Data on this ESU were limited. The PFMC (1994) states that: Inside harvest of coho is not available for any river system in California. Spawning escapement estimates are available for Klamath River Basin hatcheries, but not for coho spawning in natural areas. Recent population estimates were compiled under contract to NMFS Southwest Region (Brown and Moyle 1991) and have since been published (Brown et al. 1994). Recent status reviews of coho salmon in California by NMFS (Bryant 1994) and the California Department of Fish and Game (CDFG 1994) have expanded on the work of Brown and Moyle (1991) for estimates of abundance and trends in coho populations throughout much of the state.

In compiling estimates of recent spawner abundance, Brown and Moyle (1991) used a 20-fish rule: If a stream with historic accounts of coho salmon lacked recent data, it was assumed to still support a run of 20 adults; if coho salmon were present in recent stream surveys, they used 20 adults or the most recent run estimate, whichever was larger. While these estimates are crude, in most cases they are the best data available, and they are generally comparable with other estimates (Bryant 1994, CDFG 1994, Maahs and Gilleard 1994). Unless otherwise indicated, the recent abundance data are taken from Brown et al. (1994).

California statewide (including areas outside this ESU) coho salmon spawning escapement apparently ranged between 200,000 and 500,000 adults per year in the 1940s (Brown et al. 1994). By the mid-1960s, statewide spawning escapement was estimated to have fallen to about 100,000 fish per year (CDFG 1965, California Advisory Committee on Salmon and Steelhead Trout 1988), followed by a further decline to about 30,000 fish in the mid-1980s (Wahle and Pearson 1987). From 1987 to 1991, spawning escapement averaged about 31,000, with hatchery populations making up 57% of this total (Brown et al. 1994).

Brown and Moyle (1991, Brown et al. 1994) estimated average coho salmon spawning escapement in the central California coast ESU as 6,160 naturally spawning coho salmon and 332 hatchery spawned coho salmon for the period from 1987 to 1991 (Table 7). Of the naturally spawning coho salmon, 3,880 were from tributaries in which supplementation occurs (Noyo River and coastal streams south of San Francisco). Only 160 fish in the range of this ESU (all in Ten Mile River) were identified as native fish, lacking a history of supplementation with non-native hatchery stocks. Based on redd counts, the estimated run of coho salmon in Ten Mile River during the 1991-92 spawning season was 14 to 42 fish (Maahs and Gilleard 1994).

Table 7. Regional summary of recent (1980s) average coho salmon spawner abundance in California. Numbers are subdivided by the apparent origin of the fish (probably native, mixture of native and naturalized, or hatchery). Based on data from Brown et al. (1994).

*A few minor coastal streams in Humboldt County south of Punta Gorda are included in this subtotal.

Recent data exist for 133 of 186 streams in the region identified by Brown et al. (1994) as having historic records of adult coho salmon. Of these 133 streams, 62 have recent records of occurrence of adult coho salmon and 71 no longer have coho salmon spawning runs (Table 8). (Note that the summaries by county made by Brown et al. (1994) excluded a few streams at the northern boundary of this ESU and included the Sacramento River, which is outside this ESU.) Nehlsen et al. (1991) provided no information on individual coho salmon stocks in this region, but identified stocks in small coastal streams north of San Francisco as at moderate risk of extinction, and those in small coastal streams south of San Francisco as at high risk of extinction. Higgins et al. (1992) considered only drainages from the Russian River north but identified four coho salmon stocks within this ESU as at risk: three of special concern and one (Gualala River) as at high risk of extinction.

2) Southern Oregon/Northern California Coasts

Data are also limited for this ESU. No regular escapement estimates exist for natural spawning in California streams, and information in Oregon is limited to angler catch summaries for all rivers, dam-passage and seine-survey counts in the Rogue River, and observations of coho salmon during chinook salmon spawner surveys.

Most information for the northern California region of this ESU was recently summarized by the California Department of Fish and Game (CDFG 1994). They concluded that coho salmon in California, including hatchery stocks, could be less than 6 percent of their abundance during the 1940's, and have experienced at least a 70 percent decline in numbers since the 1960's (CDFG 1994, p. 5-6). They also reported that coho salmon populations have been virtually eliminated in many streams, and that adults are observed only every third year in some streams, suggesting that two of three brood cycles may already have been eliminated.

The Klamath River Basin (including the Trinity River) historically supported abundant coho salmon runs. In both systems, runs have been greatly diminished and are now composed largely of hatchery fish, although there may be small wild runs remaining in some tributaries (CDFG 1994). Of 396 streams within this ESU identified as once having coho salmon runs, Brown et al. (1994) were able to find recent survey information on 115 (30%) (Table 8). Of these 115 streams, 73 (64%) still supported coho salmon runs while 42 (36%) did not. The streams identified as presently lacking coho salmon runs were all tributaries of the Klamath and Eel River systems (Brown et al. 1994). The rivers and tributaries in the California area of this ESU were estimated to have average recent runs of 7,080 natural spawners and 17,156 hatchery returns, with 4,480 identified as native fish occurring in tributaries having little history of supplementation with non-native fish (Table 8).

^aA few minor coastal streams in Humboldt County south of Punta Gorda are included in this subtotal.
^bIncludes Sacramento River.

Data for coastal Oregon south of Cape Blanco include adult passage counts at Gold Ray Dam in the upper Rogue River (Cramer et al. 1985; Kostow 1994 App.), angler catch estimates for all Oregon rivers (ODFW 1992, 1993b), and seine-survey estimates of adult coho salmon run-size in the Rogue River (Satterthwaite 1992). Recently, most production of coho salmon in this region of Oregon has been in the Rogue River Basin. Recent run-size estimates for this basin (1979-91) (Satterthwaite 1992) have ranged from approximately 450 to 19,200 naturally produced adults, and from 200 to 9,400 hatchery produced adults (Fig. 45). Average run sizes for this period were 3,400 natural and 3,300 hatchery fish, with the total run averaging 49% hatchery fish. (The majority of the hatchery component probably returns to Cole Rivers hatchery rather than spawning in the wild; average hatchery rack returns for this period were 2,800.)

Adult passage counts at Gold Ray Dam provide a long-term view of coho salmon abundance in the upper Rogue River. In the 1940s, passage counts averaged approximately 2,000 adults per year (Fig. 46). Numbers declined and fluctuated during the 1950s and early 1960s, then stabilized at an average of fewer than 200 adults during the late 1960s and early 1970s. In the late 1970s, dam counts increased with returning fish produced at Cole Rivers Hatchery.

Angler catch of coho salmon in the Rogue River has fluctuated considerably, ranging from less than 50 (late 1970s) to a peak of about 800 in 1991; average annual catch over the last 10 years has been about 250 fish. Angler catch in other rivers in southern Oregon has been low, representing only a minor fraction of the total south of Cape Blanco. While there have been no directed surveys for coho salmon in this region, the species would be expected to be observed in the annual chinook salmon spawner surveys. Few coho salmon have been observed in these surveys; for example, in 23 years of chinook salmon surveys over 6 segments of the Elk River, the highest count of coho salmon was 20 adults in 1971. In Oregon south of Cape Blanco, Nehlsen et al. (1991) considered all but two coho salmon stocks to be at high risk of extinction; of the remaining two, one (Euchre Creek) was identified as extinct and the other (Hunter Creek) was not mentioned. (The status of coho salmon in Euchre Creek is in some doubt: no surveys have been conducted recently, but Oregon Department of Fish and Wildlife (ODFW) biologists believe there may be a small coho salmon population there.) South of Cape Blanco, all Oregon coho salmon stocks were rated by Nickelson et al. (1992) as depressed.

Combining recent run-size estimates for the California portion of this ESU with the Rogue River estimates provides a rough minimum run-size estimate for the entire ESU of about 10,000 natural fish and 20,000 hatchery fish.

3) Oregon Coast

In this region, we examined the extensive spawner survey records provided by ODFW (Cooney and Jacobs 1994) in conjunction with estimates of ocean harvest rates for coho salmon (PFMC 1993c). Cramer (1994) provided an extensive analysis of these data. There are three related sets of spawner survey data. Beginning in 1950, ODFW conducted annual surveys of peak coho salmon spawning abundance ( peak counts, PC) in standard survey segments of coastal rivers from the Necanicum River to the Coquille River. In 1980, surveys began on a more comprehensive set of river segments: counts were conducted multiple times each season, providing area-under-the-curve (AUC) estimates of total spawning over the season, and a better representation of local population abundance than the PC estimates.

In 1990, a study was initiated to examine potential biases and statistical validity of the survey program (Jacobs and Cooney 1990, 1991, 1992, 1993). This (not yet completed) study involves stratified random sampling (SRS) to provide AUC estimates of coho spawning abundance throughout coastal Oregon north of Cape Blanco. All three survey data sets provide estimates of local spawner density in terms of spawners per stream mile. The PC estimates have been used to estimate coastwide spawner population size for fishery management purposes, but substantial biases in this approach led to the initiation of the SRS study. Preliminary results of this study provide statistically valid estimates of coastwide spawner abundance for the 1990-93 seasons (Jacobs and Cooney 1991, 1992, 1993; ODFW 1995).

Based on historical commercial landing statistics and estimated exploitation rates, Mullen (1981) estimated escapement of coho salmon in coastal Oregon to be nearly 1 million fish in the early 1900s, with harvest of nearly 400,000 fish. In a more extensive analysis of similar data, Lichatowich (1989) concluded that coho salmon abundance in the same region at that time was about 1.4 million fish. Lichatowich also concluded that current production potential (based on stock-recruit models) for coho salmon in coastal Oregon rivers was only about 800,000 fish, and he associated this decline with a reduction of nearly 50% in habitat capacity. Recent spawning escapement estimates based on SRS survey results (Table 9) indicate an average spawning escapement of about 39,000 adults. While the methods of estimating total escapement are not comparable between the historical and recent periods, these numbers suggest that current abundance of coho salmon on the Oregon coast may be less than 5% of that in the early part of this century. The ODFW (1995) presented estimates of coho salmon abundance at several points of time from 1900 to the present. These data show a decline of about 75% from 1900 to the 1950s, and a further decline of about 90% since the 1950s.

Trends in coastal Oregon coho salmon populations can best be assessed from spawner survey information. Long-term trend indications are available only from the PC survey standard segments (Fig. 47, 48). Statewide average peak spawners per mile have declined substantially since the early 1950s, but the majority of this decline occurred in the early 1970s, and spawner counts have remained relatively stable since that time. Since the late 1970s, stocks in the north and central portions of this ESU have exhibited stable to declining escapements, while those in the southern portion of the ESU (Umpqua and Coos-Coquille management districts) have exhibited increasing escapements (Fig. 48). Spawner-to-spawner return ratios based on peak counts (Fig. 47) have fluctuated around replacement over the entire data period but have been below replacement in 5 of the most recent 6 years for which we had data. Despite the relative stability of spawner densities after 1975, recruits to the fishery (estimated by dividing spawner density by 1 minus the ocean harvest rate) continued to decline until the mid-1980s. Based on these statewide average peak count data, preharvest recruit-to-spawner ratios also have been declining up to the present (Fig. 47), although this result is not consistent with results of Cramer's (1994) analysis of individual rivers that shows no recent decline in recruit-to-spawner ratios. The ODFW (1995) estimates that this decline in recruits per spawner averaged 5% per year for brood years 1978 to 1991.

More recent AUC estimates (Fig. 49) may provide a better indication of recent trends than do the peak counts. These estimates show little trend in either spawners or recruits since 1980, although recruits are low for the last 3 years. Return ratios based on AUC estimates are extremely low for the most recent 6 years for which we had data, with spawners failing to replace themselves in all 6 years. As for the PC estimates, return ratios based on AUC estimates also exhibit a decline since the 1980 brood year. We also examined statewide average PC and AUC spawner estimates based only on streams designated as wild coho salmon production by ODFW; trend and replacement ratio estimates for these streams were similar to those calculated for all streams.

The observed pattern of declining recruits per spawner is a serious concern. If the recent trend were to continue into the future, we could expect that recruitment would fall below replacement levels in the near future, even with no harvest. However, such a projection is uncertain. The harvest rate estimates used in this analysis are very inexact, since they are based on questionable assumptions regarding distribution of fishing effort and the relative magnitude of hatchery and wild production in the Oregon Production Index (OPI) area. Recent analyses by ODFW (1995) suggest that recent harvest rate estimates may be substantially revised when new methods based on analysis of stock-specific, coded-wire tagged returns are used. Also, we have not attempted to evaluate the causes for this apparent decline in recruits per spawner, so we cannot evaluate the likelihood that those causes will continue into the future. The ODFW (1995) suggests that changes in ocean production since 1976 are largely responsible for the decline in return ratios and concludes (p. 49) that there is no clear indication of trends in the post-1975 period and no indication of an impending switch to better conditions. However, there are other factors that could contribute to declining production of coho salmon along the Oregon coast, including loss and degradation of freshwater habitat (Lawson 1993). Lacking knowledge of these factors, their effects and future direction, our best estimate of the future is a simple projection of past trends. Based on information available at this time, our projection suggests a substantial risk that coho salmon in this ESU may soon face recruitment failure.

Kostow et al. (1994) provided estimates of hatchery composition of naturally spawning coho salmon in several coastal Oregon rivers, ranging from 18 to 62% (Table 10). These estimates are for rivers known to have high hatchery influence and, therefore, do not represent the average condition along the Oregon coast. We compiled hatchery composition estimates for a larger sample of rivers and lakes, which show a wide range from less than 10% hatchery fish in lake samples to more than 75% in two rivers (Table 11). These data also illustrate a general north-south trend of decreasing hatchery influence on natural spawning. The extensive presence of hatchery-origin adults spawning in several coastal rivers is a cause for concern about the sustainability of natural production in these systems.

4) Lower Columbia River/Southwest Washington Coast

The status of lower Columbia River coho salmon stocks outside of the Willamette River Basin was reviewed extensively by Johnson et al. (1991) and is not reconsidered here. The ODFW conducts annual coho salmon spawning surveys in the lower Columbia River Basin (Fennell 1993). These surveys indicated that natural spawning of coho salmon in this region declined precipitously in the early 1970s and has remained at extremely low levels.

As noted earlier, the Clackamas River, a tributary of the Willamette River, may support a native run of coho salmon that is a remnant run of fish native to the lower Columbia River Basin (Cramer and Cramer 1994). Abundance of this run has been measured since 1950 by adult passage at River Mill (1950-57) and North Fork (1958-present) Dams, and total run size (native and hatchery) has ranged from 416 (1950) to 4,700 (1968). The native portion of the run has ranged from 309 (1958) to 3,588 (1968) (Fig. 50).

Cramer and Cramer concluded that production of the native population is depressed due to a variety of factors. They further concluded that under current harvest rates, the population is likely to remain stable but is vulnerable to overharvest. Johnson et al. (1991) briefly reviewed abundance data for this population and, although they concluded that it had a low risk of extinction if population parameters remained stable, they recommended close monitoring of the population.

In the Columbia River Basin, all coho salmon stocks above Bonneville Dam (except Hood River) were classified by Nehlsen et al. (1991) as extinct. Hood River, Sandy River, and all other lower Columbia River tributary stocks were classified as at high risk of extinction, except the Clackamas River stock, which was classified as at moderate risk of extinction. This historic ESU also included portions of the southwest Washington coast. Nehlsen et al. (1991) identified coho salmon stocks in Willapa Bay as at high risk of extinction. WDF et al. (1993) identified the Willapa Bay stocks as of unknown status, but of mixed origin and composite production. They identified all stocks in Grays Harbor tributaries as healthy, but of mixed origin and composite production.

The largest production of coho salmon in this area is in the Chehalis River Basin. Hiss and Knudsen (1993) estimated current coho salmon run sizes (before terminal harvest) in this basin (including the Humptulips River) at about 266,000 adults, of which 135,000 are naturally produced and 131,000 are of hatchery origin. They noted that hatchery influence on these runs has increased rapidly since 1970.

Coho salmon in the Chehalis River basin exhibit two run timings: normal, with spawning in early December throughout the basin, and late, with spawning in January and February in lower Chehalis River tributaries. Hiss and Knudsen (1993) suggested that the normal run is composed of a mixture of hatchery and wild fish, while the late run is virtually all wild fish (but they did not specify whether wild implies native fish, or simply natural production regardless of origin). The two runs were treated as a single stock for fishery management purposes, and we have no separate abundance estimates for them. Hiss and Knudsen (1993) identified three streams known to have late-run fish (Bingham Creek, the upper Wynoochee River, and the Wishkah River) and noted that this run has always been less abundant than the normal run but has been particularly small in recent years. No escapement estimates are available for other streams in Grays Harbor or Willapa Bay.

5) Olympic Peninsula

Data on terminal run size for stocks in this ESU are collected cooperatively by WDFW and the coastal tribes. Spawning escapements to most streams are estimated from cumulative redd counts on index reaches of the streams, assuming each redd represents the spawning activity of 1 female, and a 1:1 sex ratio. Peak counts in supplemental reaches are multiplied by the ratio of cumulative counts to peak counts in the index reaches. Escapement to a basin is calculated by multiplying the estimated escapement in surveyed (index + supplemental) areas by the ratio of surveyed spawning habitat to total habitat in the basin.

The WDFW and tribal biologists believe that redd counts provide the most reliable estimates of total escapement in these coastal streams, which typically have highly variable flows during the spawning season (PFMC 1990). These natural escapement estimates, combined with hatchery escapements, form the basis for escapement summaries for the Olympic Peninsula (WDF et al. 1993, PFMC 1994). However, no attempt was made to estimate the number of hatchery produced fish that spawn naturally.

We reviewed assessments of these stocks by Nehlsen et al. (1991) and WDF et al. (1993). Nehlsen et al. (1991) identified only one at-risk coho salmon stock (Lake Ozette) in this area, as of special concern. Most coho salmon stocks in this area were considered by WDF et al. (1993) to be healthy or of unknown status, representing a mixture of native, mixed, and non-native origins, and wild or composite production. Some stocks along the Strait of Juan de Fuca were identified as depressed. The WDF et al. (1993) report identified eight stocks of native origin with wild production in this ESU, four of healthy status and four of unknown status.

Other evidence examined for this ESU included trends in terminal run size, hatchery contribution, trends in ocean exploitation rate, and trends in the size of fish in terminal landings. No trends were detected in terminal run size, and there is no evidence for trends in ocean exploitation rates. In the stock complexes monitored and reported by PFMC, hatchery returns accounted for 50% of the spawning escapement in the period from 1982 through 1992 (Table 12), with the majority of hatchery production contributing to the Quillayute summer-run, Quinault, and Queets stocks (PFMC 1994).

Of these stocks, the Quinault and the Salmon River (tributary of the Queets) were identified by WDF et al. (1993) as of mixed origin, while the majority of other stocks were identified as of native origin. Average recent run sizes for some of these stocks are given in Table 12. The total average terminal run size for these main stocks is about 34,000 natural and 34,000 hatchery fish. We have found no historical run-size estimates for these stock complexes to compare with recent abundance, but there have presumably been substantial declines in coho salmon production as a result of well-documented habitat degradation since European settlement.

6) Puget Sound/Strait of Georgia

Spawning escapements in Puget Sound are estimated primarily by AUC estimates of live spawners in index reaches of selected streams. Spawner surveys are conducted by WDFW. Estimates of spawning escapement are obtained by multiplying the mark-recapture estimate of spawning escapement in a base year by the ratio of AUC index to the AUC index in the base year (PFMC 1990).

We reviewed assessments of these stocks by Nehlsen et al. (1991) and WDF et al. (1993). Nehlsen et al. (1991) identified three coho salmon stocks in this region as at high risk of extinction, and one (Nooksack River) to be possibly extinct. Stocks in this region were considered by WDF et al. (1993) to range from healthy to critical in status, be predominantly of mixed origin, and be predominantly of composite production. None of the stocks in this region that they identify as healthy were of strictly native origin. Two stocks (Deer Creek and Sumas/Chilliwack) were identified as of native origin with wild production but were of unknown status.

Other lines of evidence included long-term trends in escapement to counting facilities, hatchery contribution rates, ocean and total exploitation rates, and trends in the size of fish in the terminal landings. Only three rivers have long-term (extending back to the 1930s or 1940s) escapement data from which to estimate trends. Long-term trap counts at Baker River and White River generally showed declining trends in the 1960s and 1970s, with some evidence of recovery in the 1980s. The number of adults passed above the hatchery racks on the Samish River showed neither increasing nor decreasing trends over a 55-year period. More recent spawner survey data are available for numerous rivers within this ESU, but no reliable breakdown of natural and hatchery production is available for these data.

Trends in abundance for additional stocks were evaluated by examining terminal run size derived from cohort reconstruction (WDF et al. 1993). Terminal run data included terminal harvest within Puget Sound, but not harvest in ocean fisheries. In addition to naturally produced fish, terminal runs included hatchery fish from hatchery and off-station releases. Of the stocks identified by WDF et al. (1993), abundance estimates were available for the period from 1965 through 1993 for 17 stocks. Trends in abundance for these stocks were estimated by fitting an exponential trend to time series of terminal run size. Of the stocks examined for this review, two stocks had significant downward trends, five had significant upward trends, and the remainder had no significant trend (Table 13).

Table 13. Average Puget Sound run size from 1965 through 1993, and trends in abundance for selected Puget Sound stocks. Runs size estimated by run reconstruction. Data from WDF et al. (1993).

* denotes trends that significantly (P < 0.05) differ from zero.

Ocean exploitation rates on wild coho salmon from the Deschutes River, Snohomish River, and Big Beef Creek declined from the late 1970s through the mid-1980s and have increased since then but remained in the range of 0.3 to 0.5. Total exploitation rates have shown no apparent trend but have fluctuated in the range of 0.6 to 0.9. The average hatchery contribution rate for stocks monitored and reported by the PFMC for the period 1981 to 1992 has been 62%, with Nooksack/Samish and South Puget Sound stock complexes managed for, and clearly dominated by, hatchery production.

Average recent run sizes are summarized for major stock complexes in the ESU in Table 14. The total average run size for these main stocks is about 479,000 natural and 776,000 hatchery fish. Bledsoe et al. (1989) examined changes in run sizes of Puget Sound salmon since 1896. They did not find a statistically significant general decline in run sizes for wild runs of coho salmon in this period, although they did report a dramatic 85% decline of coho salmon terminal runs in south Puget Sound from 1935 to 1975. They attributed this decline at least in part to an increasing catch in nonterminal fisheries.

Overall, catches of coho salmon in Puget Sound fisheries show a substantial decrease from 1896 to the early 1940s, but this is largely attributed to the prohibition of fishing for this species with purse seines and fish traps starting in 1935. Overall catch within Puget Sound has increased gradually since that time, but has not returned to earlier levels, possibly as a result of greater interceptions of coho salmon in ocean fisheries (Bledsoe et al. 1989).

As noted above, between 1972 and 1993 the average size of fish in the terminal landings underwent a sharp decline, from an average of about 4 kg to about 2 kg (Fig. 16). This dramatic decline in average fish size, which could result from any of several causes, could seriously reduce the fecundity and fitness of naturally spawning fish.

The ESU that includes Puget Sound extends into southern British Columbia, for which we have not received detailed abundance information. Northcote and Atagi (in press) have reviewed abundance trends for all salmon species in various regions of British Columbia, two of their regions include parts of this ESU. Coho salmon have shown both historical (1800s to 1953-92 average) and recent (1953 to 1992) declines both on Vancouver Island and along the south-central British Columbia coast (excluding the Fraser River).

In both areas, the historical decline was roughly twofold. On Vancouver Island, coho salmon escapements have recently declined from more than 300,000 in the mid-1950s to about 150,000 at present. Along the south-central British Columbia coast, escapement declines in the same period have been more dramatic, from about 500,000 in the mid-1950s to less than 100,000 at present. This is a much more severe decline than the trends documented in the U.S. portion of the ESU.

Northcote and Atagi did not address levels of hatchery production for British Columbia coho salmon. However, there has been a substantial increase in coho salmon releases from British Columbia hatcheries since 1975 (Hilborn and Winton 1993).

Conclusions

In general, there is a geographic trend in the status of coho salmon stocks south of the Canadian border, with the southernmost and easternmost stocks in the worst condition. Throughout the regions reviewed, there have been recent declines in coho salmon abundance, and 1994 runs were predicted to be the worst on record in many river basins. (At the time of this report, we have received no 1994 run-size or escapement estimates for most river basins.) Conclusions for specific ESUs follow and are summarized in Table 15.

1) Central California Coast

All coho salmon stocks south of Punta Gorda are depressed relative to past abundance, but there are limited data to assess population numbers or trends. The main stocks in this region have been heavily influenced by hatcheries, and there are apparently few native coho salmon left in this region. The apparent low escapements in these rivers and streams, in conjunction with heavy historical hatchery production, suggest that natural populations there are not self-sustaining. The status of coho salmon stocks in most small coastal tributaries is not well known, but these populations are small. There was unanimous agreement among the BRT members that natural populations of coho salmon in this ESU are presently in danger of extinction.

2) Southern Oregon/Northern California Coasts

All coho salmon stocks between Punta Gorda and Cape Blanco are depressed relative to past abundance, but again there are limited data to assess population numbers or trends. The main stocks in this region (Rogue River, Klamath River, and Trinity River) are heavily influenced by hatcheries and, apparently, have little natural production in mainstem rivers. The apparent declines in production in these rivers, in conjunction with heavy hatchery production, suggest that the natural populations are not self-sustaining. The status of coho salmon stocks in most small coastal tributaries is not well known, but these populations are small. There was unanimous agreement among the BRT that coho salmon in this ESU are not in danger of extinction but are likely to become endangered in the foreseeable future if present trends continue.

3) Oregon Coast

There are extensive survey data available for coho salmon stocks in this region. Overall, spawning escapements have declined substantially during this century and may now be at less than 5% of their abundance in the early 1900s. Average spawner abundance has been relatively constant since the late 1970s, but preharvest abundance has declined. Average recruits-per-spawner may also be declining. Coho salmon populations in most major rivers have heavy hatchery influence, but some tributaries may sustain native stocks. The BRT concluded that coho salmon in this ESU are not at immediate risk of extinction but are likely to become endangered in the future if present trends continue.

4) Lower Columbia River/Southwest Washington Coast

The BRT concluded that we cannot at present identify any remaining natural populations of coho salmon in the lower Columbia River (excluding the Clackamas River) or along the Washington coast south of Point Grenville that warrant protection under the ESA, although this conclusion would warrant reconsideration if new information becomes available. The Clackamas River produces moderate numbers of natural coho salmon. The Clackamas River late-run coho salmon population is relatively stable under present conditions, but depressed and vulnerable to overharvest. Its small geographic range and low abundance make it particularly vulnerable to environmental fluctuations and catastrophes, so this population may be at risk of extinction despite relatively stable spawning escapements in the recent past. As noted above, the BRT could not reach a definite conclusion regarding the relationship of Clackamas River late-run coho salmon to the historic lower Columbia River ESU. However, the BRT did conclude that if the Clackamas River late-run coho salmon is a native run that represents a remnant of a lower Columbia River ESU, the ESU is not presently in danger of extinction but is likely to become so in the foreseeable future if present conditions continue.

5) Olympic Peninsula

Coho salmon abundance within this ESU is moderate, but stable. These stocks have been reduced from historical levels by large-scale habitat degradation in the lower river basins, but there is a significant portion of coho salmon habitat in several rivers protected within the boundaries of Olympic National Park. This habitat refuge, along with the relatively moderate use of hatchery production (primarily with native stocks), appears to have protected these coho salmon stocks from the serious losses experienced in its adjacent regions. While there is continuing cause for concern about habitat destruction and hatchery practices within this ESU, the BRT concluded that there is sufficient native, natural, self-sustaining production of coho salmon that this ESU is not in danger of extinction and is not likely to become endangered in the foreseeable future unless conditions change substantially.

6) Puget Sound/Strait of Georgia

Coho salmon within this ESU are abundant and, with some exceptions, run sizes and natural spawning escapements have been generally stable. However, artificial propagation of coho salmon appears to have had a substantial impact on native, natural coho salmon populations, to the point that it is difficult to identify self-sustaining, native stocks within this region. In addition, continuing loss of habitat, extremely high harvest rates, and a severe recent decline in average size of spawners indicate that there are substantial risks to whatever native production remains. There is concern that if present trends continue, this ESU is likely to become endangered in the foreseeable future. However, the size data examined are heavily influenced by fishery data from the Puget Sound. These fisheries target primarily hatchery stocks, and it is not known at this time to what extent the trends in size are influenced by hatchery fish. The extent of hatchery contribution to the natural spawning escapement and to natural production is unclear, as are the potential effects this contribution may have on the population genetics and ecology of this ESU. Further consideration of this ESU is warranted to attempt to clarify some of these uncertainties.