Atlantic salmon (Salmo salar) COSEWIC assessment and status report: chapter 11

Population Sizes and Trends

The data compiled for the analysis of all Canadian DUs were provided by the Canadian Department of Fisheries and Oceans and the Quebec Ministère des Resources naturelles et de la Faune. Spawning escapement estimates (the number of fish available to spawn each year after all fisheries have taken place) were used throughout the trend analysis. Escapement was chosen over pre-fishery abundance based on COSEWIC criteria to use “mature individuals who are capable of reproducing”. Within COSEWIC, definitions of mature individuals are further defined as follows: “Mature individuals that will never produce new recruits should not be counted”. Assuming a significant proportion of the salmon captured historically in commercial and recreational fisheries would have reproduced, the use of spawning escapement data in trend analysis would, relative to the abundance of fish before the fisheries occur, will underestimate the extent of decline in several DUs (compare the trends shown in Figures 13 and 14). However, when spawning escapement is used for the trends analysis, the effectiveness of management actions such as fishery closures (described in the next section) is taken into account in the analysis. Canadian abundance reconstruction suggests significant declines in pre-fishery abundance across all DUs and the North American population as a whole (Chaput 2009; Figure 14). This decline appears to have stabilized in most northern regions during the last 3 generations (DUs 1-3, 5-7), but not in the south.

The analysis of population trends was standardized to provide consistent assessments across DUs. Catch data were used primarily in the analysis despite the potential error associated with these types of data (O’Connell 2003) as it was widespread and common to most areas. These data do, however, carry significant risk and uncertainty. O’Connell (2003) demonstrated that major differences can occur when using recreational catch data to infer total returns. He showed that in one case returns were overestimated by approximately 60% in four of seven years. A review of the status of salmon (Dempson et al. 2006) stated that stocks for which only angling data were available are not routinely evaluated, in the Newfoundland-Labrador region. Reasons for this included changes in daily and season bag limits, the introduction of split seasons and quotas in some areas in some years, the switch from DFO Guardian-provided recreational catch data to that obtained from a licence stub return system, the complexities and confusion of interpreting catch-and-release statistics over the years, and the fact that in some areas and years 35-65% of all potential fishing days may be unavailable owing to environmental closures. O’Connell et al. (1998) also showed there could be substantial differences between angling data derived from the licence stub system versus that provided by DFO Guardians for years when the two methods overlapped. This depended on the year and area in question, and was much more pronounced for released fish rather than retained salmon. Despite these well-documented potential problems these were the only data available for all DUs that would allow nation-wide comparison. In some areas, data were limited (e.g. DUs 1 and 2) and/or better info was available (DUs 13, 14). Details on sampling effort and data quality issues are provided for each DU. River-specific trend data from other sampling methods are presented graphically where available. Where the catch data trends diverge from river-specific data, the differences are noted in the DU text.

COSEWIC specifies time frames of 10 years or three generations (whichever is longer) in the examination of population trends. The complex and variable life history of Atlantic Salmon results in different generation times within and among rivers. A DU-specific generation time was determined by averaging the modal smolt age for the rivers presented in Chaput et al. (2006a)xv and adding 1 or 2 years for the marine phase of life, depending on whether MSW fish were common in the specific DU. This approach would slightly underestimate generation time in populations where repeat spawning frequency is high. Smolt ages were typically consistent or within one year of other rivers within a DU. Abundance trends were analyzed using a time series for which the length was determined by multiplying the generation time by three and roughing up to the whole number. For example, if the generation time was 4.1 years, the trend was analyzed over 13 years.

Abundance trends were assessed with a general linear model using a negative binomial error distribution (all statistics computed using R; R Development Core Team (2007)). Values for the calculation of percent change in abundance were taken from the predicted values of the general linear model (latest year and that from 3 generations previous). These estimates of change isolate temporally driven change and are more robust to spurious results. The statistical significance of the estimates trends was assessed at the 95% confidence level. Forward projections have not been provided due to the known dangers of predicting outcomes beyond the range of the data collected. They would also require unrealistic assumptions of static conditions and the absence of abundance-dependent phenomena such as depensation (which would hasten the decline) or compensation (which would slow or halt the decline). Because significant declines have occurred during the last four decades (Reddin 2010; Figure 14), and because for some DUs, the inclusion of just one extra generation resulted in significant trends that were not detected in analyses using three generations, where available longer time series are presented graphically for each DU.

The estimate of abundance for Canada is based on the sum of all DU-specific data and should be considered a minimum value as full abundance estimates were not available for DUs 1, 13 and 14. The ‘complete’ data set spans 1993-2007. The Canadian estimate of abundance of spawning, wild adult Atlantic Salmon was 524,288, in 2007. Of these 414,163 were small salmon and 110,154 were large salmon. Where data were available, 2008 appeared to have improved returns versus 2007. The lowest estimate over the data set was 364,373 in 1994 while the highest was 611,405 (1996). Overall, the model-based estimate of total abundance appears to have increased slightly since 1993 (by 11%), but the trend in the data was non-significant (P = 0.41; Figure 15). Small salmon abundance has increased by 19% from 1993 levels, while large salmon abundance has decreased by 14% of 1993 levels. Neither trend was significant over three generations (P = 0.246 and 0.136 respectively). However, within this broad assessment there are population components and regions that are experiencing significant declines (i.e., MSW salmon and DUs 4, 8, 9, 14, 15, 16; Table 2) or are extinct (DU 11xvi). Regions at the southern extent of the Canadian range (Nova Scotia Southern Upland, DU 14; inner and outer Bay of Fundy, DUs 15 and 16) have undergone marked declines. Trends from individual DUs suggest that small and large salmon may be on differing trajectories of abundance, although neither trend is significant at the Canadian scale in the last three generations. Reddin and Veinott (2010) also suggest that small salmon are increasing in abundance while large salmon are declining. The analysis used in this report was applied to the data for Newfoundland and Labrador, presented by Reddin and Veinott(2010) and Reddin (2010), and it was determined that the increasing trend in small salmon abundance was marginally significant (P = 0.061) and the declining trend in large salmon abundance was highly significant (P < 0.001). The overall trend for total salmon was not significant (P = 0.302). Large salmon have declined to 59% of 1993 levels. The divergent trends for MSW and 1SW salmon abundance are difficult to explain, but the data suggest that the risk of extended periods at sea may be relatively higher than it was historically. Repeat spawners (with the exception of DUs 14-16) have experienced improved survival in recent years (e.g. Cameron et al. 2009).

Figure 13: Posterior Distributions from Monte Carlo Simulation of Estimated Returns to the Rivers/Coast (after sea fisheries of Newfoundland and Labrador and St. Pierre and Miquelon) of Large Salmon (upper) and Small Salmon (lower) for Eastern North America, 1971 to 2007

Box plots showing posterior distributions from Monte Carlo simulation of estimated returns to the rivers and/or coast of large salmon (upper panel) and small salmon (lower panel) for eastern North America, 1971 to 2007.

Box plots are interpreted as follows: dash is the median, rectangle defines the 5th to 95th percentile range, vertical line indicates minimum and maximum values from 10,000 simulations (taken from Chaput 2009).

Figure 14: Posterior Distributions from Monte Carlo Simulation of Estimated Pre-fishery Abundance of Large Salmon (upper) and Small Salmon (lower) from Eastern North America, 1971 to 2007

Box plots showing posterior distributions from Monte Carlo simulation of estimated pre-fishery abundance of large salmon (upper panel) and small salmon (lower panel) from eastern North America, 1971 to 2007.

Pre-fishery abundance for large salmon is only available to the 1SW year of 2006. Box plots are interpreted as follows: dash is the median, rectangle defines the 5th to 95th percentile range, vertical line indicates minimum and maximum values from 10,000 simulations (taken from Chaput 2009).

Figure 15: Small, Large and Total Atlantic Salmon Escapement for Canada (small: top panel; large: middle panel; total: bottom panel) Over the Past 3 Generations (15 years)

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Canada over the past three generations. Superimposed is the general linear model used to determine trends in abundance.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance.

Fisheries Managementxvii

The abundance of Atlantic salmon in Canada has been significantly influenced by fisheries management policy. To provide further context, a brief overview of fisheries management is presented.

As early as the 1970s, fisheries managers began placing restrictions on commercial Atlantic salmon harvests to replenish depleted stocks (May 1993). When pronounced declines in abundance were observed in the 1980s, a wide range of additional management measures were introduced for conservation purposes. The closures of commercial fisheries were expanded in 1984 to include all the commercial fisheries of the Maritime Provinces and portions of Quebec. Further reductions were introduced through the late 1980s and early 1990s, leading to a moratorium on commercial Atlantic Salmon fishing for insular Newfoundland in 1992, followed by a moratorium in 1998 for Labrador, and culminating with the closure of all commercial fisheries for Atlantic Salmon in eastern Canada in 2000.

In 1984, mandatory catch and release in recreational fisheries of all large Atlantic Salmon was introduced in the Maritime Provinces and insular Newfoundland. Since then, more restrictive angling management measures have been introduced in an attempt to compensate for declining survival and Atlantic Salmon abundance, including reduced daily and season bag limits, mandatory catch and release of large and in some cases all sizes of Atlantic Salmon, and in large portions of the Maritimes, the total closure of all directed fisheries.

The need for increasingly severe restrictions on harvests over the past decades reflects the chronically unrealized expectations of Atlantic salmon stock recovery. Though population increases did occur, they were often short-lived (e.g. Dempson et al. 2004). Over longer terms, harvest restrictions in most DUs have generally contributed to the stabilization of declining populations or slowed declines (the exceptions being DUs 2 and 5). As stated previously, the positive contributions of these management restrictions may have had the effect of lessening the degree of reduction in the productive capacity of Atlantic salmon populations, as indicated by spawning escapement indices, but could mask the actual decline in overall abundance of salmon based on the indicators of total returns or pre-fishery abundance.

Table 2: Trends in Atlantic Salmon spawner abundance for designatable units of eastern Canada

Designatable Unit Recent Abundance (Year) Small Salmon % change over 3 generations (p-value) Large Salmon % change over 3 generations (p-value) Total Salmon % change over 3 generations (p-value)
1 - Nunavik DD DD DD DD
2 - Labrador 235,064 (2008) +443.9 (<0.001) +127.9 (0.016) +380 (<0.001)
3 - NE Newfoundland 80,505 (2007) -11.0 (0.569) +1.7 (0.946) -9.6 (0.619)
4 - S Newfoundland 21,866 (2007) -37.3 (0.063) -26.2 (0.293) -36.1 (0.071)
5 - SW Newfoundland 44,566 (2007) +132.1 (<0.001) +143.7 (<0.001) +133.6 (<0.001)
6 - NW Newfoundland 31,179 (2007) -4.2 (0.838) +41.7 (0.126) 0.0 (0.999)
7 - Qc E North Shore 5,901 (2008) -26.3 (0.085) 50.8 (0.115) -13.79 (0.287)
8 - Qc W North Shore 15,135 (2008) -34.0 (0.031) -20.1 (0.143) -24.4 (0.013)
9 - Anticosti Island 2,414 (2008) -31.7 (0.076) -48.7 (0.017) -40.2 (0.007)
10 - St. Lawrence 4,169 (2008) -1.8 (0.951) +11.5 (0.429) +5.27 (0.772)
11 - Lake Ontario Extinct1 - - -
12 - Gaspé-Gulf 103,149 (2007) -34.0 (0.119) -18.5 (0.217) -27.8 (0.100)
13 - E Cape Breton* 1,150 (2008) -7.9 (0.789) -14.5 (0.542) -28.9 (0.202)
14 - NS Southern Upland* 1,427 (2008) -58.6 (0.002) -74.0 (0.001) -61.3 (<0.001)
15 - I Bay of Fundy <200 - - -
16 - O Bay of Fundy 7,584 (2008) -56.6 (0.024) -81.6 (<0.001) -64.3 (0.001)

Probability values associated with inferred trends are given in parentheses. Note that DUs annotated with asterisks reflect abundance estimates for a subset of rivers. DD - Data Deficient. Table summarizing trends in Atlantic Salmon spawner abundance for designatable units of eastern Canada.
1 Currently assessed as Extirpated (COSEWIC 2006a); however, this report proposes that it be revised to Extinct, in keeping with the implication of the current COSEWIC guidelines for recognizing DUs, that loss of an entire DU represents an extinction event, not an extirpation.

Designatable Unit 1 – Nunavik

Data were limited to the sporadic angling effort and catch statistics for Ungava Bay (MRNF 2009, MRNF unpublished data). The limitations of these data restricted the analysis to assessment of catch per unit effort (CPUE). As with all fishery-dependent data, the assumptions of constant catchability of the salmon and the equivalence of effort over the data set are likely to be violated. However, given that the fishery is limited to angling, changes in fishing gear and techniques are less of a factor than in commercial fisheries. Unfortunately, catchability of Atlantic Salmon is heavily influenced by water conditions. Angler data are the only type consistently available for almost all salmon populations, thus a broad assessment requires its utilization.

The data for Ungava Bay was from four of the five known salmon rivers during the time period 1984 – 2008. Mean rod-days per year was 1,014 with a range of 415-1,615. Effort has generally been declining over the time series. No estimate of abundance could be calculated. There also was a significant increasing trend in CPUE over the time series (GLM on catch with effort offset: P=0.007). While the data only include four rivers with commercial angling activities, salmon have been reported from other rivers in this DU. The George River and the Koksoak River had substantially higher CPUE estimates than the Feuilles and Baleine rivers, suggesting higher abundances over the time series. There have been no known extirpations in this area.

Designatable Unit 2 – Labrador

Data for the Labrador DU were diverse. There were commercial catch data (1969-2001) (Reddin 2010) and count data from four counting fences (2002-2008). These data were used in conjunction with habitat data to estimate abundance per habitat unit over time, which was then scaled up for the whole region, which includes 85 Labrador salmon rivers (Reddin 2010). The five rivers from Quebec that are part of DU 2 have spawner abundance time series, based on catch data, that were added to the Labrador data to derive an abundance time series for the entire DU.

There is considerable uncertainty associated with these data since it assumes the four index rivers in southern Labrador are representative of a huge geographical region (scaling from ~1,700 to 65,500 km2), which includes varying intensities of Aboriginal fishing and habitat quality. Furthermore, information from Quebec rivers is based on angler data (MRNF 2009, MRNF unpublished data ) and habitat scaling (Caron and Fontaine 1999) that are also characterized by considerable uncertainty.

The most recent estimate of adult abundance for DU 2 is 235,064 with 206,093 being small salmon (<63 cm) and 28,970 being large salmon (>63 cm). The lowest abundance during the last three generations was 30,555 in 1991. The highest abundance over the same time frame was 242,758 in 2005. During the last three generations there have been significant increases in abundance of small (P<0.001), large (P=0.016) and total salmon (P<0.001) (Figure 16). The abundance of small salmon (based on the curve fit in Figure 16) is 443.9% greater than the 1990 abundance while large salmon abundance is up by 127.9% over the same period. Total salmon are at levels 380.0% of those in 1990 (Figure 16). Data for counting fence facilities in DU 2 (English River, Muddy Bay Brook, Sandhill River and Southwest Brook) are provided in Figure 17. Additional river-specific abundance data are provided in Appendix 1 (see Big Brook, Pinware, Forteau and du Vieux Fort rivers).

As with all following DUs (except DU 11), it should be noted that using statistics of adult salmon spawners as a measure of population health has the disadvantage of potentially masking the severe declines observed in pre-fishery abundance. In this case, when commercial fishery-related mortality is accounted for, current levels of salmon abundance in DU 2 are much lower than expected (Reddin 2010).

The only known population to be lost from this DU was Bobby’s Brook, located near the Alexis River. There has been no evidence of re-colonization of this tributary to date (D. Reddin, Dept. of Fisheries and Oceans, pers. comm.).

Figure 16: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 2 (1969-2007)

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 2, from 1969 to 2007. Superimposed is the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations. Note that pre-1984 data for Quebec components of DU 2 were unavailable and are not included in this plot. Since 1984, the Quebec component only contributed an average of 4% of the run (range: 1-12%).

Figure 17: Salmon Abundance in Four Index Rivers in Southern Labrador

Charts showing salmon abundance in four index rivers (English, Muddy Bay Brook, Sand Hill and Southwest Brook) in Southern Labrador.

(taken from Reddin 2010). Note that the time periods are not identical among the panels and that the Sand Hill data include breaks in the time periods.

Designatable Unit 3 – Northeastern Newfoundland

The data available for DU 3 consists of angler (1969-2007) and commercial (1969 – 1992) catch data, and counts from 6-8 counting fences (mean of 7 per year). Estimates of abundance for the entire DU were calculated based on angler catch and effort data, adjusted for catch rates based on data from rivers with counting fences (Reddin and Veinott 2010, but see O’Connell 2003). Rivers with no angling catch were not included in the abundance estimates provided. Another challenge with these data is the large increase in abundance of salmon in the enhanced Exploits River, where extensive unused habitats were made available (Mullins et al. 2003). In some years, the Exploits and Gander rivers can account for nearly half the population of this DU and this swamping effect should be considered when examining trends for DU 3.

DU 3 has 127 documented salmon populations, with a substantial number of small streams that appear to have transient populations (juveniles are always present but adults return sporadically; C. Bourgeois, Dept. of Fisheries and Oceans, pers. comm.). The most recent estimate of adult abundance for DU 3 is 80,505 (51,883-109,267) from 2007, with 68,654 being small salmon, and 11,851 being large salmonxviii. The lowest abundance during the last three generations was in 2002 with 58,584 (Figure 18). The highest abundance during the last three generations was 141,968 in 1996. There were no significant trends in abundance for small, large or total salmon for this DU over the last three generations (P = 0.569, 0.947, and 0.618 respectively). The abundance of total salmon has declined by 9.5% over this time period (based on the curve fit in Figure 18), while small are 9.6% less abundant than three generations ago in 1994 (Figure 18). Large salmon abundance is estimated to have increase by 1.7% during this time period. As in Labrador, the non-significant trends in abundance, presented here for the past three generations, seem incomplete without considering the effects of commercial fishery closures that occurred in 1992 and remain in effect now. The returns data presented here do not include the commercial removals that were very high in the years up to 1991 (Reddin and Veinott 2010). Inclusion of these data is problematic because the landings include some salmon not originating from rivers within the DU. Reconstruction of pre-fishery abundance paints a picture of a substantial decline that has stabilized during the past 3 generations (DFO 2008). Additionally, more recent runs have not met increased expectations associated with improving escapement levels post-moratorium. Freshwater productivity has remained stable (DFO 2008) and there have been no reported extirpations of salmon in DU 3. Data from individual rivers monitored with counting fences (Exploits River, Gander River, Middle Brook, Terra Nova River and Campbellton River) are provided in Figure 19. Supplementary abundance data (for Indian Bay Brook, Northwest River and Indian River) are provided in Appendix 1.

Figure 18: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 3 (1969-2007)

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 3 from 1969 to 2007. Superimposed is the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Figure 19: Small (left panels) and Large (right panels) salmon Abundance from Counting Fence Facilities (Exploits, Gander, Middle, Terra Nova and Campbellton) of DU 3

Charts showing small (left panels) and large (right panels) salmon abundance from counting fence facilities on five rivers (Exploits, Gander, Middle, Terra Nova and Campbellton) for Designatable Unit 3.

(Taken from Reddin and Veinott 2010.)

Designatable Unit 4 – South Newfoundland

The data available for DU 4 consisted of angler (1969-2007) and commercial (1969 – 1992) catch data, and counts from 5 counting fences (mean of 4 per year) (Reddin and Veinott 2010). Angler catch data was based on a mean estimate of 20,527 rod days per year with a range of 12,208 – 32,642. There are 104 known rivers in this DU, with no known extirpations and one introduced population (Rocky River). Conne River had the highest estimated abundance over the time series, peaking at just over 10,000 returning adults. Most rivers in this DU appear to have mean abundances of less than 500 spawning adults (Dempson et al. 2006). Angling effort has declined by nearly 50% over the last 15 years. Estimates of abundance for the DU were calculated based on angler catch and effort data, adjusted using the catchability data from the rivers with counting fences (Reddin and Veinott 2010). The fishery-independent data from this DU are heavily biased to the eastern side of the DU and may not be representative of the entire DU. Furthermore, rivers with no angling catch were not included in the abundance estimates provided.

The most recent estimate of adult abundance for DU 4 is 21,866 (14,021-29,711) from 2007, with 18,633 (12,411-24,854) being small salmon, and 3,233 (1,610-4,857) large (Figure 20). The lowest abundance during the last three generations was in 2001 with 18,409. The highest abundance during the last three generations was 60,008 in 1996. The abundance of small salmon (based on the curve fit in Figure 20) declined by 37.3% since 1994. The abundance of large salmon has declined by 26.2% since 1994, and total salmon abundance has declined by 36.0% (Figure 20). Estimated declines in the abundance of small and total salmon are marginally insignificant (P = 0.063 and 0.071 respectively), but the estimated decline in large salmon abundance is not significantly different from zero (P = 0.293). It is worth noting that while trends in abundance were similar between catch data and counting facility data for this DU, the counting facility data and total catch information suggest that 2007 was the lowest year on record not 2001. Additionally, these decline rates are sensitive to the length of the time series used. Extending the time series back one additional year yields decline rates of 52.5% and 50.1% for small and total salmon respectively, both of which are statistically significant (P <0.01).

Previously published trends for individual populations, where counting fences exist, can be found in Figure 21. Supplementary abundance data (for Biscay Bay River) are provided in Appendix 1.

The Conne River has exhibited the most substantial decline, strongly influencing the total abundance for DU 4.

Figure 20: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 4 (1969-2007)

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 4 from 1969 to 2007. Superimposed is the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Figure 21: River-specific Trend Data from the Five Active Counting Facilities (Northeast Trepassey, Conne, Rocky, Northeast Placentia, and Little Rivers) in DU 4

Charts showing river-specific trend data from active counting facilities on five rivers (Northeast Trepassey, Conne, Rocky, Northeast Placentia and Little) in Designatable Unit 4. Data for small salmon (left panels) and large salmon (right panels) are presented separately for each river.

Data for small (left panels) and large salmon (right panels) are presented separately for each river (taken from Reddin and Veinott 2010).

Designatable Unit 5 – Southwest Newfoundland

The data available for DU 5 consisted of angler (1969 – 2007) and commercial (1969 – 1992) catch data, and counts from two counting fences. Five of the DU 5 rivers are also assessed with annual swim-through surveys. Angler catch data was based on a mean estimate of 25,899 rod days per year with a range of 18,544-38,487. Angling effort has increased significantly (P= 0.004); by nearly 240% over the data set. Estimates of abundance for the entire DU were calculated based on angler catch and effort data, adjusted using catch rate data from rivers with counting fences (Reddin and Veinott 2010). Furthermore, where angling data were unavailable, abundance was scaled according to available habitat. While these fishery-dependent data are corrected with fishery-independent data, estimates should be considered with the same caveats described above.

DU 5 has an estimated 40 rivers with salmon populations. There have been no known extirpations in this DU. The most recent estimate of adult abundance for DU 5 is 44,566 (32,143-56,988) from 2007, with 37,679 (27,828-47,531) being small salmon, and 6,886 (4,315-9,457) being large salmon. The lowest abundance during the last three generations was in 1991 with 15,488 salmon while the highest abundance was 68,441 in 2006. There was a significant increase in the abundance of small, large and total salmon (all P values < 0.001). The abundance of small salmon (based on the curve fit in Figure 22) is 132.1% greater than three generations previous. Over the same time period, the abundance of large salmon increased by 143.7, while total salmon abundance is 133.6% greater (Figure 22). Despite increasing trends and four of five monitored rivers meeting conservation requirements, population abundance in these rivers is considered low (DFO 2008). Trends for individual populations where counting fences exist can be found in Reddin and Veinott (2010). The Humber River is the largest population in this DU with abundance estimates ranging from 6,125 to 32,118 salmon. Abundance in populations south of the Humber, in the Bay St. George region, ranged from 235 to 3,684 salmon, with Harry’s River having the highest abundance estimates. Data for snorkel-surveyed rivers (Harry’s, Robinsons, Crabbes, Fischells and M. Barachois) are provided in Figure 23. Supplementary abundance data (for Highlands, Flat Bay, Humber and Grand Bank rivers) are provided in Appendix 1.

Figure 22: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 5 (1969-2007)

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 5 from 1969 to 2007. Superimposed is the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Figure 23: Abundance Estimates for Atlantic Salmon in Snorkel-surveyed Rivers of DU 5

Charts showing abundance estimates for Atlantic Salmon in snorkel-surveyed rivers (Harry’s, Crabbe’s, Fischells, Robinsons, M. Barachois) for Designatable Unit 5.

(Taken from Reddin and Veinott 2010.)

Designatable Unit 6 – Northwest Newfoundland

The data available for DU 6 consisted of angler (1969 – 2007) and commercial (1969 – 1992) catch data, and counts from three counting fences; although data are not available from the three fences in all years (Reddin and Veinott 2010). Angler catch data was based on a mean estimate of 15,517 rod days per year with a range of 10,386-19,695. Angling effort has decreased significantly (P= 0.004) to 82% of mid-90s values. The Torrent River has had a substantial amount of habitat made available as part of an enhancement project. Significant increases in abundance of this population may influence overall trends in the DU. Estimates of abundance for the entire DU were calculated based on angler catch and effort data, adjusted using catch rate data from rivers with counting fences (Reddin and Veinott 2010). Estimates should be considered with the same caveats described above.

There are 34 salmon rivers in DU 6, of which none have been extirpated. The most recent estimates of adult abundance for DU 6 is 31,179 (20,061-42,296) from 2007, with 26,603 (17,786-35,420-9,457) being small salmon, and 4,576 (2,275-6,876) being large salmon (Figure 24). Abundance estimates during the last three generations range from 19,369 salmon in 1994 to 51,570 salmon in 1996. There were no significant trends in the abundance of small, large or total salmon (P = 0.838, 0.125, and 0.999 respectively). The abundance of small salmon (based on the curve fit in Figure 24) has decreased by 4.2% over the last three generations. The abundance of large salmon is 41.7% greater over the same time period, and the trend line for the abundance of total salmon has a slope of zero over this time period (Figure 24). Abundance estimates were available from two monitored rivers in this DU in 2008 (Torrent River and Western Arm Brook) and both were above the conservation requirement (DFO 2008). Supplementary abundance data (for Lomond, Torrent rivers and Western Arm Brook) are provided in Appendix 1.

Figure 24: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 6 from 1969 to 2007

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 6 from 1969 to 2007. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the fit from the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Designatable Unit 7 – Quebec Eastern North Shore

Data from Quebec are derived from various methods, including direct counts (fence and snorkel surveys), extrapolations from index rivers (based on available habitat) and angler data (MRNF 2009, MRNF unpublished data ). The Ministère des Ressources naturelles et de la Faune in Quebec assigns a classification to the data for each river C1-C6 (C1 being the highest quality data) that rates the quality of the abundance data. Many of these classifications can include multiple data types (e.g., counting fences and snorkel swim-throughs). The general data classifications for the rivers in each DU are presented for DUs 7-10. DU 7 had four C3 rivers, three C5 rivers and eight C6 rivers.

All 15 salmon rivers of DU 7 were represented in the data set over the time period 1984 – 2008. Mean rod-days per year was 2,402 with a range of 1,892-3,230. Effort has been declining over the time series (P<0.001). The most recent estimate of adult abundance for DU 7 is 5,901 salmon in 2008, of which 69% were small salmon (Figure 25). Abundance estimates during the last three generations range from 4,026 salmon in 1997 to 7,785 salmon in 1993. There were no significant trends in small, large and total salmon abundance (P=0.085, P=0.115; P=0.297 respectively). The abundance of small salmon (based on the curve fit in Figure 25) declined by -26.3% during the last three generations; however, this decline was partially offset by a 50.8% increase in the abundance of (more fecund) large salmon, with the total number of salmon down by 13.8% (Figure 25). Supplementary abundance data (for the Musquanousse and Vieux Fort) are provided in Appendix 1.

Figure 25: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 7 from 1984-2008

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 7 from 1984 to 2008. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Designatable Unit 8 – Quebec Western North Shore Population

Data from Quebec are derived from various methods, including direct counts (fence and snorkel surveys), extrapolations from index rivers (based on available habitat) and angler data (MRNF 2009, MRNF unpublished data). DU 8 has three C1 rivers, nine C3 rivers, three C4 rivers, seven C5 rivers, and seven C6 rivers (See DU 7 for description of river data classification).

The 29 salmon rivers of DU 8 are represented over the time period 1984 – 2008. The most recent estimate (2008) of adult abundance for DU 8 is 15,135, of which 73% are large salmon. Abundance estimates during the last three generations range from 9,865 salmon in 2002 to 17,341 salmon in 1995. There were significant declines in small and total salmon abundance (P=0.031, P=0.013 respectively). A significant trend was not associated with large salmon abundance (P=0.143). Over the last three generations, the abundance of small salmon (based on the curve fit in Figure 26) declined by 33.9%, while large salmon declined by 20.1% and total salmon by 24.4% (Figure 26).

Data for de la Trinité river, an index river monitored with a fish ladder, is provided in Figure 27. Supplementary abundance data (Laval, Mistassini, Godbout, de la Trinité, aux Rochers, Jupitagon, Mingan, de la Corneille, Piashti, Watshishou, Petite Rivière de la Watshishou, des Escoumins) are provided in Appendix 1. There have been no populations lost from DU 8.

Figure 26: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 8 from 1984-2008

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 8 from 1984 to 2008. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the fit from the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Figure 27: Quebec Index Rivers (Saint-Jean and Trinité)

Chart showing counting fence data for Quebec index rivers (Saint-Jean and Trinité) from 1984 to 2008.

Counting fence data from 1984-2008. Note the Saint-Jean lies within DU 12 while the Trinité is within DU 8.

Designatable Unit 9 – Anticosti Island

Data from Quebec are derived from various methods, including direct counts (fence and snorkel surveys), extrapolations from index rivers (based on available habitat) and angler data (MRNF 2009, MRNF unpublished data). Salmon abundance data is available from 25 rivers on Anticosti Islandand 24 of them were classified according to the type of data available. DU 9 has one C1 river, one C3 river, 19 C4 rivers, and three C6 rivers (See DU 7 for description of river data classification).

The most recent estimate (2008) of adult abundance for DU 9 is 2,414 salmon, comprised of 1,362 small and 1,052 large salmon. Abundance estimates during the last three generations range from 1,390 salmon in 2005 to 4,855 salmon in 1996. The declining trend in abundance detected for small salmon (Figure 28) was marginally insignificant (P = 0.077), and statistically significant declines in large and total salmon were observed (respective P-values: 0.017 and 0.007). The abundance of total salmon (based on the curve fit in Figure 28) has declined by 31.7% over the last 3 generations. The abundance of both large (48.7%) and small (40.2%) salmon has declined during this period. Supplementary abundance data (á l’Huile, MacDonald, á la Patate, Vaureal, aux Saumons, du Renard, Petite rivière de la Loutre, Bell, Box, Dauphine, Petite rivière de la Chaloupe, Maccan, de la Chaloupe, Ferree, Martin, du Pavillon, aux Plats, Chicotte, Galiote, du Brick, Jupiter, à la Loutre, Bec-scie ) are provided in Appendix 1. There have been no populations lost in DU 9.

Figure 28: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 9 from 1984-2008

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 9 from 1984 to 2008. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the fit from the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Designatable Unit 10 - Inner St. Lawrence

Data from Quebec are derived from various methods, including direct counts (fence and snorkel surveys), extrapolations from index rivers (based on available habitat) and angler data (MRNF 2009, MRNF unpublished data). The nine known salmon rivers of DU 10 are represented in the dataset. DU 10 has six C1 rivers, and three C4 rivers (See DU 7 for description of river data classification).

The most recent estimate (2008) of adult spawner abundance for DU 10 is 4,169 salmon, the highest over the last three generations, consisting of 2,230 small salmon and 1,939 large salmon.The lowest spawner abundance during the last three generations was in 2007 (2,208 salmon). There were no significant trends in abundance for small, large or total salmon (small: P=0.951; large: P=0.429; total: P=0.772; Table 2). The abundance of large and total salmon (based on the curve fit in Figure 29) has increased by 11.5% and 5.3% respectively since 1997, while small salmon abundance has declined by 1.8% during this time period. Supplementary abundance data (Ouelle, Malbaie, St.-Jean, à Mars, Ste.-Marguerite principale, Ste.-Marguerite NE) are provided in Appendix 1.

Despite relatively stable trends, effective population sizes for salmon in the rivers of DU 10 are relatively low (Dionne et al. 2007). Furthermore, many populations in this area have been supplemented by stocking (M. Dionne, Quebec Ministère des Ressources naturelles et de la Faune, pers. comm.). To date, all known salmon rivers contain populations.

Figure 29: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 10 from 1984-2008

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 10 from 1984 to 2008. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the fit from the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Designatable Unit 11 - Lake Ontario

The Lake Ontario DU has been assessed as extirpatedxix (COSEWIC 2006a). Attempts are ongoing to re-establish populations through stocking. Since no known genetic material remains from the original populations, different strains are being used for restoration efforts. These efforts have not yet succeeded in producing self-sustaining, naturally reproducing populations.

Designatable Unit 12 – Gaspé–Southern Gulf of St. Lawrence

DU 12 has 78 rivers that contain salmon populations distributed across four provincial jurisdictions (Quebec, PEI, Nova Scotia, and New Brunswick). The data available for DU 12 came from a variety of sources as the DU is comprised of several Quebec and Gulf Salmon Fishing Areas. The specific data sources and collection details can be found in (Breau et al. 2009, Cairns et al. 2009, MRNF 2009, MRNF unpublished data, Cameron et al. 2009, Chaput et al. 2010, Fournier and Cauchon 2009, Secteur Faune Québec 2009, Dionne et al. 2010). Broadly, the data consist of angler catch statistics (1970-2008), counts from up to nine counting fences (range 6 - 9), snorkel surveys, and mark-recapture estimates. The primary estimate of abundance for the whole DU is based on the angler-catch data. While these fishery-dependent data are corrected with fishery-independent data, estimates should be considered with the same caveats described above.

The latest estimate (2007) of adult spawner abundance for DU 12 is 103,149 salmon. The lowest abundance during the last three generations was in 1999 with 77,323 salmon, while the highest abundance was 213,329 salmon in 1993. There were no statistically significant trends in the abundance of small, large or total salmon in this DU (P values: 0.119, 0.217 and 0.100 respectively). The abundance of small, large and total salmon (based on the curve fit in Figure 30) has decreased by 34.0%, 18.5% and 27.8% respectively over the last three generations. These values are sensitive to the length of the time series. For example, increasing or decreasing the length of the time series for total salmon changes the decline rate estimates to 46% or 1.5% respectively. The Miramichi River accounts for the majority of salmon in this DU (>50% of the total DU population in the majority of years). The swamping effect of this single large river should be considered when examining these data. In general, juvenile distribution and densities are good and most rivers are known or are suspected of meeting conservation requirements (Breau et al. 2009, Cameron et al. 2009, Chaput et al. 2010). Southern areas of SFA 16 and PEI are exceptions, as distribution of juveniles is sparse and densities are low (Cairns et al. 2009, Chaput et al. 2010). Adult salmon abundance in the latter areas is also considered to be below conservation levels (Cairns et al. 2009, Chaput et al. 2010). Furthermore some small rivers of the Northumberland Strait also appear to be in decline (Gibson et al. 2006). PEI in particular is experiencing significant habitat degradation, related to land-use issues and its indigenous stocks have likely been largely replaced by stocked fish in at least some rivers (D. Cairns, Dept. of Fisheries and Oceans, pers. comm.). Abundance data from counting fence facilities and/or dominant rivers of DU 12 are provided (Figures 31-35). Supplementary abundance data (Matapedia, Cascapedia, Petite rivière Cascapedia, Bonaventure, Petite rivière Port Daniel, Port Daniel du Milieu, Port Daniel Nord, du Grand Pabo Ouest, du Grand Pabo, du Petit Pabo, Grande Rivière, St.-Jean, York, Dartmouth, Madeleine, Ste.-Anne, Cap Chat, Matane, Mitis, Restigouche, Nepisiguit, Tabusintac, Bouctouche, Morell, Philip, East Pictou, Sutherlands, West Antigonish) are provided in Appendix 1.

Figure 30: Atlantic Salmon Returns (small: top panel; large: middle panel; total: bottom panel) for DU 12 Over the Past 3 Generations

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 12. Superimposed is the fit from the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance.

Figure 31: Counts of All Adult Salmon at the Northwest Upsalquitch Barrier (upper) and Causapscal Barrier (bottom), Restigouche River

Charts showing counts of all adult salmon at the Northwest Upsalquitch Barrier (upper panel) and Causapscal Barrier (bottom panel) on the Restigouche River.

Taken from Cameron et al. 2009.

Figure 32: Counts of Salmon at the Jacquet River Barrier

Chart showing counts of salmon at the Jacquet River barrier.

Square black symbols show years with incomplete counts due to fence washouts or early removal due to inclement weather (taken from Cameron et al. 2009).

Figure 33: Counts of Salmon (size groups combined) at the Two Headwater Barriers in the Southwest Miramichi (upper panel), at the Single Headwater Barrier in the Northwest Miramichi (middle panel) and Catch per Rod Day from the Crown Reserve Angling Waters of the Northwest Miramichi (lower panel)

Charts showing counts of salmon (size groups combined) at the two headwater barriers in the Southwest Miramichi (upper panel), at the single headwater barrier in the Northwest Miramichi (middle panel) and catch per rod day from the crown reserve angling waters of the Northwest Miramichi (lower panel).

(Taken from Chaput et al. 2010.)

Figure 34: Estimates of Returns of Small Salmon (upper), Large Salmon (middle) and Size Groups Combined (lower) to the Miramichi River, 1971 to 2007

Chart showing estimates of returns of small salmon (upper panel), large salmon (middle panel) and size groups combined (bottom panel) to the Miramichi River from 1971 to 2007.

Trend line is an exponential function for the most recent 15 years (1993 to 2007) (taken from Chaput et al. 2010).

Figure 35: Estimated Returns of Large (upper series with error bars) and Small Salmon (lower series with error bars) to the Margaree River, 1987 to 2008

Chart showing estimated returns of large (upper series with error bars) and small salmon (lower series with error bars) to the Margaree River from 1987 to 2008.

The conservation requirement for large salmon is depicted with a solid line and for small salmon with a dashed line (taken from Breau et al. 2009).

Designatable Unit 13 – Eastern Cape Breton

The data available for DU 13 came from a variety of sources including angler catch statistics (1970-2008), fishway counts (1 river), snorkel surveys on four rivers 1994-2008 (except Clyburn 1987-2008) and mark-recapture estimates. Where angler data has been used, its utility as an index has been validated using fishery-independent methods. Data reflect both returns and escapement – depending on the data source. There was no total estimate of abundance available for this DU, but low angler effort on other rivers suggests much of the salmon abundance in this DU is within assessed rivers (Gibson and Bowlby 2009). The spawner abundance data presented here are a sum for rivers with estimates (based on the data provided in Gibson and Bowlby 2009). Since Grand River data was not provided in terms of small and large salmon, data from this river are included only for total salmon. As such the results provided for total salmon do not equal the sum of small and large individuals.

There are 30 rivers in DU 13 with reported recreational catch. The most recent (2008) estimate of adult abundance for DU 13 is 1,150 salmon, of which 407 were small, and 743 were large. During the last three generations, total abundance in the five assessed rivers has ranged from 513 salmon in 2002, to 1,825 salmon in 1996. There were no significant trends in the abundance of small, large or total salmon (P = 0.789, 0.542, and 0.202 respectively) when the abundance time series for this DU are analyzed in aggregate. The abundance of small salmon (based on the curve fit in Figure 36) has declined by 7.9% since 1993, whereas the abundance of large salmon is 14.5% below 1993 levels. The abundance of salmon for both size categories combined has decreased by 28.9% during this time period (Figure 36). Despite the lack of a statistically significant declining trend over three generations, four of five DU 13 rivers were below conservation requirements in 2008 and two had “marked” declines (Gibson and Bowlby 2009). Furthermore, a declining trend can be detected for small (39.6% over four generations; P = 0.058), large (67.2%; P < 0.001) and total (69.1%; P < 0.001) salmon when the data series is extended by five years (four generations). The difference in the trends in total abundance from the large and small abundance series reflects the large decline in abundance in the Grand River (Figure 37), which was not included in the small and large abundance series. Data for individual river systems are plotted in Figure 37. Juvenile abundance levels in the region are not high in comparison to DU 12 rivers, although juveniles remain widespread (Gibson and Bowlby 2009). To date, there have been no known extirpations in this DU.

Figure 36: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 13 Over the Past 3 Generations

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 13 over the past three generations. Superimposed is the fit from a general linear model used to determine trends in abundance.

Superimposed is the fit from a general linear model (+/- 2SE prediction intervals) used to determine trends in abundance. Note contributions from the Grand River are not included in small and large salmon plots due to data limitations.

Figure 37: Adult Atlantic Salmon Abundance Time Series (size categories combined) for Five Eastern Cape Breton Rivers

Charts showing Adult Atlantic Salmon abundance time series (size categories combined) for five eastern Cape Breton rivers (North, Baddeck, Clyburn, Grand, Middle).

The solid line is the estimated abundance from a log-linear model fit to data for the last three generations. The dashed line shows the 5-year mean abundance for 2 time periods separated by 15 years. The points are the observed data (taken from Gibson and Bowlby 2009).

Designatable Unit 14 – Nova Scotia Southern Upland

The data available for DU 14 come from a variety of sources including angler catch statistics (1970-2008), fishway counts (3 rivers), and mark-recapture estimates (1 river). The trend data used for this section rely entirely on fishery-independent data: the sum of the spawner escapement counts on the two main index rivers was used to assess trends. Abundance estimates from the assessed rivers are not extrapolated to the entire DU using the recreational catch because most rivers are closed to fishing. As such there is no total estimate of abundance available for this DU. The abundance data presented here are a sum for rivers with estimates (based on data in Gibson et al. 2009). In recent years, the monitored rivers are biased towards systems with lower acidification impacts. Such rivers, however, are thought to currently contain the majority of salmon in this DU.

Within the previous century, 63 rivers with this DU are known to have contained salmon, although presently, salmon are extirpated from many. The most recent estimate (2008) of adult abundance for the two index rivers is 1,427 salmon, consisting of 1,264 small and 164 large salmon. The lowest abundance during the last 3 generations was 755 salmon in 2007, while the highest abundance was 3,557 salmon in 1996. Abundance of salmon in this DU during the 1980s at times exceeded 10,000. There has been a significant decline in the abundance of small (P = 0.003), large (P = 0.002) and total salmon (P < 0.001) in this DU based on the curve fit in Figure 38. Small salmon abundance declined by 58.6% since 1996 (Figure 38). The abundance of large salmon was down by 74.0%, and total salmon declined by 61.3% during that period. Since recent counts represent systems with relatively low levels of acidification, declines in acidified rivers of DU 14 are expected to be greater (Gibson et al. 2009). DU 14 has experienced a substantial decline in the number of individual populations. DFO (2000) predicted that 55% of rivers in this DU are extirpated with an additional 36% at risk of extirpation.

A comparison of juvenile abundance estimated from electrofishing surveys between 2000 and 2008 (Gibson et al. 2009) are indicative of ongoing declines and low juvenile abundance (Figure 39). These surveys were similar in terms of total effort and coverage, although marginally more sites were completed in 2008 (143 vs. 128), but one less river was visited (51 rather than 52). Total shocking time was slightly greater in 2008 (143,385 seconds vs. 104,331 seconds), but the total area surveyed was lower (98,019 m2 vs. 128,841 m2). Approximately one-quarter as many juvenile salmon were captured in 2008 (977 salmon) than in 2000 (3,733 salmon). In 2000, juvenile Atlantic Salmon were found in 54% of the rivers (28 of 52), but were only found in 39% (20 of 51) of the rivers in 2008.

Under current conditions, maximum lifetime reproductive rates (indicative of the compensatory reserve) of salmon in this DU are very low and abundance will likely continue to decline because the populations have little intrinsic capacity to rebound following events that further lower abundance (Gibson et al. 2009). Only a few populations (e.g. the LaHave and St. Mary’s rivers) may be viable under current conditions and then only at low population size (Gibson et al. 2009). Because of their low reproductive rates, these populations may also be at risk as a result of stochastic processes. Annual salmon counts at the Morgan Falls fishway on the LaHave River, the primary index of abundance in this DU, are provided in Figure 40. Supplementary abundance data (Liscomb, St. Marys, and East River (Sheet Hbr.)) are provided in Appendix 1.

Figure 38: Atlantic Salmon Escapement from 1980 to 2008 (small: top panel; large: middle panel; total: bottom panel) for DU 14

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 14 from 1980 to 2008. Superimposed is the general linear model used to determine trends in abundance over the past three generations.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance over the past 3 generations.

Figure 39: Box Plots Showing the Density of Atlantic Salmon in Southern Upland Rivers Based on Electrofishing During 2000 and 2008

Box plots showing the density of Atlantic Salmon in Southern Upland rivers based on electrofishing during 2000 and 2008.

The dot shows the median density and the box shows the inter-quartile spread. Open dots indicate that no salmon were captured in the river. The whiskers are drawn to the minimum and maximum. “N” is the number of sites that were electrofished in each river (adapted from Gibson et al. 2009).

Figure 40: Counts of Atlantic Salmon at Morgans Falls Fishway on the LaHave River, NS, from 1974 to 2008, Divided into the Proportions of Wild-origin and Hatchery-origin 1SW and MSW Adults

Chart showing counts of Atlantic Salmon at Morgans Falls fishway on the LaHave River, from 1974 to 2008, and indicating the proportions of wild-origin and hatchery-origin one-sea-winter and multi-sea-winter adults.

(Taken from Gibson et al. 2009.)

Designatable Unit 15 – Inner Bay of Fundy

This DU has been designated as Endangered under the SARA. A full status report was prepared in 2006 (COSEWIC 2006b). Current estimates for this DU (2008) suggest the total number of wild fish is likely to be less than 200 individuals.

Designatable Unit 16 – Outer Bay of Fundy

Small and large returns to the Saint John River from 1993 to 2008 were calculated by using the estimated returns to the Nashwaak River (upriver of the counting fence), raised by the amount of habitat available in the Saint John River downstream of Mactaquac Dam plus the total returns destined for above Mactaquac Dam. The returns to the other outer Bay of Fundy rivers were determined by using the total returns to both the Magaguadavic and St. Croix rivers raised by the amount of habitat available to salmon between the Saint John River and the Maine border. Added to the estimated Saint John River returns, these estimates provided the total estimated 1SW and MSW returns to DU 16 (Jones et al. 2009).

There are 17 salmon rivers in DU 16. The most recent estimate of adult abundance for DU 16 is 7,584 from 2008. Of these 6,629 were small and 955 were large. The lowest abundance during the last three generations was in 2007 (3,486 salmon). The highest abundance during the last three generations was 20,010 salmon in 1996. There have been significant declines in the abundance of large (P < 0.001), small (P = 0.024) and total salmon (P = 0.001). The abundance of small salmon (based on the curve fit in Figure 41) has declined by 56.5% since 1996 (Figure 41). The abundance of large salmon has declined by 81.6% of 1996 abundance and total salmon are down by 64.3%. Adult escapement is well below conservation requirements for the entire area and juveniles, though well distributed, are also at low densities (Jones et al. 2009). While all monitoring facilities show strong declining trends, the St. Croix and the Magaguadavic rivers have been effectively extirpated of wild fish. Data from the Saint John River (Mactaquac), Magaguadavic River and St. Croix River are provided in Figures 42-44.

Figure 41: Atlantic Salmon Escapement (small: top panel; large: middle panel; total: bottom panel) for DU 16 Over the Past 3 Generations

Charts showing small (top panel), large (middle panel) and total (bottom panel) Atlantic Salmon escapement for Designatable Unit 16 over the past three generations. Superimposed is the general linear model used to determine trends in abundance.

Superimposed is the general linear model (+/- 2SE prediction intervals) used to determine trends in abundance.

Figure 42: Estimated Total Adjusted Returns of Wild and Hatchery 1SW and MSW Salmon Destined for Mactaquac Dam, Saint John River, 1970–2008

Charts showing estimated total adjusted returns of wild and hatchery one-sea-winter (upper panel) and multi-sea-winter (lower panel) salmon destined for the Mactaquac Dam, Saint John River, from 1970 to 2008.

(Taken from Jones et al. 2009.)

Figure 43: Trends in Abundance of Adult Atlantic Salmon in the Magaguadavic River During the Last 15 Years

Chart showing trends in abundance of adult Atlantic Salmon in the Magaguadavic River during the last 15 years.

The solid line is the predicted abundance from a log-linear model fit by least squares. The dashed lines show the 5-year mean abundance for 2 time periods separated by 15 years. The points are the observed data (taken from Jones et al. 2009).

Figure 44: Trends in Abundance of Adult Atlantic Salmon in the St. Croix River During the Last 15 Years Assessed (1992-2006)

Chart showing trends in abundance of adult Atlantic Salmon in the St. Croix River during the last 15 years assessed (from 1992 to 2006).

The solid line is the predicted abundance from a log-linear model fit by least squares. The dashed lines show the 5-year mean abundance for 2 time periods separated by 15 years. The points are the observed data (taken from Jones et al. 2009).




Footnotes

xv Modal smolt ages were derived from Appendix 3 (large and small salmon combined) in Chaput et al. (2006a), except for DU 7 (Appendix 1 - small salmon), DU 8 (Appendix 2 - large salmon) and DU 10 (estimated from Figure 3), where data were lacking.
xvi This DU was listed as “extirpated” in COSEWIC 2006a; however, current interpretation of the meaning of “DUs” requires that if a DU is lost, its unique elements cannot be recovered. As such, the authors have been advised that “extinct” is a more appropriate description.
xvii Elements of this section have been copied, abstracted and/or synthesized from DFO and MRNF (2009).
xviii Large salmon in DU 3 are comprised almost exclusively of repeat spawning grilse as opposed to maiden multi-sea -winter fish.
xix This DU was listed as “extirpated” in COSEWIC 2006a; however, current interpretation of the meaning of “DUs” requires that if a DU is lost, its unique elements cannot be recovered. As such, “extinct” is a more appropriate description.

Page details

Date modified: