Skip booklet index and go to page content

COSEWIC assessment and status report on the coho salmon (Interior Fraser population) in Canada

Population Sizes and Trends

Aboriginal Traditional Knowledge

Aboriginal Traditional Knowledge (ATK), sometimes referred to as Traditional Ecological Knowledge, describes the knowledge originating with First Nations peoples pertaining to their immediate environments, and the cultural practices that build on that knowledge (Ford and Martinez 2000). Communities with a long history of resource use can acquire a deep but qualitative knowledge about the resource that they depend upon (Kurien 1998).

Interior Fraser coho return to spawn primarily within the traditional territories of the Secwepemc people (North and South Thompson and Clearwater rivers) and of the Nlaka’pmux, Sce’exmx and Okanagan people of the upper Fraser canyon and Nicola valley. Some coho spawning also takes place within the traditional territories of the St’at’imc, (Lillooet/Bridge River areas) and Tsilhqot’in (Chilcotin river system). The Secwepemc Fisheries Commission (SFC) and the Nicola Valley Stewardship and Fisheries Authority (NWFSA) represent bands with knowledge of traditional fisheries. In addition, there are various bands not affiliated with these organizations that also possess ATK. ATK pertaining to some natural resources in the interior Fraser has been assembled (e.g. Turner et al. 2000) but no thorough review of ATK has been undertaken for salmon. Irvine et al. (2000) review issues concerning the role of ATK in the assessment of interior Fraser coho salmon.

Abundance Estimates

We have no estimates of the abundance of coho salmon in the interior Fraser prior to the arrival of Europeans. Northcote and Burwash (1991) calculated that the average annual abundance (catch plus spawners) for Fraser River coho salmon in the 1920’s to early 1930’s was approximately 1.2 million. Assuming ~1/3 of these fish were from the interior Fraser, the abundance of interior Fraser coho during this period was ~400 000 and assuming as Northcote and Burwash did that 50% of these fish were harvested, the annual escapement of coho was in the order of 200 000 fish. In their report on fish of the Fraser River basin, Northcote and Burwash (1991) estimated that coho salmon in the Fraser watershed underwent a 7.7 fold decrease between the 1920’s and the period between the 1950’s and the 1980’s. However, they cautioned that data for coho were the least reliable of the salmon data.

Spawning coho salmon are notoriously difficult to count. Although escapement estimates (number of salmon escaping marine fisheries and returning to freshwater to spawn) exist for some streams in the interior Fraser as far back as 1951, older estimates are of unknown accuracy and precision. Consequently they are of little use for time series analysis, other than to confirm species presence in a watershed.

A variety of techniques are currently used to estimate coho salmon escapements for ~71 streams in the interior Fraser River watershed. Although essentially all known spawning streams in the Thompson watershed are assessed, many in the upper Fraser are not.

Precision is more important than accuracy in time series analysis, and high precision requires that field and estimation methods should be consistent through time.Two approaches were used to examine recent trends in spawner numbers, an escapement indicator approach and an approach using an adjusted escapement series.

The escapement indicator approach relied on escapement estimates to unenhanced North and South Thompson streams with reasonably consistent monitoring. Total escapements to the 10 North Thompson escapement indicator streams followed similar temporal patterns as escapements to the 16 South Thompson escapement indicators (Irvine et al. 2001). Numbers appeared to be relatively stable during the late 1970’s, increased during the early 1980’s, peaking in the latter part of the decade. Since then, escapements declined rapidly until 1996, after which a small increase may have occurred. Using procedures documented in Simpson et al. (2001), the adjusted escapement series consisted of total numbers of coho salmon spawners returning to the major basins within the Thompson watershed, and included hatchery-origin fish (Table 1). Similar to the wild indicator data set, the adjusted escapement estimates for the North and South Thompson also peaked in the mid-1980’s, declined until about 1996, and were then stable or increasing (Fig. 6). It was difficult to reconstruct the time series for lower Thompson/Nicola and non-Thompson populations. A significant portion of returns to the lower Thompson/Nicola is enhanced, and many estimates are unreliable. Escapements to the lower Thompson appear less variable than to other parts of the Thompson, although they also may have increased in recent years (Fig. 6). The time series for non-Thompson streams is too short to assess temporal patterns.

At the peak of our time series of reliable estimates, escapements to the interior Fraser were in the order of 100 000 and abundance in two years appeared to exceed 300 000 (Table 1). The size of the total population spawning more recently was estimated by averaging escapement estimates for each of the five subpopulations (i.e. North Thompson, South Thompson, lower Thompson/Nicola, Fraser canyon, and upper Fraser) during 1998-2000 (Table 1). All estimates will be biased low to a small degree simply because it is not possible to see all fish. Of greatest concern is the degree of bias for upper Fraser streams. Results from a fishwheel mark recapture program imply that this bias may be in the order of 57% (Irvine et al. 2001). The total population of interior Fraser coho spawning annually in recent years is probably in the order of 24 000 (23 914) fish (Table 2), of which ~15% are of hatchery-origin. The North and South Thompson populations collectively include slightly more than half the total population of coho in the interior Fraser

Figure 6: Adjusted Historical Escapement Estimates for Coho Salmon (Wild and Hatchery Fish) Returning to the South, Lower, and North Thompson Watersheds

Figure 6: Adjusted historical escapement estimates for coho salmon.

Data are in Table 1.


Table 1 (South Thompson): Estimated fishery exploitation rates, escapements, marine fishery catches, and total abundances for interior Fraser coho salmon (hatchery and wild fish combined)
Estimated fishery
exploitation rates
EscapementsMarine fishery
19750.685 86412 49018 354
19760.683 9208 34912 268
19770.688 49018 08226 572
19780.687 99617 03225 028
19790.6810 19821 72031 918
19800.687 02514 96421 989
19810.684 1208 77512 895
19820.685 84912 45918 308
19830.686 19613 19619 392
19840.6815 39432 78948 183
19850.6816 99836 20553 204
19860.6616 52131 66548 186
19870.5421 08724 47845 564
19880.7124 42660 37684 802
19890.6517 20831 28848 496
19900.748 60924 06932 677
19910.684 1608 73712 896
19920.8111 88652 23964 125
19930.881 87313 17215 045
19940.434 4853 4307 915
19950.563 6224 6398 261
19960.831 7608 90610 667
19970.402 0341 3843 418
19980.074 9463755 321
19990.093 0743053 379
20000.033 7851343 919
20010.0713 23999614 235

Data updated from Irvine et al. 2001. Exploitation was not measured prior to 1987 but was assumed to be the same as the average exploitation during 1987-1996. 2001 estimates are preliminary.


Table 1 (continued - North Thompson): Estimated fishery exploitation rates, escapements, marine fishery catches, and total abundances for interior Fraser coho salmon (hatchery and wild fish combined)
Estimated fishery
exploitation rates
EscapementsMarine fishery
19750.6822 28647 46869 754
19760.6820 67544 03764 713
19770.6842 80491 171133 975
19780.6839 09583 269122 364
19790.6847 819101 851149 670
19800.6810 54222 45432 996
19810.6820 61543 90964 524
19820.6842 29590 087132 382
19830.6835 08674 731109 816
19840.6869 552148 141217 692
19850.6845 16096 188141 349
19860.66104 267199 846304 113
19870.5454 88463 710118 594
19880.7170 612174 539245 150
19890.6530 67755 77986 455
19900.7425 69771 84497 542
19910.6814 58530 63345 217
19920.8122 04296 875118 917
19930.889 66967 99977 667
19940.4310 0317 67117 702
19950.5622 47728 79451 272
19960.8312 31962 32574 645
19970.406 7224 57311 295
19980.079 1256859 810
19990.098 9168859 801
20000.037 0322507 282
20010.0726 4291 98928 418

Data updated from Irvine et al. 2001. Exploitation was not measured prior to 1987 but was assumed to be the same as the average exploitation during 1987-1996. 2001 estimates are preliminary.


Table 1 (continued - Lower Thompson): Estimated fishery exploitation rates, escapements, marine fishery catches, and total abundances for interior Fraser coho salmon (hatchery and wild fish combined)
Estimated fishery
exploitation rates
EscapementsMarine fishery
19840.685 15512 05017 205
19850.681 9134 0605 973
19860.662 2114 3006 511
19870.544 2084 9459 153
19880.714 0139 83013 843
19890.653 4236 3409 763
19900.744 42112 60017 021
19910.683 7948 82512 619
19920.814 90521 00025 905
19930.888 41661 50069 916
19940.435 2523 9659 217
19950.561 9842 5254 509
19960.831 2095 9007 109
19970.404 2172 8207 037
19980.072 6282002 828
19990.095 0074955 502
20000.034 4591574 616
20010.079 82874010 568

Data updated from Irvine et al. 2001. Exploitation was not measured prior to 1987 but was assumed to be the same as the average exploitation during 1987-1996. 2001 estimates are preliminary.


Table 1 (continued - Non-Thompson/Fraser): Estimated fishery exploitation rates, escapements, marine fishery catches, and total abundances for interior Fraser coho salmon (hatchery and wild fish combined)
Estimated fishery
exploitation rates
EscapementsMarine fishery
19980.078 1476108 757
19990.095 3895355 924
20000.034 7231444 867
20010.0713 5151 01814 533

Data updated from Irvine et al. 2001. Exploitation was not measured prior to 1987 but was assumed to be the same as the average exploitation during 1987-1996. 2001 estimates are preliminary. The Fraser Canyon and the Upper Fraser subpopulations have been merged into Non-Thompson Fraser.


Table 2: Mean number (percentages) of spawners escaping to the major basins of the interior Fraser during 1998-2000
OriginSouth Th.North Th.Lwr Th./
Wild + hatchery3 935 (18)8 358 (37)4 031 (18)4 092 (18)1 994 (9)22 410
Wild + hatchery3 935 (16)8 358 (35)4 031 (17)4 092 (17)3 498 (15)23 914
Wild only3 904 (19)7 202 (35)1 617 (8)4 092 (20)3 498 (17)20 313

In the lower two rows, the number of spawners estimated in the upper Fraser was divided by 0.57 to account for fish in streams that were not surveyed.


Hatchery production of coho salmon began in the early 1980’s, peaking during the mid- to late 1980’s when enhancement strategies were being tested for coho in the Eagle, Salmon, and Coldwater systems. The main objectives of enhancement during this period were to evaluate the effectiveness of different strategies and to assess the impact of enhanced production on natural stocks (Perry 1995, Pitre and Cross 1993). These authors concluded that fry releases might be a useful supplementation strategy when progeny from natural spawning do not fully occupy available habitat. However, negative interactions between wild and hatchery coho salmon can occur, particularly when the release of hatchery coho fry results in the carrying capacity of a stream being exceeded.

There are no large production facilities for coho salmon in the interior Fraser. There are ~13 small enhancement projects producing coho as well as habitat restoration at various sites. Enhancement efforts currently focus on rebuilding depressed stocks and obtaining assessment information that can be used for both wild and enhanced stocks. Since temporal patterns for the adjusted escapement series (Fig. 6) that included hatchery fish were similar to those of the wild indicator series, it appears that enhancement had a relatively minor effect on overall population trends. Bradford and Irvine (2000) reached a similar conclusion. They found that the presence of hatchery activities had no significant effect on rates of decline. Enhancement activities are described in more detail by Irvine et al. (1999a, 2000).

Since many of the coho salmon released from hatcheries are marked, and most of the salmon returning to these systems are examined for marks, it was possible to estimate the proportion of the escapement that are of hatchery origin. During 1998-2000, hatcheries contributed a significant proportion of the fish returning to streams in the lower Thompson/Nicola watersheds (Table 2), but relatively few fish elsewhere.

Survivals and Fishery Exploitations

Temporal patterns of salmon abundance are often analysed by partitioning survival into freshwater and marine components. Freshwater survival estimates are not available for interior Fraser coho salmon. Marine survivals declined during the decade following the late 1980’s, a pattern that has been documented for many coho south of northern British Columbia (Coronado and Hilborn 1998). While marine survivals may have improved in the last two years, they are still generally less than 3%, much less than they were in the late 1980’s (Irvine et al. 2000).

Estimates of fishery exploitations up to 1997 are available through the MRP, and since then from an analysis of DNA from tissue samples from fish taken in fisheries (Irvine et al 2001). Unprecedented restrictions in Canadian salmon fisheries commencing in 1997 and increasing in 1998-2000 are apparent in the time series of exploitation rates (Table 1). Exploitation rates (i.e. catch/catch plus escapement) during 1998-2000 averaged 6.5% (Canada and USA combined). In contrast, the mean annual exploitation during the ten-year period from 1987-1996 was 68% (Table 1).

Rates of Decline

Rates of decline were computed based on:

  1. The escapement time series consisting of data from the 10 North Thompson and 16 South Thompson indicator streams, which have relatively few missing data and are unaffected by hatchery activities.4
  2. The adjusted total escapement estimates for the North and South Thompson watersheds (Table 1).

Declines were estimated in two ways. The first approach used annual estimates of abundance and the standard COSEWIC formula for calculating declines.5 The second (smoothed) used running averages calculated over 3-year periods, the normal generation time for these fish. This approach reduced year to year variations in abundance, which are common with semelparous animals including coho salmon that reproduce primarily at one age.

Each approach resulted in relatively large rates of decline (Table 3). Estimates for the 10-year period from 1990 – 2000 ranged from 27 – 73% (overall mean 60%).6

The lowest estimated decline resulted when the smoothed approach was used to interpret escapement indicator data from the North Thompson watershed. Differences between methods and data sets were not consistent.

Annual estimates of productivity for Thompson coho were estimated as:

ran = ln [Rt/St-3]

where Rt is recruitment (i.e. catch plus escapement) and St-3 is the abundance of parent spawners (i.e. escapement). Thus r is a measure of survival from spawners to returning (i.e. prefishery) adults. The time series of ran for the mean of the 10 North and 16 South Thompson indicator streams is presented (Fig. 7).


Table 3: Rates of decline (percentages) for South and North Thompson coho salmon escapement indicator data (wild coho only) and adjusted estimates of total escapements (wild and hatchery) during 1990-2000
MethodS. Thompson

Esc. Indic.
S. Thompson

S. Thompson

N. Thompson

Esc. Indic.
N. Thompson

N. Thompson

Overall Mean

Estimates were calculated using the standard COSEWIC formula, as well as using 3 year smoothed data.

Figure 7: Time Series of ran, the Annual Rate of Population Growth of Thompson Coho Salmon

Figure 7: Time series of ran, the annual rate of population growth of Thompson coho salmon.

From Irvine et al. 2001. Each point is the average (±SE) of two time series (North and South indicator stream aggregates). When r<0, populations are unable to replace themselves, even in the absence of fishing.

There was an overall decline in ran from the mid-1980’s until 1997-1998 (Fig. 7). During some years (1991 and possibly 1995, 1997, and 1998) ran may have been < 0, meaning that some populations were unable to replace themselves, even if fishing mortality was zero. Fortunately, the average ran for the 1999 and 2000 returns was positive.

Reference Points

Biological reference points are benchmarks against which the status of fish populations can be measured (Collie and Gislason 2001). Irvine et al. (2001) computed provisional reference points for coho from the North Thompson watershed (see also Chen et al. 2002) which are used here to identify somewhat arbitrary categories of abundance. Results are presented in terms of numbers of adult female coho salmon per km of habitat accessible to them.

The estimated number of female spawners that would produce maximum sustained yield (24.9 females/km of accessible habitat) was selected to identify the transition between zones of moderate and high abundance (Fig. 8). The mean of two lower reference points was chosen as the boundary between critical and poor categories of abundance. One lower reference point was the minimum escapement that the population has recovered from (6.1), and the second was a value computed as the theoretical 10% probability of extinction for a single brood line in one generation (4.3) (Irvine et al. 2001; Chen et al. 2002). The zone between poor and moderate abundance was two times this mean value (10.4).

As can be seen in Fig. 8, with one exception spawner numbers were in the moderate and abundant categories from 1975-1990, were in the poor or low moderate zones from 1991-1996, and were near but generally below the poor-critical boundary from 1997-2000 (Fig. 8).

Figure 8: Annual Estimates of Numbers of Adult Female Coho Salmon (Wild and Hatchery) per Kilometre within the North Thompson Watershed

Figure 8: Annual estimates of numbers of adult female coho salmon (wild and hatchery) per kilometre within the North Thompson watershed.

Adapted from Irvine et al. 2001). Horizontal lines indicate boundaries between abundance categories.

Future Scenarios

The future for interior Fraser coho is highly uncertain and depends on impacts due to two categories of human-induced activities, fishing and habitat perturbations, and one category that is largely out of our control, climate-related changes in salmon survival. The human population in the Pacific Northwest (including British Columbia) is expected to increase by 2-7 fold this century (Lackey 2001). Hartman et al. (2000) discuss how human activities affect salmon at local, regional, and global levels and conclude that human population growth likely represents the greatest threat to Pacific salmon.

In Fig. 9, the outlook for coho in the North Thompson watershed is forecast beginning in 2001 under three simplistic scenarios. In the most optimistic scenario, survivals are assumed to improve to levels recorded during 1978-1997 (recruits/spawner (R/S) = 3.31); in the average scenario, survivals remain the same as during 1998-2000 (R/S = 1.47); and in the worst case scenario, survivals are the same as recorded for fish returning in 1998, a poor survival year (R/S = 0.96).7 It is assumed that there will be no additional impact of habitat development and that fishing will remain at the same low levels as during 1998-2000 (i.e. ~7%). In addition, spawners are assumed to be all 3 years old, and spawner numbers cannot exceed 50 female coho per kilometer, a level reached only once in the previous 25 years (Fig. 8).

For North Thompson and presumably other populations of coho, a return to survival levels experienced historically, combined with continued low fishing pressures and no additional habitat impacts, would theoretically produce rapid increases in escapements and rebuilding within two generations. If survivals continue at recent average levels, populations will take 5-6 generations to reach the abundant category. However, if survivals are as poor as they were in 1998, numbers will decrease, eventually resulting in extinction (Fig. 9).

These forecasts are simplistic and probably overly optimistic for several reasons: the capacity of the environment to produce coho salmon has probably been reduced in recent years due to habitat changes and this has been ignored in the model; the extremely low fishing pressures assumed have come at a large socioeconomic cost and may be difficult to maintain; and, finally, no variability in survival is assumed in the model. Routledge and Irvine (1999) found that even small increases to the amount of chance variation in recruitment can significantly reduce population survival rates.


4 North Thompson watershed index streams are: Barrièrre, Blue, Cook, East Barrièrre, Fennel, Lion, N. Thompson, Raft, Reg Christie, and Tumtum, while South Thompson watershed index streams are Adams (lwr), Adams (up), Bessette, Blurton, Bolean, Canoe, Hunakwa, Kingfisher, Scotch, Shuswap (lwr), Shuswap (mid), South Pass, Tappen, Trinity, Wap, and Sinmax.
5 Formula provided by COSEWIC (
6 Preliminary escapement estimates for 2001 became available as this report was being revised. When the smoothed approach was used to compute declines during 1991-2001 using the adjusted escapement data set, estimates for the South and North Thompson aggregates were 12 and 34% respectively. These estimates were more positive than estimates for 1990-2000 for two reasons: 2001 escapements were much higher than in recent years, and the 1991-2001 data set excluded the relatively large escapements of 1990 (relative to 1991).
7 Spawner recruitment data all from the North Thompson indicator stream data set.