March 23, 2009 version
Do not cite or circulate without permission
Vulnerability of Fraser River Sockeye salmon to climate change: a lifecycle perspective using expert judgments
By
Tim McDaniels, Sarah Wilmot, Michael Healey and Scott Hinch
University of British Columbia
Contact information:
Email: timmcd@interchange.ubc.ca
Tel: 604-822-9288
Fax: 604-822-3787
This is a draft manuscript, submitted for editorial review. Do not cite or distribute without permission.
Comments are welcome.
Abstract
Fraser River sockeye salmon are the basis for a major commercial fishery shared by Canada and the
United States, and an important cultural foundation for many aboriginal groups; they are also of huge
ecological significance throughout the Fraser Basin. The potential for altered aquatic habitat and
temperature regimes due to climate change is an important concern for Fraser River sockeye salmon. In
this paper we characterize the vulnerability of Fraser sockeye salmon to future climate change by an
approach that is novel on three counts. First, previous efforts to assess vulnerability of salmon to climate
change have typically focused on only part of the life cycle whereas we consider climate vulnerability at
all stages in the life cycle of sockeye salmon. Second, we use the available scientific literature to provide
a basis for structuring and eliciting judgments from fisheries science and management experts who do
research on and manage these systems. Third, we also consider prospects for mitigation of the effects of
climate change in sockeye salmon. Tests showed that participants' judgments differentiated among the
questions, life stages, spawning regions and climate scenarios in statistically significant ways. The
consensus of scientists and managers was that Fraser sockeye are most vulnerable to climate change
during the returning adult and egg stages of the life cycle. A high temperature scenario was seen as
imposing greatest risk on sockeye stocks, particularly those which migrate to the upper reaches of the
Fraser system, and spawn earlier in the summer. The inability to alter water temperature, and highly
constrained nature of sockeye management, with competing gear types and sequential fisheries over a
long distance, all suggest that the potential to mitigate these adverse effects is be limited. Fortunately,
Fraser sockeye already demonstrate a great deal of adaptive capacity in utilizing heterogeneous habitats
in different river basins. Hence, this adaptability points to the potential value of policies that seek to make
stocks more resilient to uncertain futures.
1. Introduction
For thousands of years salmon have spawned in the Fraser River, their annual spawning runs from the
ocean providing an abundant harvest for aboriginal and, more recently, commercial fisheries. Northcote
and Larkin (1989) have described the Fraser as the most important salmonine-producing river system in
North America. The fishery for sockeye salmon (Oncorhynchus nerka) returning to the Fraser River is the
most economically valuable commercial salmon fishery in Canada. This harvest is shared with fishers
from the United States under the terms of the Pacific Salmon Treaty (Williams, 2007). Aboriginal
communities along the Fraser River also depend on the sockeye runs to provide salmon for food and
ceremonial purposes. Sockeye salmon are the second most abundant salmon species in the Fraser
River, after pink salmon (O. gorbusha), which has a much lower economic value.
As well as being economically and culturally important, the annual runs of salmon are ecologically
significant; the spawning fish provide food for predator species from bears to eagles, and their
decomposing carcasses release important nutrients into floodplain, riverine and lacustrine ecosystems
(Helfield and Naiman 2001, Finney et al. 2000, Johnston et al. 2004). Most Pacific salmon spawn in the
mainstem and tributaries of the lower 200 km of the Fraser. However, sockeye salmon, which return from
the ocean each year in the 100s of thousand to millions, migrate to spawning beds distributed from near
the river mouth to more than 1000 km upstream in important tributaries in the Fraser. Thus, sockeye
distribute their protein and nutrients through much of the 230,000 km 2 of the Fraser basin.
In the past, sockeye have been able to weather both human and environmental assaults. The species
has, for example, persisted through the many periods of glaciation that occurred over the past 2 million
years. Fraser sockeye have also been resilient in response to intensive aboriginal fisheries that
developed after the Wisconsin glaciation and the even more intense colonial fisheries that developed in
the late 1800s. However, recent events and anticipated future conditions under global climate change
raise concerns about whether sockeye may be commercially abundant in the Fraser 50 or 100 years from
now or indeed anywhere in British Columbia.
The role of changing environments in driving down sockeye abundance is uncertain; however,
exceptionally warm temperatures in both the river and ocean have been implicated (Cooke et al. 2004,
Macdonald et al. 2000). Extremely poor survival of sockeye in the river during 2004 has been linked to
exceptionally high temperatures (Williams, 2005) although the environmental analysis could account for
less than half the actual loss. Fraser sockeye survival in 2005-2007 was also poor (for various reasons in
each year), leading to highly constrained commercial fisheries (M. Lapointe, Pacific Salmon Commission,
pers.comm, 2008; Pacific Salmon Commission, 2009,a,b,c). Continued increases in the temperature of
freshwater migratory areas, nurseries and oceanic feeding areas associated with global warming have
the potential to severely depress Fraser River sockeye stocks. Various authors have examined the
potential impact of global warming on particular life stages of sockeye salmon (e.g., Henderson et al.
1992, Farrell et al. 2008, Rand et al. 2006, Hinch et al. 1995a, b, Swansburg et al. 2002) as well as on
overall production rates (e.g., Klyashtorin 1998, Levin 2003) and concluded that the consequences may
be severe.
The objective of this paper is to characterize the vulnerability of Fraser sockeye salmon to future climate
change. Its approach is novel on three counts. First, previous efforts to consider vulnerability of salmon to
climate change have generally only focused on part of the life cycle, (e.g. Hinch et al. 1995b; Henderson,
et al, 1992; Rand et al 2006.). Pacific salmon are anadromous, occupying various fresh water and marine
environments at different life stages. Hence, over their life cycle, they integrate the effects of climate
change on fresh water and marine ecosystems, and the consequences for one generation can carry
forward to influence reproductive success, abundance and ecological fitness in subsequent generations.
To date, no one has explored the way changes at one life stage propagate through the life cycle to affect
other stages or accumulate across generations1. Here we consider the effects of climate change on
vulnerability of sockeye salmon to changing habitat at all life cycle stages. Second, while expert
judgments are used extensively in risk and decision analysis, often in the form of probabilities of specific
events, the approach adapted here is novel in considering vulnerability over life cycle stages, climate
scenarios and regions of the Fraser system. Third, we also consider prospects for mitigation of the effects
of climate change in sockeye salmon.
2. Scientific Understanding regarding Climate Change and Fraser
Sockeye
2.1
Sockeye Salmon Abundance Cycles
Fraser River sockeye populations are characterized by a 4-year cycle of abundance with one or two years
in four having much greater abundance than other years. The causes of this cycle remain unknown
1
M. Healey has a manuscript in preparation that will explore the cumulative effects of climate change on
salmon production.
(Ricker 1997). Fraser sockeye also fluctuate in abundance in relation to decades-long cycles in ocean
environmental conditions (Mantua et al. 1997), increasing in abundance during some ocean regimes and
decreasing during others.
Presumably in response to one of these oceanic regime changes, Fraser sockeye began increasing in
abundance rather dramatically at the end of the 1970s, reaching historic high abundance in the mid
1990s. After 1995, abundance in terms of annual harvest surplus began to decline, as ocean conditions
changed again (M. Lapointe, pers. Com.)
The Fraser River (as well as several other BC salmon systems) has experienced an average increase in
peak summer water temperature of >1.5°C over the past 40 yrs. Eight of the past 10 summers have been
the warmest on record and climate models predict even warmer peak temperatures in the near future,
with a 2-4°C average warming of the Fraser River over the next few decades and peak summer
temperatures occurring earlier (Morrison et al. 2002). Most of the 150 stocks in the Fraser (i.e.
populations) initiate adult river migrations between July and October, encountering a range of
temperatures (The 40-year average daily temperature are lowest in July and October (~ 15°C), and
highest in August (~ 19°C); Patterson et al. 2007). For the past 10 years, stocks that once migrated in the
fall, including the world-famous Adams River run, have altered their historic migration behavior for
reasons not fully understood [though possibly related to ocean environmental changes (Cooke et al.
2006)] and have begun river migration several weeks earlier than normal, now migrating during August
and September (Cooke et al. 2004). Because of this changed behavior, and because the Fraser is much
warmer in summer than the past, segments of all Fraser sockeye stocks now encounter river
temperatures >19°C during a portion of upstream migration. This has profound evolutionary significance
because no sockeye stock anywhere in the world is known to have initiated river migration at 20°C
(Hodgson & Quinn 2002). River temperature is therefore an important factor in the search for causes of
the current declines in stock abundance (Crossin et al. 2008). Inadequate monitoring and enforcement
and illegal harvesting of salmon in the Fraser River are also implicated in the disappearance of fish prior
to spawning (Williams, 2005). Regardless of the specific causes, the impact on commercial and aboriginal
fisheries has been dramatic.
2.2
Sockeye Salmon Life Cycle
For purposes of this paper, we distinguish among eight stages in the sockeye life cycle. Different stages
require different habitat, as shown in Table 1. General details about sockeye salmon life history can be
found in Burgner (1991). Details on life-stage specific movements and migrations for sockeye salmon can
be found in Hinch et al. (2006).
Table 1.
Life Cycle Stages of Sockeye Salmon
Life Cycle Stage
Habitat
Egg
Fresh water (lakes and rivers)
Fry - emergence to lake entry
Fresh water (lakes and rivers)
Fry - lake entry to first winter
Fresh water (lakes and rivers)
Fry - first winter
Fresh water (lakes and rivers)
Fry - smolt
Fresh water, moving towards ocean
Post Smolts
Estuary and coastal ocean
Immatures
Ocean
Returning Adults
Ocean, coastal, estuary, and fresh water rivers
and lakes
To simplify our analyses, we grouped Fraser River sockeye stocks into three broad categories based on
the location of their spawning grounds along the Fraser River: northern, middle and southern. These
three groups of stocks generally return to the river at different times, have different cycles of abundance,
and occupy different habitats during the fresh water life cycle stages shown in Table 1. Our classification
scheme creates some unavoidable overlap among groups in some of these variables.
2.3
Environmental Changes Caused by Climate Change That Will Affect Fraser River Salmon
Global climate change is expected to result in higher average temperatures and altered precipitation
patterns in the Fraser River basin, which will alter the habitat for sockeye salmon throughout their life
stages. The Pacific Climate Impacts Consortium provides access to a range of climate change scenarios
for the Fraser Basin, drawing on various global circulation models (GCMs) and future greenhouse gas
emission scenarios. There is a high level of uncertainty within and among the scenarios and models,
although all indicate warmer air temperatures and changes in precipitation patterns. Over all scenarios of
emissions and economic activities, and over all models, estimates of average annual increase in air
temperature for the Fraser Basin by 2100 range from about 2.5 ° C to about 6 ° C compared to the
average for 1980 to 1999 (PCIC, 2008).
Detailed analyses have been conducted to assess likely changes specifically in Fraser River flow and
temperature patterns. Morrison et al. (2002), using historic data from the past 50 years, found that
summer temperatures have been increasing and peak flows decreasing. Based on GCM predictions,
these trends are expected to continue. Specifically, for the period 2070–2099, the flow model predicted a
modest 5% (150 m3/s) average flow increase but a decrease in the average peak flow of about 18%
(1600 m3/s). These peaks would occur, on average, 24 days earlier in the year even though for 13% of
the years the peak flow occurred much later as a result of summer or fall rain, instead of the currently
normal spring freshet. In the same period, the summer mean water temperature is predicted to increase
by 1.9 °C.
Freshwater spawning and rearing areas are expected to change. Using output from GCMs in conjunction
with empirical time-series analyses of hydrologic data, Moore (1991) and Levy (1992) argued that
within the next 80-100 years, the streams, lakes and groundwater in the Fraser Basin will warm by 1-4 °C;
there would be an increase in stream velocities in winter and spring though an decrease in summer and
fall, and earlier spring freshets. Altered runoff patterns are predicted to make the Fraser drainage more
oligotrophic, resulting in a reduction in the abundance and availability of food for juvenile sockeye which
are rearing in lakes (Henderson et al. 1992).
Key environmental features in the the North Pacific Ocean are expected to change. Simulation results
from GCMs coupled to a simplified mixed-layer ocean model predict that within the next 80-100 years,
sea surface temperatures will warm 2-4 °C, and north-south air pressure gradients will weaken leading to
reductions in surface winds, wind-driven upwelling and zooplankton biomass, Food production for ocean
sockeye may decrease by 5-9% (Hinch et al. 1995).
The broad potential impacts of climate change on sockeye habitat in the Fraser Basin are summarized in
Table 2.
Table 2.
Broad Impacts of Climate Change on Sockeye Habitat
Habitat
Impact of Climate Change
Spawning streams

during spawning
and incubation
Spawning streams will be warmer during the normal late
summer/autumn spawning time and through the incubation period.

Spring freshet will be earlier and may be of shorter duration due to the
decrease in snow pack.

Discharge may be somewhat lower in later summer/autumn but may
remain higher over the winter if more winter precipitation falls as rain.

Coastal spawning streams will be warmer and may be subject to
stronger winter freshet if winter storms intensify.

Temperatures in summer will be higher, particularly in smaller streams,
which may cause stress for stream resident species like coho
(Oncorhynchus kisutch ) and Chinook (Oncorhynchus tschawytscha)
but may be less important for sockeye.
Nursery lakes

Interior lakes will lose ice cover earlier in the year (or may not be ice
covered at all), may stratify thermally earlier and remain stratified longer
and will have higher summer epilimnion temperatures.

Coastal nursery lakes may also stratify longer and have higher summer
temperatures.
Coastal marine

habitats
Coastal waters of British Columbia will be warmer and productivity will
probably be higher.

Spring plankton blooms will probably occur earlier.

Along the coast of Alaska, where downwelling conditions occur, there
may be nutrient limitation due to reduced upwelling in the central Gulf of
Alaska.
Gulf of Alaska

Surface temperatures will be warmer and gyre circulation may be
slower due to weakening of the Aleutian low.

Upwelling may be reduced and production lower.
3. Method
3.1
Role of Judgments from Scientific and Management Specialists
Any method for assessing and managing risks will necessarily rely on the informed judgments of technical
specialists and scientists to some extent. Methods have been developed and applied to characterize
judgments about uncertainties in the form of probabilities for specific well-defined events (Morgan and
Henrion, 1990; Keeney and von Winterfeldt, 1989; 1986). Probability elicitation has been employed in
various climate change contexts, including, for example, changes in ocean circulation patterns ( Zickfeld,
et al 2007). However these methods are time intensive, require extensive commitment of those who
provide the judgments, and typically only focus on one variable at a time. Other approaches to eliciting
judgments include seeking consensus among a nominated group through iterative responses to specific
questions ( e.g. the Delphi method (Linstone and Turoff, 1975), or consensus reports derived through
negotiation and review, such as the Intergovernmental Panel on Climate Change process. In the present
study the concern is with interactions over a wide range of variables for eight different life cycle stages
and three different regions given climate regimes well outside the current range of variability. Hence, a
judgment-based approach was used to characterize vulnerability to climate change, but not in
probabilistic terms. Unlike some of the approaches noted above, we also saw no reason to attempt to
force a consensus in the judgement process, through iteration or negotiation, because previous studies
have shown that consensus approaches tend to narrow the range of differences among expert views
(Morgan and Henrion, 1990). Hence, we sought an approach that could be implemented in a group
context, allowing for individual expression of views, followed by interpretation over the opinions of
recognized experts on Fraser sockeye salmon and climate change.
3.2
Characterizing scientific knowledge
Three approaches were used in this research to structure scientific knowledge and thus provide the basis
for the judgmental assessment of vulnerability. First, interviews were conducted with scientific specialists,
including two of the authors of this paper (MCH and SGH), to confirm that considering sockeye salmon
vulnerability to climate change throughout their whole life cycle is sensible and relevant as a basis for
structuring vulnerability judgments and that managers and scientists wanted to consider the vulnerability
of specific geographic groups of Fraser sockeye stocks. The second step was the preparation of an
extensive background paper summarizing available scientific information regarding the effects of climate
change on sockeye (much of this literature is summarized above in Section 2.3). The information was
structured in terms of a conceptual model that considers how changes to sockeye due to climate change
at one life stage may propagate through the life cycle to affect other stages or accumulate across
generations. The paper also included a series of influence diagrams (Clemen & Reilly, 2000; Howard,
1989) illustrating possible linkages among variables that could influence the possible effects of climate
change on each life cycle stage (Figure 1 provides one example). The figures are not deterministic, nor
do they imply the linkages necessarily hold. They are representations of beliefs in the form of a model on
paper. A third step was the creation and testing of a survey instrument that was meaningful for the
participants, workable in terms of the tasks required, and relevant for the management tasks at hand.
This survey instrument is described below.
Figure 1.
An example of an Influence Diagram for Returning Adults
3.3
Participants
A one-day workshop was held in December 2006 to obtain judgments regarding sockeye vulnerability.
Participants included academic scientists, federal and provincial government staff members, former
political leaders concerned with fisheries, and non-government organization staff members knowledgable
about Fraser sockeye and climate change. Confirmed attendees were sent the background paper so they
could become familiar with the available literature prior to the workshop. Thirteen experts participated in
the workshop (seven categorized themselves as biologists, four as policy makers, one as “both”, and one
as “neither”).2 The mean years of experience was over 16 years, with the biologists having more
experience (20 years) on average than the policy-makers (7 years).
3.4
Structure and Survey Instrument Design
The workshop began with brief presentations on key topics including an overview of the literature dealing
with environmental trends that are likely to ensue as climates continue to change, to ensure participants
had a common level of understanding regarding the tasks and a common definition of vulnerability.
Vulnerability was defined as “…the extent to which a natural or social system is susceptible to sustaining
damage from climate change. Vulnerability is a function of the sensitivity of a system to changes in
climate and the ability to adapt the system to changes in climate. Under this framework, a highly
vulnerable system would be one that is highly sensitive to modest changes in climate.” (IPCC, 1997
definition cited on The Vulnerability Network http://www.vulnerabilitynet.org/))
For the first half of the workshop, participants used a workbook to rate: 1) the vulnerability of sockeye at
each life cycle stage, 2) the vulnerability of sockeye over their entire life cycle (overall vulnerability); and
3) their knowledge and confidence regarding the previous two exercises. Participants were given the
2
The individual who self-identified as neither a biologist nor manager is a former federal Minister of
Fisheries, and a knowledgeable member of non-governmental organizations regarding fisheries issues.
opportunity to discuss their responses and adjust their answers if their opinions changed as a result of the
discussion.
The workbook questions were designed so that responses could be analyzed for differences between the
lifecycle stages and between the individual stages and overall. Pre-defined scales were used to describe
various levels of vulnerability and the same scale was used for all questions. A sample question is shown
in Figure 2 for one geographic area.
Fry- emergence to lake entry
As described above, the faster development of embryos may lead to early emergence of
smaller fry. This early emergence may result in a mismatched timing with the spring bloom
of zooplankton. This mismatch may limit feeding and growth opportunity for sockeye.
When the nursery lake is upstream, fry must migrate against the current. Smaller fry may be
less able to make such an upstream migration successfully.
As a result, the quantity and/or quality of sockeye at the end of this lifecycle stage
would be impaired.
How vulnerable are fry, from their emergence to lake entry, to this reduction in
quantity and/or quality?
Sockeye that spawn in the upper Fraser system
o
Vulnerability with 2 C average temperature increase
1
2
3
4
5
Not vulnerable
Highly vulnerable
(not affected at all or
(reproductive capacity
status quo)
severely disrupted)
o
Vulnerability with 4 C average temperature increase
1
2
3
4
5
Not vulnerable
Highly vulnerable
(not affected at all or
(reproductive capacity
status quo)
severely disrupted)
Figure 2.
3.5
Sample question for one lifecycle stage and one geographic area
Structuring the Judgments
The questions elicited judgments of vulnerability in relation to lifecycle stage, spawning region, and
climate forecast. The first set of questions asked participants to rank how vulnerable they thought the
sockeye would be at each lifecycle stage under two different climate change scenarios (2 oC and 4 oC
average temperature increases), and for three separate spawning regions (upper, mid and lower Fraser
systems). This created a total of six combinations.. The two scenarios were framed around a relatively
low and relatively high temperature increase reflecting a likely minimum and maximum range which is
forecasted based on the available literature examining this for the Fraser Basin and North East Pacific
Ocean (Morrison et al. 2002, Hinch et al. 1995; Levy 1992, Moore 1991; Henderson et al. 1992; PCIC
2008). Clearly there are other environmental changes associated with these temperature changes (e.g.
rise in river temperature is associated with a decrease in river discharge; rise in ocean temperature is
associated with a decrease in wind strengths and upwelling) and the more extreme the change in
temperature, the more extreme the change in the other associated environmental features. We asked
participants to consider the temperature change as an indicator of the myriad of environmental changes
that may occur simultaneously with temperature changes. Nonetheless, temperature changes are
expected to have some of the most direct and dramatic effects on salmonids. The accuracy of the 2 ºC
and 4ºC scenarios was not the concern of this workshop, which focused instead on how vulnerable
sockeye were to adverse effects given these two reasonably understandable scenarios.
Participants were asked to consider the effects over a 60-year time frame (15 cycles of sockeye salmon),
and to consider the effects summarized in Table 2 as a group. The questions were organized by lifecycle
stage and the issue of vulnerability focused on the ability of the sockeye to move on to the next stage of
their life cycle in good state (i.e. considering both quantity and quality) under each climate scenario.
Participants circled a number between 1 and 5, where 1 indicated no vulnerability (i.e. no different from
the status quo) and 5 indicated a high degree of vulnerability (i.e. reproductive capacity is severely
disrupted). The second exercise asked the participants to rank the overall vulnerability of sockeye
salmon (on the same 1-5 vulnerability scale) to 2 or 4 ºC climate warming in each of the three regions
over a 60-year time frame (15 cycles of sockeye salmon). Answers to this question were later compared
to the answers for individual lifecycle stages. Finally, participants were asked to rate their knowledge of
and confidence in their responses to each question. [Note, join the above 4 paragraphs together]
3.6
Eliciting and Revising Judgments
All judgments for a given life cycle stage were elicited at one time over both temperature scenarios and
the three spawning regions. Before each life cycle stage was considered, we reviewed the structure of the
influence diagram and content of the background paper relevant for that question. After all participants
had completed their judgments for a given set of questions, we encouraged participants to discuss their
answers. For some questions, lively discussion ensued, and for others, little was said. Participants were
permitted to revise their judgments at any time. A review of the workbooks showed that 11 of the 13
participants had revised at least one answer; some had made many revisions. We have no way of
knowing what led to the changes, but they do indicate the participants took the tasks seriously.
3.7
Statistical Approach
After the workshop was completed, responses to the questions were compiled into a database. We used
the non-parametric Mann-Whitney U rank sum test to compare the distribution of responses for pairs of
questions or question sets assuming a null hypothesis of no difference in the ranks. (n=13)3 . We
conducted three sets of Mann Whitney tests.
To investigate differences between lifecycle stages when other variables were held constant, we
compared vulnerability estimates for three lifecycle stages (eggs, fry-first winter and returning adults)
within all six temperature-region combinations.4 For example, we compared the vulnerability of the eggs
life cycle stage in the upper Fraser region under a 2ºC temperature increase scenario with the
vulnerability of the fry first winter life stage and the returning adults life cycle stage in the same region and
at the same temperature increase. This analysis involved a total of 18 Mann Whitney tests (three for each
of the six temperature-region combinations). [Need to say briefly why these 3 life stages were selected for
comparison.] To investigate differences between temperature scenarios and regions, we compared
overall vulnerability estimates for the six combinations of spawning region and temperature increase. The
3
A total of 63 data points out of the potential total of 936 were missing because some participants did not answer certain questions,
Of these, 46 were left blank by one individual, who self-identified as a policy maker and not a fisheries expert. We handled these
missing responses by replacing them with the mode response for that question, in keeping with the ordinal, categorical data
structure.
4 We selected these three life stages for comparison because, a priori, they seemed to have the most potential to be affected by
climate change.
overall vulnerability estimates for a particular temperature scenario and region were compared to the
overall vulnerability estimates for the same region but at the other temperature. The overall vulnerability
for a particular temperature and region was also compared to the same temperature scenario but in
different regions. These tests investigated the impact of changing the location and temperature scenario
on overall vulnerability.
The third set of Mann-Whitney tests investigated differences in knowledge and confidence for the three
aforementioned life cycle stages (eggs, fry first winter and returning adults).
4. Results of Vulnerability Assessment
4.1
Histograms of the responses for each question
Figure 2 shows the histograms of responses for all questions, with the missing data imputed as noted in
footnote 4 . The histograms are all unimodal. The variation in responses ranged from questions for which
all responses were two adjacent categories (indicating high agreement) to questions for which responses
were spread over five categories (indicating a strong lack of agreement). For two of the judgment sets for
returning adults spawning in the Middle Fraser region, one participant judged vulnerability much lower
than the other participants (Fig 2, centre panels for returning adults). The judgments showing high
agreement (varying over only two categories) and high vulnerability of stocks were (i) eggs in the midFraser under the 4oC scenario, and (ii) Upper Fraser stocks over the whole life cycle under the 4 oC
scenario.
Figure 2. Histograms of responses for each question
Visual inspection of the histograms does not indicate that participants were any more confident of the
impact of a 4oC increase than of a 2oC increase, nor do they appear to be more confident about
vulnerability over the entire life cycle than for any specific stage.
4.2 Life cycle vulnerability ratings
Participants rated sockeye as most vulnerable to climate change during egg and returning adult life
stages. Eggs had a mean score of 3.67 and returning adults had a mean score of 3.82 (for all regions
and temperature scenarios). Mid-life cycle stages such as fry-first winter had, in contrast, lower
vulnerability means (range 2.58 to 3.13). The means for each life cycle stage are shown in Table 3,
broken out by location and temperature scenario, and shown for each life cycle stage overall.
The difference in vulnerability between the eggs and fry-first winter was statistically significant at p < 0.01
for all six combinations of spawning region and temperature increase. None of the differences in
vulnerability between egg and other life stages were statistically significant.
Likewise, the difference in vulnerability between the returning adults stage and fry-first winter was
statistically significant at p < 0.01 for all six region-temperature combinations. The difference in
vulnerability between returning adults and fry-smolt was significant for upper Fraser at 2oC and 4oC, and
lower Fraser at 2oC. None of the other differences in vulnerability were statistically significant.
Table 3.
Mean vulnerability (on a scale of 1-5) of upper, mid and lower sockeye salmon stocks at
different temperatures and by life cycle stage
Mean Vulnerability scores for each region and
temperature increase
Means for all
Upper -
Upper -
Mid -
Mid -
Lower -
Lower -
regions and
Life cycle stage
2ºC
4ºC
2ºC
4ºC
2ºC
4ºC
temperatures
Eggs
3.17
4.21
3.17
4.13
3.25
4.13
3.67
Fry - emergence to lake entry
2.42
3.54
2.67
3.79
2.67
3.71
3.13
Fry - lake entry to first winter
2.42
3.75
2.67
3.75
2.42
3.67
3.11
Fry - first winter
2.36
3.18
2.27
3.18
1.91
2.55
2.58
Fry - smolt
2.45
3.27
2.55
3.36
2.36
3.09
2.85
Post smolts
2.46
3.46
2.63
3.63
3.04
3.79
3.17
Immature sockeye in ocean
3.04
3.96
2.96
3.88
2.95
3.95
3.46
Returning adults
3.69
4.58
3.35
4.27
3.12
3.92
3.82
Means for all lifecycle stages
2.75
3.74
2.78
3.75
2.71
3.60
The mean vulnerability scores of the most vulnerable life cycle stages (eggs and returning adults) were
not statistically different from the mean score for the overall life cycle.
4.3
Overall vulnerability ratings
Table 4 shows the means from the questions related to overall vulnerability (i.e. not broken down by
lifecycle stage). Sockeye were viewed as substantially more vulnerable to climate change under a 4ºC
temperature increase scenario than a 2ºC increase scenario. The vulnerability of sockeye to climate
change increased from the lower to the upper Fraser but the differences in vulnerability between regions
was statistically significant only for upper vs lower Fraser at 4 C. The difference between the 2ºC 4ºC
temperature scenarios within each region was statistically significant (at p < 0.01) for all regions.
Table 4 Mean overall vulnerability (on a scale of 1-5) of upper, mid, and lower sockeye salmon stocks
at different temperatures
Temperature
Upper Fraser
Mid Fraser
Lower Fraser
Spawning Grounds
Spawning Grounds
Spawning Grounds
2ºC
4ºC
2ºC
4ºC
2ºC
4ºC
3.58
4.73
3.42
4.50
3.27
4.19
Overall Vulnerability
mean
4.4
Uncertainty
Participants rated their overall knowledge of the workshop topics and their overall confidence regarding
their workbook responses on a scale of 1 to 5, with 5 being most knowledgeable or confident and 1 being
least. The mean knowledge ranking was 3.13 and the mean confidence ranking was 2.42, suggesting the
participants did not have high confidence in either their knowledge or their judgments. This is in spite of
the fact that the participants were among the best informed technical and management specialists in the
world regarding these fish resources. Participants were roughly equally uncertain about their judgements
regarding each life stage except for returning adults, where they felt more knowledgeable and more
confident. Differences between the knowledge ratings for the returning adult stage and every other stage
were statistically significant at p < .01. Likewise, the differences in self-reported confidence between the
returning adult stage and other stages were statistically significant (at p < 0.05 for immature sockeye in
ocean and eggs, and p < .01 for the other lifecycle stages). Note that participants’ self-assessment of
their overall knowledge was higher than the mean of reported knowledge over all lifecycle stages.
Table 5 shows participants’ knowledge and confidence ratings for each life cycle stage.
Table 5.
Mean ratings of own knowledge and confidence regarding sockeye vulnerability issues,
by life cycle stage and overall
Mean
Mean
Knowledge
Confidence
Rating
Rating
Eggs
2.54
2.54
Fry - emergence to lake entry
2.54
2.42
Fry - lake entry to first winter
2.62
2.42
Fry - first winter
2.46
2.17
Fry - smolt
2.54
2.42
Post smolts
2.46
2.29
Immature sockeye in ocean
2.69
2.54
Returning adults
3.62
3.23
Means for all lifecycle stages
2.68
2.50
As reported overall
3.13
2.42
Life Cycle Stage
5.
Findings – Mitigation
The second major task was to elicit the participants' views regarding the potential to mitigate any adverse
effects of climate change on sockeye salmon reproductive capacity. We addressed two broad questions
in a facilitated discussion: (i) what options are available to mitigate the effects of climate change on
Fraser sockeye stocks, and (ii) how practical are these options?
The discussion began with the adult stage of the salmon life cycle, as participants explained that reaching
spawning grounds was crucial to every other stage. Each lifecycle stage was then examined in turn,
moving in reverse chronological order from returning adults back to eggs. For each life stage, participants
were asked to think of and discuss two kinds of options: (i) mitigation measures that involve changes in
policy, or how fisheries are managed, and (ii) mitigation measures that involve building infrastructures of
some kind.
Table 6 summarizes the results of the discussion. In general terms, the kinds of policy measures that
could be pursued for returning adults are similar to those now being considered or implemented by
fisheries managers in the Department of Fisheries and Oceans current wild salmon initiatives. Policy
measures for other life cycle stages are unclear, except for perhaps predator management for the stage
from lake entry to smolts. The only infrastructure measures that were viewed as practical included the
construction of additional spawning channels to aid returning adults and eggs. Investments in
temperature management and flow augmentation could be considered at this level in systems without
lake influence. Importantly, the prospect of temperature management in the main stem of the Fraser
through construction of low level outlets on the existing Kenny Dam, or construction of dams on other
tributaries for purposes of temperature management were all viewed as impractical, given the volumes of
water needed to affect temperature in the main stem of the Fraser system.
In sum, there are relatively few alternatives available to mitigate the effects of climate change on sockeye
salmon in the Fraser system. These fisheries are already tightly managed and nearly fully committed to
various user groups. There are few or no new structural alternatives, and little in terms of new
management directions that could help mitigate the vulnerability of Fraser River Sockeye salmon to
climate change,
On the other hand, Fraser River sockeye already demonstrate a great deal of adaptive capacity in
utilizing heterogeneous habitats in different river basins. Hence, this adaptability points to the potential
value of policies that seek to make stocks more resilient to uncertain futures. Some authors have
discussed the potential for future adaptation of salmon to climate changes (Crozier, et al, 2008), Healey
(2009) considers resilience attributes of salmon and measures to enhance them.
Table 6 Mitigation prospects for Fraser sockeye given climate change
Life cycle stage
Returning adults
Policy Measures to assist with
Infrastructure to assist with
mitigation
mitigation

Adopt a balanced “fish friendly”

approach that fostered adaptive
management and policy

Create additional spawning
channels, \

Cold-water release (to lower
coordination across species.
the temperature of the Fraser
Alternative fishing strategies to
main stem or spawning
achieve escapement goals
streams) is not a viable option,
(thereby conserving genetic
given the volumes of water that
diversity) including: quota
would be required.
management, limited harvest
and removal rates, more
efficient fishing strategies, and
an increased number of smaller
fisheries.
Immatures in

ocean
Not clear. At this stage the fish

None suggested

Measures to aid in recovery of
are widely dispersed and not
necessarily within Canadian
jurisdiction.
Post smolts,

None suggested
estuary, and
estuary ecosystems
coastal
Fry: Lake Entry to

Smolts
Predator management to stop

None suggested

Constructing spawning
the spread and distribution of
predators (e.g., preventing
fishermen from moving bass
from lake to lake or preventing
the natural migration of
predators).

Participants argued for a multispecies monitoring program
that could help identify
predators early on, before they
became established in a new
ecosystem.
Eggs

None suggested
channels and off-channel
spawning places can best
assist eggs. Investments in
temperature and flow
management should be made
in systems without lake
influence.
6.
Discussion
We first provide discussion of the substantive results, followed by discussion of the elicitation process and
its potential for further development and application.
Participants clearly thought that sockeye salmon in the Fraser River system are vulnerable to climate
change to the extent that reproductive capacity could be impaired. This was particularly true for climate
scenarios that predict greater temperature increase and for stocks spawning in the upper Fraser system.
Participants believed that returning adults and eggs were the life cycle stages most vulnerable to climate
change. The returning adults life cycle stage was rated as the most vulnerable, and was also the life
stage for which the participants had the highest knowledge and confidence ratings. Overall, participants
expressed the following points in clarifying their views on sockeye vulnerability:

There are many uncertainties regarding vulnerability estimates.

While it is easy to predict that there will be an impact on sockeye due to climate change,
predicting the specific impact is more difficult.

There are many factors that can affect sockeye vulnerability, but one factor alone may be
decisive. (For example, predator impact on smolts in coastal areas may be so important as
to render all other factors irrelevant)

It is important to address specific regions, stocks and runs when estimating vulnerability
because each subspecies/run is different and because secondary impacts are different in
each region.

Because the exact impacts due to climate change are not certain, there is a need to focus on
robust mitigation policies and actions that preserve/conserve that ability of fish to adapt to
climate change (i.e. policies that conserve genetic and habitat diversity).

Sockeye in the Fraser system are clearly highly adapted to specialized habitat conditions,
which may differ greatly from one river system to the next. Hence it s possible that the
adaptive capacity they have demonstrated in the past may lead to spontaneous adaptation of
warmer conditions through behaviour changes. For example, delayed run timing for Upper
stocks have been recorded in recent years.
Mitigation of the impacts of climate change on sockeye salmon may be difficult because the resource
management systems are already highly constrained, with overlapping demands on the resource from
multiple gear types and sequential fisheries. Participants believed that efforts should be concentrated on
fostering resiliency to climate change impacts instead of predicting specific levels of change. To facilitate
adaptation, participants suggested improved information collection and strategic science-based learning
to improve understanding of current system functions. This new information could help experts develop
robust policies that are better able to cope with the irreducible uncertainties of climate change.
Turning to the process developed and applied here, this is to our knowledge the first use of a structured
process to elicit expert judgments of vulnerability to climate change for a specific species, in a specific
habitat system, over its life cycle, differentiating among sub-groups. Ecological vulnerability to climate
change has largely been addressed over large geographic areas (e.g., such as the Pacific northwest
states, as in Miles, et al, 2007) or over many species (e.g., Walther et al 2002; Parmesan and Yohe,
2003). These assessments have largely considered impacts of climate change to the present, without
considering vulnerability of specific species to future change in particular locations. Here we used
approaches informed by methods of probability elicitation (Morgan and Henrion, 1990), but within a more
direct and less complex judgment task, to obtain judgments of vulnerability that integrate over many
variables and scenarios. We employed the individual judgements of each participant, rather than
attempting to create a consensus within the elicitation process, because of the importance of
understanding the range of views (Morgan and Henrion, 1990). Based on the experience in this research,
the method appears promising for similar judgment tasks regarding vulnerability and climate change. The
method could benefit from further applications and critical examination, as well experimental testing of the
kind that is sometimes employed for probability assessment. Any experimental tests would however face
methodological hurdles similar to that arising for probability elicitation, such as the difficulty of judging
calibration without rapid feedback on the results (Clemen and Winkler, 1999).
Acknowledgement
This research was made possible by Cooperative Agreement SES – 0345798 between the US National
Science Foundation and the Climate Decision making Centre at Carnegie Mellon University, and a subgrant to the University of British Columbia. The contributions by M. Healey were supported by an NSERC
Discovery grant.
We greatly thank the participants who provided their time and valuable perspectives for this project.
References
Burgner, R. L. (1991). Life history of sockeye salmon (oncorhynchus nerka). In C. Groot, & L. Margolis
(Eds.), Pacific salmon life histories (pp. 1-117) UBC Press, Vancouver.
Clemen, R. T., Reilly, T. 2000, Making Hard Decisions with Decision Tools, Duxbury Press; Belmont
California
Cemen, R. T. and R. Winkler. (1999) Combining Probabilities From Experts In Risk Analysis. Risk
Analysis. 19, 2. 187-203.
Cooke, S. J., Hinch, S. G., Crossin, G. T., Patterson, D. A., English, K. K., Shrimpton, J. M., et al. (2006).
Physiology of individual late-run Fraser River sockeye salmon(oncorhynchus nerka) sampled in the
ocean correlates with fate during spawning migration. Canadian Journal of Fisheries and Aquatic
Sciences, 63(7), 1469-1480.
Cooke, S. J., Hinch, S. G., Farrell, A. P., Lapointe, M. F., Jones, S. R. M., Macdonald, J. S., et al. (2004).
Abnormal migration timing and high en route mortality of sockeye salmon in the Fraser River, British
Columbia. Fisheries, 29(2), 22-33.
Crossin, G. T., Hinch, S. G., Cooke, S. J., Welch, D. W., Patterson, D. A., Jones, S. R. M., et al. (2008).
Exposure to high temperature influences the behaviour, physiology, and survival of sockeye salmon
during spawning migration. Canadian Journal of Zoology, 86(2), 127-140.
Crozier, L. G., A. P. Hendry, P. W. Lawson, T. P. Quinn, N. J. Mantua, J. Battin, R. G. Shaw and R. B.
Huey (2008). “Potential responses to climate change in organisms with
complex life histories: evolution and plasticity in Pacific salmon” Ecological Applications. 252-270
Farrell, A.P., Hinch, S.G., Cooke, S.J., Patterson, D.A., Crossin, G.T., Lapointe, M., and Mathes, M.T.
(2008) Pacific salmon in hot water: applying metabolic scope models and biotemetry to predict the
success of spawning migrations. Physiological and Biochemical Zoology 82: in press.
Finney, B. P., Gregory-Eaves, I., Sweetman, J., Douglas, M. S. V., & Smol, J. P. (2000). Impacts of
climatic change and fishing on Pacific salmon abundance over the past 300 years. Science,
290(5492), 795-799.
Healey, M. C. 2009. Resilient Salmon, Resilient Fisheries for British Columbia, Canada.
Ecology and Society 14 (1): 2. [online] URL:http://www.ecologyandsociety.org/vol14/iss1/art2/
Helfield, J. M., & Naiman, R. J. (2001). Effects of salmon-derived nitrogen on riparian forest growth and
implications for stream productivity. Ecology, 82(9), 2403-2409.
Henderson, M. A., Levy, D. A., & Stockner, J. S. (1992). Probable consequences of climate change on
freshwater production of Adams river sockeye salmon (oncorynchus nerka). GeoJournal, 28(1), 5159.
Hinch, S. G., Cooke, S. J., Healey, M. C., & Farrell, A. P. (2005). Behavioural physiology of fish
migrations: Salmon as a model approach. In K. Sloman, S. Balshine & R. Wilson (Eds.), Fish
physiology: Behaviour and physiology of fish (pp. 239-295) Elsevier Press.
Hinch, S. G., Healey, M. C., Diewart, R. E., & Henderson, M. A. (1995a). Climate change and ocean
energetics of Fraser River sockeye (oncorhynchus nerka). Climate Change and Northern Fish
Populations, 439-445.
Hinch, S. G., Healey, M. C., Diewert, R. E., Thomson, K. A., Hourston, R., Henderson, M. A., et al.
(1995b). Potential effects of climate change on marine growth and survival of Fraser River sockeye
salmon. Canadian Journal of Fisheries and Aquatic Sciences, 52(12), 2651-2659.
Hodgson, S., & Quinn, T. P. (2002). The timing of adult sockeye salmon migration into fresh water:
Adaptations by populations to prevailing thermal regimes. Canadian Journal of Zoology, 80(3), 542555.
Howard, R. A. (1989). Knowledge maps. Management Science, 35(8), 903-922.
Johnston, N. T., MacIsaac, E. A., Tschaplinski, P. J., & Hall, K. J. (2004). Effects of the abundance of
spawning sockeye salmon(oncorhynchus nerka) on nutrients and algal biomass in forested streams.
Canadian Journal of Fisheries and Aquatic Sciences, 61(3), 384-403.
Keeney, R. L., & von Winterfeldt, D. (1986). Improving risk communication. Risk Analysis, 6(4), 417-424.
Keeney, R. L., & von Winterfeldt, D. (1989). On the uses of expert judgment on complex technical
problems. Engineering Management, IEEE Transactions on, 36(2), 83-86.
Klyashtorin, L. B. (1998). Long-term climate change and main commercial fish production in the Atlantic
and Pacific. Fisheries Research, 37(1-3), 115-125.
Lapointe, M, (2008), personal communications, December, 2008.
Levin, P. S. (2003). Regional differences in responses of chinook salmon populations to large-scale
climatic patterns. Journal of Biogeography, 30(5), 711-717.
Levy, D.A. 1992. Potential impacts of global warming on salmon production in the Fraser River watershed. Can. Tech.
Rep. Fish. Aquat. Sci. No. 1889.
Linstone, H. A., and M. Turoff, (1975). Delphi Method: Techniques and applications. Addison-Wesley: Massachusetts,
640 p.
Macdonald, J. S., Foreman, M. G. G., Farrell, T., Williams, I. V., Grout, J., Cass, A., et al. (2000). The
influence of extreme water temperatures on migrating Fraser River sockeye salmon (oncorhynchus
nerka) during the 1998 spawning season. Canadian Technical Report of Fisheries and Aquatic
sciences/Rapport Technique Canadien Des Sciences Halieutiques Et Aquatiques.Imprint Varies,
(2326), 131.
Mantua, N. J., Hare, S. R., Zhang, Y., Wallace, J. M., & Francis, R. C. (1997). A Pacific interdecadal
climate oscillation with impacts on salmon production. Bulletin of the American Meteorological
Society, 78(6), 1069-1079.
Miles, E. L., Snover, A. K., Hamlet, A. F., Callahan, B., and Fluharty, D. (2007) Pacific Northwest
Regional Assessment: The Impacts of Climate Change and Climate Variability on Water Resources
of the Columbia River Basin. Journal of the American Water Resources Association. 36.2. 399-420.
Moore, R.D. 1991. Hydrology and water supply in the Fraser River basin. In Water in sustainable development: exploring
our common future in the Fraser River basin. Edited by A.H.J. Dorcey and J.W. Griggs. Westwater Research Centre, The
University of British Columbia, Vancouver, B.C. pp. 21-40.
Neitzel, D.A.,
Morgan, M. G., & Henrion, M. (1990). Uncertainty: A guide to dealing with uncertainty in quantitative risk
and policy analysis Cambridge University Press.
Morrison, J., Quick, M. C., & Foreman, M. G. G. (2002). Climate change in the Fraser River watershed:
Flow and temperature projections. Journal of Hydrology, 263(1-4), 230-244.
Northcote, T. G., & Larkin, P. A. (1989). The fraser river: A major salmonine production system.
proceedings of the international large river symposium. Canadian Special Publication of Fisheries
and Aquatic Sciences,
Pacific Climate Impacts Consortium (PCIC), 2008, PCIC regional analysis tool, accessed August, 2008 at
http://pacificclimate.org/tools/select
Patterson, D. A., Macdonald, J. S., Skibo, K. M., Barnes, D., Guthrie, I., & Hills, J. (2007). Reconstructing
the summer thermal history for the lower Fraser River 1941 to 2006, and implications for adult
sockeye salmon (oncorhynchus nerka) spawning migration. Canadian Technical Report of Fisheries
and Aquatic Sciences, In press
Pacific Salmon Commission 2009a. Report of the Fraser River Panel to the Pacific Salmon Commission
on the 2005 Fraser River Sockeye and Pink Salmon Fishing Season. Vancouver, B.C. XX p.
Pacific Salmon Commission 2009b. Report of the Fraser River Panel to the Pacific Salmon Commission
on the 2006 Fraser River Sockeye Fishing Season. Vancouver, B.C. XX p.
Pacific Salmon Commission 2009c. Report of the Fraser River Panel to the Pacific Salmon Commission
on the 2007 Fraser River Sockeye and Pink Salmon Fishing Season. Vancouver, B.C. XX p.
Rand, P. S., Hinch, S. G., Morrison, J., Foreman, M. G. G., MacNutt, M. J., Macdonald, J. S., et al.
(2006). Effects of river discharge, temperature, and future climates on energetics and mortality of
adult migrating Fraser River sockeye salmon. Transactions of the American Fisheries Society,
135(3), 655-667.
Ricker, W. E. 1997. Cycles of abundance among Fraser River sockeye salmon (Oncorhynchus nerka).
Canadian Journal of Fisheries and Aquatic Sciences. 54:950-968
Swansburg, E., Chaput, G., Moore, D., Caissie, D., & El-Jabi, N. (2002). Size variability of juvenile Atlantic
salmon: Links to environmental conditions. Journal of Fish Biology, 61(3), 661-683.
Walther, G-R., Post, E., Convey, P. Menzel, A. Parmesan, C. Beebee, T. J. C., Fromentin, J-M.
Hoegh-Guldberg, O. and Bairlein, F. 2002. Ecological Responses to Recent Climate
Change. Nature. 416. 389-395.
Pamesan, C. and Yohe, G. 2003. A globally coherent fingerprint of climate change impacts
across natural systems. Nature. 421. 37-42.
Williams, A. 2007. The Pacific Salmon Treaty: A Historical Analysis and Prescription for the
Future. Journal of Environmental Law and Litigation. 22. 153-195
Williams, B. 2005. 2004 southern salmon fishery post-season review. Part one, Fraser River
sockeye report. Fisheries and Oceans Canada, 91p.
Zickfeld, K., Levermann, A., Morgan, M. G., Kuhlbrodt, T., Rahmstorf, S., & Keith, D. W. (2007). Expert
judgements on the response of the Atlantic meridional overturning circulation to climate change.
Climatic Change, 82(3), 235-265.