population resilience of the species. This method provides another method to examine and organize available information for the evaluation of risk and an opportunity to compare the results of the methods.
Major risks to the survival of Pacific herring include overharvesting, predation by pinnipeds, birds and fish, adverse climatic conditions, loss or degradation of habitat, and pollution-related effects as identified in West (1997). Each of these risk factors will be described in the following sections.
Most human harvest of Pacific herring since about 1970 has been for the lucrative roe market. Policies on human harvest of herring for roe are similar for California, Washington, British Columbia, and Alaska. No harvest is allowed below specified thresholds. When mature biomass is above a threshold, allowed exploitation rates are gradually increased to a maximum of 20%. The level of exploitation is thought to be sufficiently low to protect the resource. Minor bait and subsistence harvests are also allowed, and Washington does not control tribal fisheries which have been very minor in recent years. The tribes cooperate with WDFW in management of their harvest. Much larger harvests were allowed prior to 1970. In recent years, however, Washington has not allowed a sac roe fishery because of low biomass in the Cherry Point population. Harvests in recent years appear to be much lower than the amount consumed by natural predators.
The WDFW fishing regulations have reduced human exploitation rates to low levels, however, pinniped exploitation of herring may have increased. Herring-survey biologists have observed increased nocturnal occurrences of harbor seals near schools of herring and concurrent changes in herring schooling behavior (West 1997).
Two species of pinnipeds, California sea lion (Zalophus californianus) and Pacific harbor seal (Phoca vitulina), that are common in Puget Sound and British Columbia exploit herring. Schmitt et al. (1995) estimated that herring comprised 6% of the diet of California sea lions in Puget Sound during the 1986-1994 period. Total fish consumption by California sea lions in Puget Sound was estimated to be 830 mt per year (NMFS 1997). Thus about 50 mt (0.06 X 830) of herring were consumed by California sea lions per year in Puget Sound during the 1986-1994 period. Large aggregations of California sea lions were not reported in Puget Sound until 1979 (Schmitt et al. 1995), numbers increased through 1986, and then fluctuated without trend (Schmitt et al. 1995, and J. Laake6). California sea lions also occur in British Columbia waters but estimates of their consumption of herring were not available.
More data are available on Pacific harbor seals in British Columbia waters than for Puget Sound. Olesiuk et al. (1990) estimated that harbor seals consumed 3,206 mt of herring in the Canadian Strait of Georgia (CSG) during 1988, which represented 9.6% of the herring spawning biomass (Table 5). They estimated that herring comprised 32.4% of harbor seal diet in 1988. NMFS (1997) estimated that harbor seals consumed 14,997 mt of prey in Washington inland waters during 1993. If herring comprised 32.4% of the diet of Washington harbor seals, they would have consumed 4,859 mt (5,356 tons) of herring in Washington’s inland waters in 1993, which represented 34.9% of estimates of spawning biomass of herring for 1993 (Table 6). While herring biomass was not estimated for all areas and harbor seals consume immature as well as mature herring, it appears that harbor seals could be a significant source of mortality for Washington populations of herring and could account for some of the increases estimated for non-fishing mortality (Bargmann 1998). NMFS (1997) estimated that harbor seals in Washington waters increased at 7.7% annually between 1978 and 1993. They did not provide rate of increase for inland waters alone. Herring are also reported to be an important prey item for harbor porpoises in the Strait of Juan de Fuca (Gearin et al. 1994).
Forage fish, including herring, are prey for some birds. Fish-eating species documented for inland waters of Washington and/or British Columbia included red-throated loon, Pacific loon, common loon, horned grebe, eared grebe, pied-billed grebe, red-necked grebe, western grebe, double-crested cormorant, Brandt’s cormorant, pelagic cormorant, great-blue heron, common merganser, red-breasted merganser, Caspian tern, common murre, pigeon guillemot, rhinoceros auklet, and tufted puffin (Mahaffy et al. 1994). Estimates of herring consumption by these birds were not available. Palsson (1984) studied the consumption of herring eggs by birds in Puget Sound. He found 14 species of ducks and gulls feeding over spawning beds. The most important egg-predators were surf scoters, white-winged scoters, glaucous-winged gulls, and Bonaparte’s gulls. The birds were mainly attracted to areas with high egg densities and bird densities dropped soon after consumption reduced the abundance of eggs in an area. It is difficult to estimate the impact of birds on egg mortality because there are other sources of high egg mortality, such as suffocation when there are multiple layers of eggs and predation by snails and amphipods that can occur if the eggs are not first consumed by birds (Palsson 1984).
Herring are estimated to comprise 71% of lingcod, 62% of chinook salmon, 58% of coho salmon, 53% of Pacific halibut, 42% of Pacific cod, 32% of Pacific hake, 18% of sablefish, and 12% of dogfish diets off the west coast of Vancouver Island (Environment Canada 1998). Pacific hake and cod abundance in Puget Sound decreased in recent years (Gustafson et al. 2000), but West (1997) expressed the concern that increased abundance of hake in offshore waters may be affecting herring. The proportion of offshore hake that feeds off British Columbia during summer months is directly related to ocean temperatures, which tended to be higher than normal between 1976 and 1998. West (1997) also noted that spiny dogfish apparently increased in Puget Sound since 1985, and that WDFW increased releases of yearling chinook salmon from low levels in the early-1970s to more than 3,000,000 fish in most years since 1975(WDF 1993). Chinook salmon released as yearlings tend to remain in Puget Sound and could account for some of the increased non-fishing mortality estimated for Puget Sound herring, but much of the increase in mortality occurred after yearling releases had increased to present levels.
In summary, herring are important diet components for many predators. However, time series of estimates of total herring consumption by predators have not been made. While it seems safe to conclude that predation is a major source of mortality for herring, it is not possible to determine what portion of the apparent increase in non-fishing mortality of herring is due to changes in predation intensity.
The decline in some of the Puget Sound herring populations coincided with warm/dry conditions in the Pacific Northwest (Fig. 4). Similar conditions occurred during the 1930s when Chapman et al. (1941) reported that the Discovery Bay and Cherry Point herring populations were at low levels. Conditions changed to cold/wet or average during the 1940s and 1950s (Fig. 4), and by 1959 Williams (1959) reported that the Cherry Point and Discovery Bay populations had "regained their productive levels after many years without exploitation". The Cherry Point and Discovery Bay populations experienced the steepest declines observed in the Washington populations since 1986 (see Table 11). EVS (EVS Environment Consultants 1999) found significant negative correlations between biomass of the Cherry Point population and annual sea surface temperatures at British Columbia light houses at Active Pass (in Gulf Islands near Cherry Point), Race Rocks (Strait of Juan de Fuca), and Amphitrite Point (west coast of Vancouver Island). The correlations ranged from -0.64, Race Rocks, to -0.75, Amphitrite Point, which is the farthest from Cherry Point.
The Cherry Point population is thought to migrate to off the coasts of Washington and southern British Columbia during summer months (Lemberg et al. 1988). Discovery Bay is on the Strait of Juan de Fuca and thus closer to the outer coast than populations in Puget Sound. Herring year-class strength of the west coast of Vancouver Island population averaged twice as large in cool years than in warm years (Canada Fisheries and Oceans 1998). Tanasichuk (1997) found that length-at-age of herring of the southwest coast of Vancouver Island was negatively related to sea temperature during the first growing season.
The observed correlations with temperature could be caused by increased predation and/or competition resulting from increased populations of hake and Pacific sardines off the west coast of Vancouver Island during warm years or adverse direct effects of high temperature at some life stage(s). EVS (1999) reviewed the literature on temperature tolerances of herring eggs and larvae and found that "the preponderance of data in the scientific literature suggest that the upper threshold for effects is between 13 and 14oC". They collected temperature data at 12 stations in the Cherry Point area between May 14 and June 9, 1998. Station averages ranged from 11.1 and 11.9oC during the first half of the study and 11.5 and 12.1oC during the second half. The maximum observed temperature in June was 16.8oC. Environment Canada (1999) also summarized surface temperature data collected once per month by the Washington State Department of Ecology between 1978 and 1997 at a station in the Strait of Georgia (16 km west of the Cherry Point area) and a station in Bellingham Bay (13 km south of the Cherry Point area). May and June temperatures exceeded 13oC on several occasions. While data are limited, in recent years the observed temperatures at and near Cherry Point during the time of spawning and early larval stages are close to or above the upper tolerance reported in the literature suggesting that temperature related mortalities could have occurred in early life stages of Cherry Point herring. Because juvenile and adult herring are very mobile, it seems likely that they could avoid excessive temperatures.
Because herring spawn at shallow depths in nearshore waters, their spawning grounds are particularly vulnerable to human disturbance. Chapman et al. (1941) reported that herring had spawned in Port Hadlock, but no longer spawned there by 1936. They quoted local inhabitants, who "claimed that no fish have spawned in that vicinity since the digging of the ship canal between Marrowstone Island and the mainland." Herring no longer spawn in Nanaimo Harbor and the adjacent Newcastle Channel and Ladysmith Harbor, which have been heavily affected by human activities (Environment Canada 1998). Herring also have not returned since 1977 to spawn in Pender Harbor, which has had considerable waterfront residential growth (Environment Canada 1998). There are three industrial piers in the Cherry Point spawning area. The last one was completed in 1971 (EVS 1999). The piers have had modest impacts on the spawning habitat through light shading, current modifications, and wave shading (EVS 1999). There is evidence of some changes in vegetation in the area (EVS 1999).
Pacific herring almost always spawn on aquatic vegetation. Three species of plants (Zostera marina, Desmarestia spp., and Odonthalia spp.) appear to be preferred as herring spawning substrate over other plants at Cherry Point (Stick 1995 cited in EVS 1999). The percent frequency used for spawning was 58.0% for Zostera marina, 46.9% for Desmarestia spp., and 6.8% for Odonthalia spp., while their frequency of occurrence in samples was 53.6%, 30.4%, and 6.6% respectively. These data suggest that Desmarestia spp. was preferred over Zostera marina at this site. Pacific herring also sometimes spawn on gravel, rocks, and human- made structures. There is little aquatic vegetation in Squaxin Pass and the Port Susan areas and herring often spawn on rocks and gravel there (Lemberg 1997). The introduced Zostera japonica was not listed, and it is not clear that the study distinguished between the two species of eelgrass. Zostera japonica does not function well as a spawning substrate for herring, because it is an annual plant and beds are not established until after the spawning season of winter spawners. Because of this there is concern about its spread in Puget Sound (West 1997). While the Cherry Point data suggest that Desmarestia spp. was preferred over Zostera marina, WDFW staff believe that Zostera generally is a preferred spawning substrate throughout Puget Sound and that subtidal beds of Gracilaria are also important (Koenings, unpubl. data). The Washington Departments of Natural Resources and Fish and Wildlife have a policy of no net loss of vegetated habitat, but "many Puget Sound researchers and managers think that significant anthropogenic loss of eelgrass continues" (West 1997). There is inadequate information to assess the cumulative loss of eelgrass or habitat in Puget Sound during historical times. Prior losses of intertidal habitats such as are documented in Hutchinson (1988) may be a factor in the declines of herring populations due to ecosystem effects. However, many of these losses are historic (prior to 1950) and predate the recent declines in some herring stocks within the Georgia Basin.
Herring in Puget Sound, the Strait of Juan de Fuca, the southern Strait of Georgia, and elsewhere in British Columbia, have been potentially exposed to a variety of anthropogenic pollutants at various sites during the past century. Both the local and region_wide understanding of contaminant patterns, injuries, and trends are relevant because herring: 1) have local spawning sites in proximity to known pollution sources, 2) spawn at sites scattered across a wide range of urban, industrial and rural locations, and 3) are migratory, thus they may be exposed to potentially_contaminated water and food over large areas.
Contaminants of most concern include synthetic chlorinated organic chemicals, such as: 1,1,1-trichloro-2,2-bis-(p-chlorophenyl)ethane and related compounds (DDTs) and the polychlorinated biphenyls (PCBs); polycyclic aromatic hydrocarbons (PAHs) from petroleum and hydrocarbon combustion; dioxins and a host of other organic compounds; trace elements, such as mercury and lead; and organic matter and nutrients, such as nitrogen. Documented sources of contaminants have included: industrial discharges, pulp and paper mills, oil spills, sewage treatment plant discharges, runoff from urban and agriculture area, and coal mining and processing. Local and onshore sources also include creosoted pilings and seepage from hazardous waste sites. Poorly-documented sources of contaminant include: atmospheric deposition and advection of marine water from other locations. Major efforts to stem, control or re-direct pollution began in the 1940s, following World War II industrial activity. By the early-1970s there were major reductions in pollutant loads discharged by the wood and paper industries (Dexter et al. 1985). By the late-1970s there were major improvements in municipal wastewater treatment, effective industrial pre-treatment or source control, and the onset of clean-up of hazardous waste sites. However, urban and suburban development, highway construction and increased traffic continue.
Herring can be exposed to contaminants through all stages in their life cycle. Developing eggs and larvae can be exposed to contaminants in water and in the surface micro-layer. Juveniles, sub-adults, and adults that migrate through urban waters, may accumulate persistent organic chemicals though feeding. These bio-accumulated contaminants may be passed to their gametes as well as their predators (larger fish, mammals and birds).
The near-shore, shallow-water spawning habitat of herring makes them particularly vulnerable to exposure to contaminants from such sources as oil spills, urban and agriculture runoff and chronic air pollution. Herring eggs can be exposed to contaminants during embryonic development. Herring at Cherry Point, one of 18 spawning sites in the Puget Sound/Straits region, has been studied in detail for possible problems with contaminants. The ARCO refinery experienced three minor discharge permit violations (EVS 1999): fecal coliform bacteria, biochemical oxygen demand (BOD), and oil and grease. In the Ferndale refinery effluent, concentrations of mercury, copper, and lead have been problematic, but probably not of particular concern after dilution by mixing (EVS 1999). In the effluent from the Intalco refinery, concentrations of nickel, phenanthrene and fluoranthene have also exceeded standards but mixing would reduce concentrations below levels of concern (EVS 1999). Studies of the receiving waters indicate no evidence of contaminants above current standards (EVS 1999). However, standards for PAHs may not be appropriate and concentrations of PAHs may have been above 1 ppb, a concentration that caused problems in herring larvae hatched from eggs that were exposed to weathered Alaska North Slope crude oil (Carls et al. 1999, Brown et al. 1996, Kocan et al. 1996, Laur and Haldorson 1996).
Some scientists are becoming increasingly concerned about the effects of endocrine_disrupting chemicals, such as PCBs and PAHs on animals (Vos et al. 2000). Sediments near the ARCO refinery were found to be contaminated with chemicals similar to those found in coal tar epoxy, which was used to protect pier pilings (EVS 1999). The Intalco facility was ranked third highest of 49 sites in Puget Sound with sediment contamination (excluding Superfund sites) and was the highest ranking site not undergoing cleanup (EVS 1999). Two of four studies (conducted in 1990-1992, and 1998) found either lower-hatching success or a reduction in the percent of normal larvae for herring eggs collected from Cherry Point sites compared to controls (EVS 1999). The ARCO and Tosco Cherry Point facilities have reported 73 oil spills ranging from sheens reported on 12 occasions to 21,000 gallons spilled on 6/4/72 (EVS 1999).
Exposure to PAHs by adult male herring collected during spawning from five spawning sites in Puget Sound and the Strait of Georgia in 1995, 1999 and/or 2000 was evaluated by West (J. E. West7). Fish are able to extensively biotransform PAHs to more polar metabolic products, most of which are readily excreted into the bile (Varanasi et al. 1989). Therefore, exposure of male herring to PAHs was estimated by measuring fluorescent aromatic compounds (FACs) in bile. Concentrations of biliary FACs were reported as benzo[a]pyrene (BaP) equivalents, representing the high molecular weight aromatic hydrocarbons (HAHs); or as naphthalene (NPH) and phenanthrene (PHN) equivalents, representing molecular low weight aromatic hydrocarbons (LAHs), on a mg biliary protein basis. Sources of LAHs include all fossil fuels, as well as crude oil. The HAHs are also present in crude oil and fossil fuels, and another important source is combustion residue (e.g., soot) from incomplete combustion processes with fossil fuels, including natural processes, such as forest fires. The spawning sites in North Puget Sound from which herring were collected and the years in which they were sampled included: Semiahmoo Bay (1999 and 2000), Cherry Point (1999), and Fidalgo Bay (1995, data not presented here). The Puget Sound proper spawning sites from which herring were collected and the years in which they were sampled included: Port Orchard (1999 and 2000) and Johnson Point (1999 and 2000). A spawning site near Nanaimo, B.C. was sampled in 1999.
In bile samples analyzed in 1999, the highest mean concentration of BaP equivalents was found in the bile of herring from Port Orchard (Fig. 30). This result is not unexpected because of this site’s proximity to Elliott Bay. McCain et al. (2000) reported that concentrations of HAHs in sediment and BaP equivalents from the bile of English sole (Pleuronectes vetulus) from a site in Elliott Bay were among the highest found on the West Coast. Concentrations of NPH equivalents also tended to be higher in the bile of herring the Port Orchard site; however, the mean concentration was not significantly different from the Cherry Point site (Fig. 30). This higher concentration at the Cherry Point site may be the result of oil refinery-related activities at this site which could yield LAHs (EVS 1999). For samples collected in 2000, mean concentrations of BaP, NPH, and PHN equivalents tended to be higher in herring from the spawning sites in Puget Sound proper, but the differences from the Semiahmoo Bay site were not significant.
Puget Sound-wide contaminant investigations have been recently conducted at sites either with or without a history of pollution or near known point sources. Some studies have been done near a few of these herring spawning areas. For example, total PAH concentrations in Sound and Straits sediments in 1998 ranged from less than 100 ppb dry weight (dw) at sites in the Straits of Georgia, Discovery Bay, Dash Point and Port Townsend to 11,000 ppb dw in the Duwamish River (PSWQAT, 1998). In this survey, sediments at sites at Birch Point, north of Cherry Point, appeared to contain on the order of 300 ppb dw. Similar values (ca 300 ppb) were reported for a site in Port Madison and a value of over 3,000 ppb was reported for Dyes Inlet. This region (Dyes Inlet to Port Madison) hosts many kilometers of herring spawning beaches such that if PAH's are affecting herring, the Port Orchard area spawning populations should be affected.
Relatively high levels of PCBs were found in herring from spawning sites in Puget Sound proper. Whole bodies of male herring were used for analyses of PCBs. Data from analyses for PCBs are presently only available for herring collected in 1999. Concentrations of PCBs in composites of homogenized whole bodies (10 composites per site, 5 fish per composite) were consistently higher in herring from the Port Orchard and Johnson Point spawning sites compared to sites in North Puget Sound and the Strait of Georgia (Fig. 31). McCain et al. (2000) also reported concentrations of PCBs in the sediments and in the livers of English sole from Elliott Bay to be among the highest on the West Coast. The proximity of Port Orchard to Elliott Bay and Johnson Point to Olympia (another urban center in Puget Sound) may account for these higher levels of PCBs.
In the 1980s the upper layer (neuston layer) of Puget Sound waters contained concentrations of PCBs, PAHs, and metals exceeding EPA standards by orders of magnitude (Hardy and Antrim 1988). Herring eggs deposited in the inter_tidal zone are exposed to upper water surface and larvae feed on organisms that at least partially inhabit the neuston layer.
Herring populations in several areas could have been affected by contaminants in the past (prior to the 1970s), such as pulp mills discharged large volumes of primary treated effluent (including those in Port Susan, in Sinclair Inlet, near Port Townsend and in southern Puget Sound). There have been other identifiable and pollution events such as oil spills in Discovery Bay, LaConner, Port Susan, Manchester, Washington Narrows (Port Orchard), Winslow, and Budd Inlet (Dexter et al. 1985) and at Anacortes and other localities (A. J. Mearns8). Therefore, we cannot conclude that other populations are not, or have not been, at risk from chemical contaminants similar to those at Cherry Point.
Herring migrate throughout the Puget Sound/Straits region and therefore can encounter pollutants at locations other than spawning sites. Region_wide contaminant monitoring has been spotty, both geographically and over time. However, recent and historical surveys generally support several regional patterns and trends of contamination. Concentrations of organic chemical contaminants in water, sediments, shellfish and fish have been high in sub_areas of the Main Basin surrounding Seattle and Tacoma and considerably lower in the Straits of Juan de Fuca and Georgia, north to the international boundary. In 1997 and 1998, concentrations of selected contaminants were measured in mussels from 18 marine sites in Washington as part of NOAA’s Mussel Watch Program (Mearns et al. 1999) (Fig. 32). Mean concentrations of total PAHs (TPAHs) ranged from 174 ppb dw at Cape Flattery to nearly 33,000 ppb at Four Mile Rock in Seattle (Fig. 33). Mean concentrations of TPAHs in mussels from Puget Sound proper ranged from 681 to 33,000 ppm while those from sites near Bellingham and Point Roberts were about 4,000 and 1,200 ppb, respectively. Salazar (M. Salazar9) reported that composites of mussels from Cherry Point contained between 300 and 400 ppb TPAHs. This site is about halfway between Bellingham and Point Roberts and thus actually appears low compared to surrounding areas.
Mussel Watch sites near Port Townsend, in Hood Canal and at the south end of Vashon Island were located near known herring spawning sites. These sites are: Kilisut Harbor, Port Gamble, Quartermaster Harbor, respectively. Mussels from all three of these sites had TPAH concentrations comparable to or greater than the sites in the northern Strait (Fig. 33). The extremely high mean concentrations of TPAHs in mussels from Port Townsend (8,861 ppb) is of considerable interest because the site is located within a few meters of an active sand lance spawning beach and within 1_4 km of Kilisut Harbor. Mussels at the site in northern Hood Canal contained unusually high mean concentrations of TPAHs (1,836 ppb dw). This site is at the "old" ferry landing which is 3_4 km south of the Hood Canal Bridge and across the canal from the herring spawning sites in and outside Port Gamble. Although, it is possible Hood Canal Bridge traffic contributes PAH's to this region.
Similar patterns were observed for PCBs, with higher concentrations in mussels from the central Puget Sound area (123 to 533 ppb dw), Sinclair inlet (233 ppb) and Port Townsend (147 ppb), compared with the mean concentrations in mussels from the two sites in the northern Strait of Juan de Fuca (24 and 38 ppb dw) (Fig. 34).
About 38% of Puget Sound proper has been identified as contaminated above state standards, evidence for potential adverse effects on marine resources. Concentrations of number of heavy metals; chlorinated pesticides, such as DDTs, chlordanes, and dieldrin; dioxins and furans; PCBs; and PAHs in sediments are monitored regularly as part of the Puget Sound Ambient Monitoring Program (PSAMP), operated by the Puget Sound Water Quality Action Team, and have also been measured by Federal agencies such as NOAA and EPA. Currently, over 15,000 acres of intertidal and subtidal lands have been surveyed in urban embayments of Puget Sound as part of the PSAMP (PSWQAT 2000). Most contaminated sediments are located in industrialized areas of the Sound, such as Elliott Bay, Commencement Bay, Sinclair Inlet, and Everett Harbor. Some contaminants of concern (e.g., PCBs, DDTs) are no longer being released into the environment, and levels are gradually declining, but high-residual concentrations are still found in many industrialized areas. Others, such as PAHs, are still being released, especially through non-point sources, and do not appear to have declined greatly over the last 20 years.
Dated sediment cores taken from the middle of the Main Basin of Puget Sound in the late-1970s and again in 1990, clearly document a Main Basin_wide rise of contamination during the period 1920 to 1960 and then decreasing contamination to the present (Lefkovitz et al. 1997). For example, the cores reflect that inputs of PCBs, PAHs, DDTs, chemicals from mills, and several trace elements (mercury, arsenic, lead) increased through the 1950s and 1960s and then progressively decreased through the 1980s. Older mass emission data, water column toxicity data and anecdotal information, confirm the past heavy input of contaminants into herring spawning regions such as the Port Susan complex (Dexter et al. 1985).
Fifteen years (1986_2000) of more recent data on contaminants in mussels, coupled with some 25_year-old data, show a variety of geographic scales, patterns, processes, and response times, some of which are rather surprising. For example, mussels within Puget Sound and the Straits are depleted in several metals (arsenic and cadmium) compared to those from open coastal sites of the Pacific Northwest (Mearns, see Footnote 8). For most metals, there are no "hot spots" or long_term trends, regardless of trends in inputs and discharges. Both cores and mussels from the Sound confirm that PCB concentrations and, presumably, inputs, have been declining over the past three to four decades. Compared to mussels from other Pacific and U.S. coastal areas, Puget Sound mussels contain comparable concentrations of PCBs, but incomparably high concentrations of PAHs.
These trends are reflected in fish and shellfish surveys and monitoring data from specific sites. For example, by 1975, PCB concentrations in flounder and sole in the Duwamish water waterway were declining at rates comparable to those seen in deep-basin sediment cores. Recent (1997_2000) surveys indicate that sediments along the coast north of Admiralty Inlet, and in Ports Susan are largely uncontaminated and non-toxic and support robust benthic assemblages (Long et al. 1999).
Counter to these trends is continued chemical contamination and biological effects of chemicals, including diseased bottomfish populations, within and around historic contaminant hot spots. These include such inshore areas of Commencement Bay, the Duwamish Waterway in Seattle, and a creosote_contaminated site at Eagle Harbor on Bainbridge Island. Focused studies at "pollutant bottlenecks" show that juvenile chinook salmon migrating through still_polluted waterways are at risk of poor health due to exposure to organic chemicals such as PAHs and PCBs. Documented adverse health effects include increased susceptibility to infectious diseases and genetic aberrations (Arkoosh et al. 1991, Stein et al. 1995). There are no similar "pollutant bottleneck" local studies for Pacific herring, except at Cherry Point. Similar studies at sites such as in Port Madison and in Port Susan would help clarify the extent to which ambient levels of contamination may pose risk to herring. Further, it would also be instructive to understand how herring fared in past decades in areas where pollution inputs were much greater than they are today.
While there are useful data on contaminants in sediments, shellfish, and fish, we still have no comparable synthesis of contaminant inputs into Puget Sound or the Straits. Inputs may be coming from sources yet to be adequately assessed, such as urban and agriculture runoff channels and ferry terminal tarmacs. Inputs data, coupled with maintenance or enhancement of dated core and mussel watch monitoring, are needed to understand and derive benefit from contaminant management actions and spill responses and, ultimately, to determine clean safety levels.
In conclusion, it is clear that herring residing in Puget Sound proper are generally exposed to higher levels of a variety of chemical contaminants compared to herring in North Puget Sound and nearby coastal areas. Many of these chemicals have known negative effects on aquatic life, and while their effects on Pacific herring stocks within Puget Sound is unknown, laboratory and field evidence suggests that negative effects are likely. The concentrations of specific contaminants vary over spatial and temporal scales. Concentrations of some contaminants, such as polychlorinated biphenyls (PCBs), are declining over time, whereas concentrations of others, such as polynuclear aromatic hydrocarbons (PAHs), are either remaining constant, or are increasing.
West’s (1997) presentation of risk factors for Pacific herring in Puget Sound points to climatic trends (high temperatures) and increased predation by pinnipeds, spiny dogfish, and Pacific salmon as the probable major factors contributing to the decline of these fish. Other factors that may be important are loss or degradation of nearshore nursery habitats; however, no recent, comprehensive Puget Sound-wide information on this exists. What information that remains comes from isolated reports of researchers reporting habitat loss. The increase in abundance in California sea lions and harbor seals may also play a role in the decline of the species, and new studies of diet for these species may shed additional light on this. Increased predation on larval and juvenile fish by delayed-release Pacific salmon may also be important. Loss of fitness due to exposure of larvae, juveniles and adults to contaminants cannot be ruled out. However, some of the contaminant loads in these fish seem to be lessening and there is no clear-cut evidence as to the effects of these contaminants on Pacific herring.
This section assesses the risk of extinction of herring populations in the Georgia Basin DPS, including Puget Sound and the southern Strait of Georgia, and examines trends in California, other British Columbia regions, and Southeast Alaska. The primary considerations are abundance, trends and productivity of Pacific herring populations. Following the presentation of this information, we will use the methods of Wainwright and Kope (1999) and Musick et al. (2000) to evaluate the magnitude of the overall risk.
Puget Sound–The Forage Fish Management Plan (FFMP) (Bargmann 1998) is the most recent management publication by WDFW on Puget Sound herring. The FFMP recognized that forage fish populations tend to be unstable even in the absence of human impacts and that management will not produce stable populations of individual species (Table 7). It stated that "Once a forage fish stock reaches a low level of abundance, recovery may require a protracted period of time, even if fisheries are curtailed or stopped." The FFMP considered herring that utilize specific spawning grounds as individual stocks and contrasted this approach with British Columbia. "In Washington, we attempt to maintain viable populations utilizing each ground each year; in British Columbia loss of herring utilizing a specific ground is not a concern, the management goal is to maintain the overall abundance" (Bargmann 1998).
The WDFW classifies populations into five status categories: 1) Healthy – recent two year mean abundance above or within 10% of the 20 year mean; 2) Moderately Healthy – recent two year mean abundance within 30% of the 20 year mean; and/or with high dependence on recruitment; 3) Depressed – recent abundance well below the long term mean, but not so low that permanent damage to the population is likely (i.e., recruitment failure); 4) Critical – abundance low enough that permanent damage to population is likely or has already occurred; 4) Extinct – no longer can be found in a formerly consistently utilized spawning ground; and 5) Unknown – insufficient assessment data to identify stock status with confidence. The FFMP classified status of the 18 inland water populations as 8 in healthy condition, 1 in moderately healthy condition, 3 depressed, 1 critical, and 5 unknown based on data through 1996 (Table 8). The FFMP also reported that estimates of natural mortality rates increased from less than 0.4 from 1976 to 1980 to more than 0.6 from 1990 to 1995. During this time period the number of age groups comprising the bulk of the populations decreased from five to two or three. While herring formerly lived to ages exceeding 10 years, fish older than 6 years are now rare (Bargmann 1998). Washington Department of Fish and Wildlife (WDFW) (K. Stick and M. O'Toole10) has updated the classifications twice using data through 1998 and 2000 (Table 8). The updates resulted in many changes in the classifications. The latest update classified status of the populations as 10 in healthy condition, 2 in moderately healthy condition, 3 depressed , 2 critical, and 1 unknown.
An earlier study also expressed concern about the status of some of the Puget Sound populations (Chapman et al. 1941). The Cherry Point population was too low to attract fishermen, but fishermen said it had previously been "one of the most productive herring grounds in Puget Sound" (Chapman et al. 1941). The Discovery Bay population also did not attract fishing, but there had been a "fishery of considerable size here" (Chapman et al. 1941). But, in 1959, Williams (1959) found that the Cherry Point and Discovery Bay populations had "regained their productive levels after many years without exploitation".
The petition submitted by Wright (1999) referred to 18 populations of herring in Puget Sound and provided detailed discussion about the "depressed" Cherry Point and "critical" Discovery Bay populations. The Cherry Point population has been the largest Puget Sound population (Table 5), and its late spawning time is unique for Puget Sound. The population is thought to migrate to productive coastal waters because members exhibit relatively rapid growth rates after age-1. Spawn deposition biomass estimates declined from 11,097 tons in 1977 to 1,530 tons in 1997. The extent of the area now used for spawning is only a small fraction of historical observations, and the current spawning area is centered in an area of industrial impacts. The Discovery Bay population is the largest Strait of Juan de Fuca population in U.S. waters and was one of the largest in Washington waters. In 1979, surveyors estimated that the biomass was 3,220 tons. It has declined to low levels of abundance and surveyors were unable to detect herring eggs at this site in 1998. The Port Susan and Port Orchard/Port Madison populations are also called depressed by Bargmann (1998). The petition states they "show the same type of distinct downward trends exhibited by the Cherry Point and Discovery Bay herring populations."
The petition cites evidence that marine ecosystems tend to have relatively few forage species and such species tend to have decadal-scale changes in abundance. An ecosystem with relatively few species at the mid-level is called a "wasp-waist ecosystem" (Rice 1995). The petition states, "A recurring theme in Bakun (1996) is the decadal-scale shifts in abundance that may occur in wasp-waist species; often, but not always, due to replacement by other forage fishes. Puget Sound herring appear to be one of the exceptions to this generalization." However, Bargmann (1998) indicated that assessments of other forage species in Puget Sound may not be adequate to verify that other forage species have not replaced herring in Puget Sound. For example, Pacific sand lance spawning habitat was virtually unknown until discovery of spawn deposits in Port Gamble Bay in 1989. Systematic surveys were developed, but reduced by budget constraints in 1997 (Bargmann 1998). Bargmann (1998) stated "Judging from the reported biology of the species, the widespread nature of their spawning grounds, spawn densities on the spawning beaches, and numbers of spawnings per spawning season, it is possible that there are thousands of tons of sand lances residing in the Puget Sound basin on a year-round basis."
The conclusions expressed in the FFMP and the petition are based upon a substantial time series of herring survey and fishery data in Puget Sound. The WDFW has estimated spawning biomass of important populations annually and minor populations tri-annually (Lemberg 1997). They usually conducted spawn deposition surveys. In some cases, both acoustic/trawl and spawn deposition surveys were made and in some cases only acoustic/trawl surveys were conducted. Spawn deposition results were used for analysis when available. Acoustic/trawl results were similar to egg deposition results when both types of surveys were made (Lemberg 1978, Lemberg et al. 1997). Acoustic/trawl methods were used for analysis when egg deposit surveys were not made. Catches up to survey dates were added to the survey estimates to obtain total run size. The California Department of Fish and Game (Spratt 1981), and Alaska Department of Fish and Wildlife (Rooper et al. 1998) adjust egg deposition estimates for post-spawn mortality. Although Palsson (1984) of WDFW stated "The management of herring stocks might be improved if egg loss rates were incorporated into the procedure of estimating spawning biomass", the department has not done so. Palsson conducted more research "which indicated other biases (e.g., vegetation width variations) that may offset the negative bias of egg loss" (Koenings, unpubl. data). Acoustic/trawl biomass estimates assume that target strength is independent of length. Average length has decreased compared to earlier years. Thus acoustic/trawl surveys could overestimate biomass in recent years relative to earlier surveys. Biomass estimates are shown in Table 6. A quantitative analysis of these trend data will be presented in a later section of this document.
Commercial landings from Washington state inland waters since 1935 are shown in Table 7. Recreational landings are trivial and commercial fishing for herring is not allowed in coastal waters. Only commercial fishing for sport bait has been allowed by the state in recent years. The state does not control treaty fishing and there has been a low-level treaty fishery for sac-roe in recent years. Treaty landings may be incompletely reported in Table 7 prior to 1973. Total landings were modest and did not exceed 1,000 tons until 1958 and then remained above 1,000 tons until 1983. Since then landings ranged from 1,076 tons in 1990 to 361 tons in 1998, which is well below peak landings of 7,171 tons in 1975.
The sport bait fishery occurs primarily in south and central Puget Sound and primarily exploits 1.5-year-old herring that are at the size preferred by recreational salmon fishermen. Sources of these immature herring are not known. Demand for bait has decreased in recent years because of restrictive salmon regulations (Bargmann 1998). The general purpose fishery occurred in northern Puget Sound and most of the catch was reduced to meal and oil, food for zoo animals, and bait (Bargmann 1998). Tagging studies (Buchanan 1985a) indicated that some fish exploited by this fishery migrated to and from Canadian waters and expansion of the fishery was encouraged in the 1960s because it was thought that it would primarily intercept fish bound for Canada. The general-purpose fishery exploited immature and mature herring, but did not exploit herring that were ready to spawn. The sac-roe and spawn-on-kelp fisheries mainly exploit fish thought to be from the Cherry Point population. The spawn-on-kelp fishery impounds herring caught by seine and releases them after spawning. Mortality of these fish was estimated to be no more than 10% (Bargmann 1998), but assumed to be harvested for management and abundance estimation purposes (Koenings, unpubl. data). The non-treaty sac-roe fishery was closed in 1980 because of low stock size (Bargmann 1998). The spawn-on-kelp fishery was closed in 1996 because of continued decreases in stock size (Bargmann 1998).
British Columbia
—Schweigert and Fort (1999) conducted the most recent published assessment of Pacific herring in waters of British Columbia. The Department of Fisheries and Oceans Canada (DFO) uses five areas for assessing and managing herring (Fig. 20). The Prince Rupert area (PR) is adjacent to Alaskan waters and Williams (1999) found some evidence that herring from PR respond to environmental conditions similarly to herring from inshore waters of the Gulf of Alaska. The west coast of Vancouver Island (WCVI) and Strait of Georgia (SG) areas are adjacent to Washington and tagging results indicate some movement of herring between British Columbia and Washington waters (Buchanan 1985a). The SG is larger than Puget Sound and landings in the combined SG and WCVI areas were considerably higher than Washington landings (Table 5). While DFO manages herring on a larger geographical scale than WDFW, some data were available for smaller geographical regions denoted sections (Fig. 20).The Department of Fisheries and Oceans Canada (DFO) made annual spawning biomass estimates using egg deposition surveys and added catches from the roe fishery to obtain total mature stock size. These data were used as input to an age-structured model to produce estimates adjusted for estimated efficiency of each survey. Both results were used to make annual management recommendations for catches from each area based on up to 20% of forecast stock size when the biomass was estimated to be above cutoff set at 25% of estimated unfished average biomass (Schweigert and Fort 1999). Only the survey-based estimates of total mature stock size were used in this review to be comparable to Washington estimates.
The southeast part of the SG is adjacent to the Cherry Point area. Herring spawn in May on Roberts Bank, British Columbia, which is just north of the May spawning area at Cherry Point (Levings 1983). Thus the "Cherry Point population" may extend into Canadian waters and WDFW biomass surveys may not include all of the areas used by this group of fish. The heaviest spawning in section 293 (see Fig. 20) was reported in the Boundary Bay area along the east shore of Point Roberts just north of the B.C.-Washington border ( http://www-sci.pac.dfo-mpo.gc.ca/sci/divisions/default_e.htm) (DFO 2000a). The DFO had permission to survey in Washington waters and their estimates for section 293 some spawning that occurred in Washington. The border did not appear to have an impact on the distribution of spawning. Most of the DFO spawning records for section 293 occurred during February and March (http://www-sci.pac.dfo-mpo.gc.ca/sci/divisions/default_e.htm) (DFO 2000a). It appears that the herring population denoted Semiahmoo by WDFW is part of a larger spawning group shared with Canada. The WDFW was aware that during peak years of the Cherry Point population spawning occurred on the west side of Point Roberts to the Canadian border and probably extended past the border. The WDFW also was aware that spawning by the Semiahmoo population extended into Canadian waters (Koenings, unpubl. data). The WDFW has not had permission to survey in Canadian waters (Mark O'Toole11).
Summaries for all herring spawn deposition surveys in SG since 1951 were obtained from http://www-sci.pac.dfo-mpo.gc.ca/sci/divisions/default_e.htm (DFO 2000a) Sections 201, 202, 280, 291, and 293 (Fig. 20) , which comprise most of the SG sections that border Washington waters, were not included in the DFO assessments for the SG (Schweigert and Fort 1999) and were not regularly surveyed since 1978. The summaries were presented as spawn habitat indices (product of total length of spawn deposits, mean width of spawn deposits, and mean layers of eggs). Survey methodologies changed over the years. Prior to 1987, survey techniques varied but were similar to those used by WDFW.
Since 1987 most major spawning areas, excluding non-assessment areas, were surveyed using self-contained underwater breathing apparatus (SCUBA). Schweigert and Fort (1999) used procedures that varied as the survey methodologies changed to first estimate total number of eggs deposited and then used fecundities and sex ratios to estimate spawning biomass, but did not give details. A plot of the ratio of spawning biomass to spawn habitat index for the assessed portion of the SG indicated that the ratio was relatively constant from 1951 to 1975 (average 0.0090), 1976 to 1986 (average 0.0055), and 1987 to 1999 (average 0.0028). These averages were used to estimate spawning biomass from three portions of the SG: South, all sections adjacent to Washington waters; Northwest, sections 141,142, 143, 161, 171, 172, and 173; and Northeast, sections 151,152, 162, 163, 164, 165, and 280. Because SCUBA surveys were rarely used in the South portion, the 1976 to 1986 spawning biomass to spawning habitat ratio was used for the 1987 to 1999 period in the South portion. The Northwest portion dominated herring spawning biomass except for a short period in the 1960s (Fig. 35). Spawning biomass in the Boundary Bay area, Section 293, comprised a dominant proportion of biomass in the entire SG region during this period of time. Biomass in the assessment portion of the southern area was relatively low, increased in 1970 then decreased to relatively low levels after 1977 (Fig. 35). Because survey effort varied over the time period, available data do not support a more quantitative conclusion than was made by Hay and McCarter (1997a) about trends in herring spawning biomass in the southeastern portion of British Columbia Strait of Georgia region. Landings of herring were dominated by the Northwest and South portions prior to the collapse of the reduction fishery in the l968 (Fig. 36). After the stocks recovered in 1973, landings were dominated be the Northwest portion. However landings were fairly high in the South portion in 1977 and 1978 (Fig. 36).
The Strait of Georgia (SG) population was at relatively high levels in recent years; however, spawn deposition decreased in the southeast parts of the SG (Hay and McCarter 1997a). Biomass in none of the British Columbia areas had downward trends as severe as observed for some of the Puget Sound populations (Table 5). Biomass was relatively high in all British Columbia areas in 1999 except WCVI.
Southeast Alaska—Alaska manages its sac-roe fisheries similarly to British Columbia (http://www.cf.adfg.state.ak.us/geninfo/finfish/herring/forecast/01_4cast.htm) (Larson et al. 2000). No exploitation is allowed if forecasted biomass in a management area falls below a threshold. If biomass is adequately above the threshold, allowable exploitation is set at 20%. If biomass is close to the threshold, allowable exploitation is set at a level that will not drop the surviving biomass below the threshold. There are four management areas in Southeast Alaska: Kah Shakes/Cat Island, Sitka Sound, Seymour Canal, and Craig. Estimates of biomass, forecasted biomasses and thresholds are shown in Table 9. The Kah Shakes/Cat Island and Prince William Sound spawning populations were considered depressed. Prince William Sound biomass decreased from 111,800 tons in 1992 to 28,100 tons in 1993 and has not recovered to high levels observed between 1980 and 1992. Herring returning to Prince William Sound in 1993 were abnormally small, had unusual behavior, and tests indicated the presence of Viral Hemorrhagic Septicemia (Funk 1995). The 1976, 1980, 1984, 1988, 1992, and 1994 year-classes dominated age compositions in both the Prince William and Sitka Sounds (Carlile, unpubl. data).
San Francisco and Tomales Bays—The only significant fishery for herring south of Puget Sound is on spawning populations in San Francisco and Tomales Bays, California (Suer 1987) . Biomass estimates (Table 10) considerably fluctuated without a strong trend, since subtidal spawning areas were first included in the surveys in 1979, but the San Francisco Bay biomass averaged about 39% higher during the 1981-1990 period than during the 1991-2000 period. Combined San Francisco Bay and Tomales Bay estimates for 2000 were the third lowest since 1975.
Musick (1999) presented guidelines based on productivity criteria for considering marine fish to be sufficiently at risk to warrant careful evaluation. These guidelines were developed in the belief that guidelines developed for other groups of animals were not appropriate for marine fish. The life-history characteristics indicate that herring are in the medium productivity category. Musick recommended considering a species at risk if a species in the medium productivity category experiences a decline of 95% or more during the longer of 10 years or three generations. Generation length for herring is about five years. We conducted trend analyses of herring biomass during the past 15 years.
Enough data were available from six of the Puget Sound populations to conduct trend analyses. Trends were estimated using the following model:
Bt+1 = 8Bt = B0e: (1)
where,
Bt = biomass in year t, and
8 = annual rate of biomass change.
Holmes (In prep.) recommended using running sums of biomass to estimate 8. The running sum of biomass at year t is
Rt = GB t+i , where i=0,..,L. (2)
Let
: = sample average of ln(Rt+1/Rt), and
F2(:) = slope of sample variance of ln(Rt+J/Rt),
where slope is the linear regression estimate of the slope in the relationship between sample variance of ln(Rt+J/Rt) and J, and J varies from 1 to 4. The estimate of 8 is then given by
8 = exp( : + F2(:)/2). (3)
Holmes (In prep.) found that the above procedures for estimating : and F2(:) were less biased than the respective maximum likelihood estimates: sample average and variance of ln(Bt+J/Bt). She recommended that L be set to less than 4 and used 3 in her examples. We also set L at 3.
We used the estimates of : and F2(:) to estimate several measures of risk, based on forecasting future stock trends. Declines to 5% of 1999 biomass in fifteen years (following Musick (1999), one ton, or one fish seemed to be significant measures to consider. A wild population of one fish would be for practical purposes extinct. A one ton population of herring would be sufficiently small to be considered very close to extinction and difficult to detect. We assumed that average weight of herring would be 0.2 lbs (Fig. 37). Mean time (Time(F)) to reach a specified decline is ln(F)/ :, where F is fraction of starting population. We also estimated the probabilities of a 95% decline in 15 years, and declines to one ton or one fish in 10, 25, 50, and 100 years (PR(F,t)). We assumed that : follows a log-normal distribution with variance F2(:) following Holmes (In prep.).
Under the assumptions of the model and if current conditions continue, the results (Table 11 and Fig. 36) indicate none of the populations are likely to decline 95% during the next 15 years, but indicate the time is 16 years for the Discovery Bay population. These estimates may also be considered the best estimates of the trend during the past 15 years. There is greater than 50% chance that the Cherry Point population will decline to 1 ton or less in 100 years, and greater than 50% chance that the Discovery Bay population will decline to one ton or less in 50 years. Chances are less than 50% that the other four populations will decline to one ton or less in 100 years. Holmes (In prep.) recommends caution in use of probabilities of extinction if the estimate of 8 is close to one. Holmes (In prep.) doesn’t provide a method for estimation of confidence limits of 8, but if the model holds, and : is normally distributed, and if : is more than 1.96 standard deviations lower than 0, it is significantly different than 0% at the 95% level of confidence. The results (Table 11) indicated that the only significant negative trend was for Discovery Bay. Since multiple tests were made, even if all of the assumptions hold the actual level of significance was less than 95%. These results indicated that the probability estimates should be interpreted with caution.
Population estimates in some of the spawning areas have always been small, and combined populations may produce more meaningful estimates of trends than individual populations. Projections indicated there is greater than 50% chance that biomass in North Sound (Cherry Point and Discovery Bay combined) will decline to 1 ton or less in 100 years (Table 11 and Fig. 38). Chances are less than 50% that biomass in South Sound (Quartermaster Bay, Port Orchard-Port Madison, Port Gamble, and Port Susan combined) (Fig. 39) or Puget Sound (North and South Sound combined) will decline to 1 ton or less in 100 years. As previously mentioned herring biomass in the southeast portion of the Canadian portion of Strait of Georgia (CSG) decreased as did the Washington portion of the Strait, but insufficient detailed data were collected by Canada to allow analysis (Hay and McCarter 1997). Biomass in the CSG is much larger than in Puget Sound and biomass in both the CSG and Georgia Basin DPS had an upward trend (Table 11).
Here we consider two aspects of a population’s productivity: recruitment of young fish into the population and mortality rate of adult fish. Age composition data are available since 1976 for the Cherry Point population, most of the populations for 1998, and for 15 years from the Port Gamble population. Age compositions for the Cherry Point population (Table 12), show the declining representation of older fish with time as described in the Wright petition. Numbers of age-2 fish were low during the first and last four years of the period. There were large fluctuations in numbers between 1980 and 1995 with relatively large values for 1980, 1993, and 1994. Because Puget Sound herring mature at age-2 or age-3, changes in numbers of 2-year-old fish could be due to differences in year-class strength and/or age at maturity. Numbers of fish older than 2 tended to decline with time. The ratio of 2-year-old fish to 3-year-old fish in the following year tended to increase after 1979, which suggests a reduction in the age at maturity.
The average age composition between 1987 and 2000 for Port Gamble fish (Table 13) was similar to Cherry Point fish, but the apparent year-class strengths were not correlated between the two areas. There was considerable variation in age compositions among 12 herring populations located in the Georgia Basin DPS in 1998 (Table 14). However either age-2 or age-3 fish dominated the composition in all areas. Older herring tend to appear and spawn before younger herring (Day 1987). Some of the differences in age composition could be results of the date of sampling effort relative to the timing of spawning of the different populations. Average weight at age tended to decrease after about 1983 (Fig. 37). It is not known if the decrease was due to changes in growth in length or condition.
The combined-age average weight of the Cherry Point population declined sharply between 1976 and 1984, and gradually declined after 1984 (Fig. 37). The decline mainly resulted from reduced numbers of older fish and partially resulted from reduced weight-at-age. The average biomass of Cherry Point herring decreased 83% from 10,973 tons in the 1976-1979 period to 1,815 tons in the 1996-1999 period (Table 15). If the average weight at age observed during the 1976-1979 period occurred during the 1996-1999 period the decline in biomass would have declined to 1,919 tons in the 1996-1999 which is very similar to the observed decline. The number of 2-year-old fish actually increased between the two periods, but the numbers of older fish decreased sharply.
Average exploitation rates tended to increase with age for the mature fish fisheries on the Cherry Point and Port Gamble populations (Fig. 40). Age composition data are not available for the sport bait fishery, but this fishery targets age-1.5-year-old fish. Exploitation rates of mature fish from the Cherry Point population decreased sharply when the non-treaty sac-roe fishery was closed in 1980 (Fig. 41). The treaty sac-roe fishery exploited the population at modest rates between 1986 and 1996. There was also a modest treaty-fishery on the Port Gamble population between 1987 and 1993 (Fig. 41). Exploitation rates are not available for the sport-bait fishery, but should be modest unless a relatively small population was heavily targeted. Estimates by WDFW of non-fishing related annual mortality rates of mature herring in Puget Sound increased sharply between 1976 and 1985 and slightly if at all since 1985 (Fig. 42). The mortality rates could include some fishing mortality caused by Canadian fisheries (Buchanan 1985b).
Because the Cherry Point population was abundant in the early-1970s, it is informative to backcalculate the number of recruits that would have been necessary to produce this level of abundance. Numbers of 3-year-old fish in the Cherry Point population were approximated for the 1971-1975 period using the numbers of 4 to 8-year-old fish in 1976 (Table 12) and the following assumptions: 1) annual rate of natural mortality was 0.2 (instantaneous rate =0.22) (Fig. 40), 2) relative age-specific fishing mortality was the same as during 1976-1980, 3) weight at age was same as during 1976-1980, 4) only the sac-roe fishery exploited Cherry Point herring, 5) the ratio of number of 2-year-old fish in year t to the number of 3-year-old fish in year t+1 was the same as during 1976-1980, and 6) fish did not survive beyond 9 years and the number of 9 year fish was the same as the 1976-1980 average. Results indicated that there were about 30, 62, 42, 22, and 31 million 3-year-old fish from 1971 to 1975. While these estimates must be considered first approximations, the numbers of 3-year-old fish during 1971 to 1975 probably averaged considerably higher than latter years.
Age compositions did not indicate that a decrease in abundance of older fish occurred in the British Columbia stocks that was comparable to the observed decrease in Puget Sound populations (Schweigert and Fort 1999). However, concern was expressed that fish were migrating from SG to the Central Coast Area (CCA), because estimates of natural mortality were considerably higher for SG than for the CCA. An analysis was made to examine the possibility and did not produce convincing results (Schweigert and Fort 1999). Weight-at-age decreased in all major British Columbia herring stocks since the mid- to late-1980s (Stocker and Kronlund 1998), which is consistent with the decline observed for the Cherry Point population (Fig37). The 1972, 1974, 1985, 1987, 1989, 1994, and 1995 year-classes were relatively strong in SG. The 1972, 1985, and 1994 year-classes were relatively strong in WCVI (Schweigert and Fort 1999).
In summary, a combination of reduced recruitment of 3-year-old herring and increased non-fishery related losses of older fish appeared to be the primary causes of the decline in biomass of Cherry Point and perhaps other Puget Sound populations of herring. Reduced weight-at-age appeared to contribute little to the decline.
The BRT utilized the process as presented by Wainwright and Kope (1999) to assess the three main risk categories: abundance and trends in population, productivity and variability as well as habitat quality change. The members of the BRT were asked to rate these risks for 1 to 5 with 1 representing very low risk and 5 as high risk of extinction in the near future due to this factor.
For the Georgia Basin DPS of Pacific herring, abundance was rated by the BRT as a modal score of 2. A score of 2 represents "Low risk. Unlikely that this factor contributes significantly to the risk of extinction by itself, but some concern that it may in combination with other factors." The range was between 1 and 2. A score of 1 represents "Very low risk. Unlikely that this factor contributes significantly to risk of extinction, either by itself or in combination with other factors." For trends in abundance, the unanimous score was also 2. For changes in habitat quality, the modal score was 2 with a range from 2 to 3. A score of 3 represents "Moderate risk. This factor contributes significantly to risk of extinction, but does not in itself constitute a danger of extinction in the near future." As a reference, other species that have been subsequently recommended for listing generally have scored in the 3 to 5 range for each factor.
Musick et al. (2000) have developed a method that compares information pertaining to the productivity of the DPS to criteria based on productivity of species and resiliency of populations. Because the risk matrix process assisted the BRT in determining that the Georgia Basin DPS of Pacific herring is not in danger of becoming extinct in the foreseeable future, the BRT utilized this method to assess whether the species might be "vulnerable" which is "(of special concern), not necessarily endangered or threatened severely, but at possible risk of falling into one of these categories in the near future" (Musick et al. 2000). The information available for use in this method is the time-at-maturity (Tmat) (2-4 yrs) and the maximum age of fish (Tmax) (4 to 10 years). Both of these parameters result in a "medium" productivity parameter (Table 16).
The BRT then utilized this overall productivity value to determine if this species is "vulnerable" to becoming extinct. The IUCN defines a species as being vulnerable if its DPS has experienced a 20% decline in 10 years or 3 generations, whichever is longer (Musick 1999). The proposed American Fisheries Society (AFS) risk criteria (Table 17) indicates that a species with medium to high productivity, such as Pacific herring, would have to undergo a decline of at least 95% in 10 years or 3 generations, whichever is longer. The Georgia Basin DPS of Pacific herring does not approach this decline, so the BRT determined that this DPS does not meet the criteria for "vulnerable" at present. However, some populations such as Cherry Point and Discovery Bay within the DPS do met the criteria.
The BRT utilized the methods presented in Wainwright and Kope (1999), West (1997) and Musick et al. (2000) to organize their conclusions regarding risk to the Georgia Basin DPS of Pacific herring. The BRT concluded, by a large majority, that the Georgia Basin DPS of Pacific herring are neither at risk of extinction nor likely to become so. However, most members expressed concern that they could not entirely rule out the possibility that this Georgia Basin DPS at present is likely to become in danger of extinction, especially because some stocks within the Georgia Basin, such as Cherry Point and Discovery Bay, have declined to such an extent that they may meet the IUCN criteria to be considered "vulnerable". Although the BRT recognized that herring populations in north Puget Sound and Puget Sound proper may be vulnerable to extinction, these populations represent a relatively small portion of the overall DPS of herring in the Georgia Basin. Moreover, because of the moderate to high productivity of herring populations and the tendency of herring to stray among spawning sites, the BRT felt that there are reasonable possibilities at present for recolonization of depleted populations associated with specific spawning sites. The BRT also expressed caution that important changes in resource management practices (e.g., increased harvest levels) and in the ecosystem (e.g., increased numbers of marine mammals or predatory fish species), as well as increased habitat degradation, could result in increased extinction risk for Pacific herring in this DPS.
However, the BRT, emphasized that while the DPS is defined at a larger scale than the "stocks" that are managed in Puget Sound by WDFW, and that the Georgia Basin DPS does not appear at risk of extinction at present, local populations are the appropriate scale for fisheries management activities, and as McQuinn (1997) emphasizes, their "conservation is essential for the preservation of spawning potential and for the viability of coastal fisheries."