Kim Hyatt, Margot Stockwell & Paul Rankin Fisheries & Oceans Canada, Salmon in Regional Ecosystems Program, Pa

1. Impact of Climate Variation and Change Events on Threatened or Vulnerable Salmonids in British Columbia: Implications for Conservation and Restoration Planning Kim Hyatt, Margot Stockwell & Paul Rankin Fisheries & Oceans Canada, Salmon in Regional Ecosystems Program, Pacific Biological Station, Nanaimo, B.C. Feb 3, 2005: CIG Work Seattle

2. Sockeye Salmon Index Stocks on the Eastern Rim of the Pacific • approximately 30% of all BC sockeye populations originate from watersheds near the southern end of the species range in North America • these populations are likely to be highly sensitive to climate change

3. Subject Stocks and Study Area

4. Trends in Adult Sockeye Returns to the Columbia River(1890 - 1999) • Okanagan sockeye are the last anadromous salmon stock of dozens that formerly returned to Canada through the Columbia River. • The stock is currently depressed and fluctuates with a time weighted average to decline.

5. S T R E S S Study Theme Climate change will challenge the scope for life history adaptation by southern sockeye populations to variations in seasonal thermal regimes and associated changes to aquatic ecosystems. • many elements of behaviour, physiology, & ecology of sockeye are controlled by temperature • physiological performance is optimal at 15 C • sockeye are increasingly stressed between 17 and 24 C • 25 C is lethal Results presented here first for Okanagan sockeye are part of a general approach to developing: (1) time series estimates of annual and seasonal changes in aquatic thermal regimes, (2) analysis of sockeye population responses to historic variations in climate, (3) inferences about global warming impacts on sockeye populations and (4) identification of potential options for effective stock management, conservation and/or restoration.

6. Changes to Aquatic Thermal Regimes in Southern Watersheds Pose Threats to Salmon at Several Life History Stages Analysis of sockeye responses to seasonal & annual variations in thermal regimes are now complete for life history stages involving: (1) adult migration (2) adult spawn timing (3) egg-fry incubation (4) juvenile rearing General procedure developed on Okanagan sockeye is to: (1) identify a biophysical model or “set of rules” for a given life history response and then (2) predict consequences of climate change

7. true water temp spawners Derivation of Temperature “Rules” for Adult Migration(Hyatt et al. 2003. Can. Wat. Res. J. 28: 689-713) • Sockeye migration through Wells Dam on the Columbia mainstem is generally “unimodal” ( ) • Sockeye migration through Zosel Dam on the Okanagan River is often “polymodal” ( ) • Disruptions & delays to entry of Okanagan coincide with seasonal temperature variations. ( ) • Adult migrations stop at 21 C when temperature is increasing • Migration resumes at 21 - 22 C when seasonal temperatures decrease

8. Predicted versus observed delays for Okanagan sockeye adults based on temperature stop & start migration model

9. Temperature-based Stop & Start “Rule” Predictions of Adult Sockeye Migration Delay Adult sockeye migration delay based on seasonal changes in water temperature of the Okanagan River • Applying temperature-migration rules suggest delays may range from 0 - 54 days & an average of 30 days per annum. The magnitude of delays alternate with PDO cycles & have been increasing steadily in association with climate warming during 1985 - 2000 compared with lesser delays in the “cool” 1947-1985 interval.

10. water temperature (interpolated) sockeye spawning water temperature (actual) water temperature 12ºC Derivation of Temperature “Rules” for Adult Spawn Timing • peak spawn timing occurs at 12 °C on average (range 8 - 15 °C)

11. Okanagan Adult Sockeye Average Peak Spawn Date • peak spawning by adult sockeye has been delayed, on average, by 9 days during recent “warm” interval

12. ATU Model “Rules” Used to Estimate Hatch & Emergence Dates for Okanagan Sockeye Eggs and Fry Eggs Hatching SPAWNING AREA OKANAGAN RIVER Fry Recruitment OSOYOOS LAKE On average, 100% hatch date is delayed by 15 days during warm intervals (i.e. 1985-2001) relative to cooler intervals (i.e. 1947-1985).

13. Derivation of Temperature “Rules” for Habitat Use by Juvenile Sockeye in Osoyoos Lake • Juvenile sockeye rear in Osoyoos Lake for 1 year. • Acoustic and trawl surveys are used to follow seasonal changes in habitat use by juvenile sockeye rearing in Osoyoos Lake. November June Winter – juvenile sockeye use most of the water column Summer – sockeye concentrate near the thermocline

14. Day versus Night Distribution Changes Reveal “Rules” for Seasonal Habitat Utilization by Juvenile Sockeye in Lakes • “17 degree rule”: night ascent limited by 17 C water • “4 mg rule”: daytime descent limited by 4 mg/l O2 level

15. North Central South Seasonal Restriction of Vertical Distribution Temperature and Oxygen extremes operate together to restrict the useable rearing volume of Osoyoos Lake. South - unsuitable for rearing through summer & fall. Sockeye avoid this basin or die. Central - unsuitable for summer rearing although some sockeye may survive. North - suitable for rearing but still poses a challenge for sockeye in late summer and fall in some years.

16. Seasonal & Annual Variations in Physical Conditions May Cause Mortality Events 5.9% loss per week = 95 % loss over 4 months

17. warm cool warm warm cool warm x = 100 days x = 82 days x = 94 days x = 68 days x = 78 days x = 85 days all year mean = 77 days all year mean = 91 days 17 ºC “Rule” Predictions of Number of Days Juvenile Sockeye are Excluded from Surface Waters of Osoyoos (B.C. Interior) and Great Central (Vancouver Island) Lakes (1924 - 1998) (a) Osoyoos (a) Great Central • temperature induced exclusions of juvenile sockeye from surface waters of Osoyoos Lake range from 40 - 120 days per year and average 78 days per year. • the temperature-oxygen “squeeze” has likely been more severe during the current “warm” period (1985 - 2000) compared to the earlier “cool” period (1945 - 1985). • these effects are especially clear in Osoyoos Lake but are similar in distant locations suggesting a regional climate “driver” associated with PDO cycles.

18. Climate Change Impacts on Sockeye • To date, our studies on Okanagan sockeye suggest: • high temperatures delay adult migration by 30 days on average (range 0-54) and average delays double in “warm” (1985-2002) vs “cool” (1947-1985) intervals. (Migration delays induce more pre-spawn mortality & reduce gamete viability). • spawn timing during “warm” vs “cool” intervals is delayed by an average of 9 days while 100% egg hatch may be delayed by 15 days on average. (Hatching delays expose eggs to additional risk of loss from redd scour events during spring “floods”). • summer temperature and O2 extremes control the quantity and quality of juvenile rearing habitat in Osoyoos Lake. (Differences in days of summer exclusion from surface waters in warm and cool intervals will influence feeding energetics and production levels of juvenile sockeye). • the temp-O2 “squeezes” may be severe enough to induce major mortality events (e.g. 1998 event). (Rearing habitats have been reduced to zero in all basins of Osoyoos Lake at least twice in the past 30 years).

19. Implications for “Threatened” Sakinaw and Cultus Lake Sockeye Salmon Sakinaw Cultus Small populations of sockeye originating from Sakinaw and Cultus lakes have exhibited more than a 90% decline within a few generations prompting an “emergency recovery” designation by COSEWIC and a petition to list them under SARA legislation in Canada.

20. cool warm cool warm warm cool warm % ND = 12.0 % ND = 0.0 % ND = 9.7 % ND = 0.0 % ND = 45.5 % ND = 29.0 % ND = 7.7 x = 19 x = 29 x = 23 x = 34 x = 5 x = 9 x = 16 Mean = 10 days Mean = 26 days (b) “Exclusion” of Sakinaw fry (d) “Exclusion” of Cultus fry warm cool warm cool warm cool warm x = 118 x = 119 x = 118 x = 93 x = 118 x = 109 x = 127 Mean = 122 days Mean = 110 days Climate Change Events Will Affect Recovery Prospects for Depressed or Threatened Sockeye Stocks Such as Sakinaw and Cultus (b) “Delay” of Cultus adults (a) “Delay” of Sakinaw adults Climate variation & change events are “non-stationary” & have a predictable influence on salmon life history events during specific periods (e.g. “warm” versus “cold” PDO intervals that each last 20-30 years) .

21. CVC Effects on Resident Salmonids in Georgia Basin Lakes

22. CVC Effects on Resident Salmonids in Marion Lake, B.C. N groundwater spring 200 m Rainbow Trout Kokanee Marion L. is a small (29 ha), shallow (6m maximum depth), sub-alpine lake in the UBC Research Forest on the north shore of the Fraser River

23. Rainbow Trout Surface Feeding Activity and Temperature Trout feeding activity at the lake surface falls off rapidly at temperatures <60C or >180C.

24. Salmonid Aggregations in a Spring-fed Thermal Refuge Kokanee and large trout routinely aggregate in cool water refuges when lake surface temperatures exceed 210C

25. Temperatures at Depth in Marion Lake and the Spring When seasonal temperatures exceed 21oC at the surface and 17oC at 4m in the lake they remain below 110C at depths > 2m in the spring

26. Climate Change Impacts on Frequency and Duration of Thermal Refuging by Marion Lake Salmonids warm cool cool warm adr = 7.8 days adr = 10.1 days adr = 7 days adr = 17 days x = 10.7 nr = 8 yr nr = 4 yr nr = 7 yr nr = 2 yr nr = no refuge (# days above 21°C = 0)

27. Interim Conclusions • Our impact assessment model of sockeye life history responses to climate variations developed for Okanagan sockeye suggests southern stocks will be highly sensitive to future climate change. • Our life history “models” developed on Okanagan sockeye may be usefully extended to quantify impacts on sockeye populations throughout the Pacific region (e.g. stocks in Sakinaw, Cultus and Marion lakes) . • Climate variation & change effects on salmon populations are “non-stationary” (e.g. PDO cycles & longer term climate warming), partially predictable & should be assessed for their importance for effective salmon conservation & restoration planning. • The latter will challenge institutional capacity and the existing socioeconomic order (e.g. Columbia R. salmon conservation & restoration experience).

28. CVC Events & Barkley Sd. Sockeye Production

29. Barkley Sound Sockeye Return Variations 2000 1800 1600 1400 1200 1000 800 600 400 200 0 Sockeye returns decline 2 yrs after strong ENSO events Sockeye (thousands) 1900 1920 1940 1960 1980 2000 • returns exhibit two multi-decadal, production “regimes” (i.e. low & high average production) • lots of within decade variation to complicate harvest management

30. 120 Above Average 100 80 60 40 20 % of Mean 0 10 20 30 40 50 60 70 80 - 20 - 40 - 60 - 80 - 120 Below Average Long term (1903 – 1986) trends in sockeye returns exhibit stanzas of below & above average production that are not related in any meaningful way to parent stock & recruitment

31. Trends in marine temperature (Year N) & Barkley Sd sockeye returns (Year N+2) suggest production is depressed or enhanced by warm & cool coastal ocean conditions respectively reflecting both ENSO & PDO cycles SND’s 1935 1960 1984 YEAR Warm PDO Warm PDO Cool PDO Year N Year N + 2

32. Partitioning of freshwater & marine survival within & among sockeye index stocks confirmed declines in early 80’s were localized to SWVI stocks and had marine, not freshwater, origins.

33. Barkley Sound sockeye are especially sensitive because of theirtransitional zone location over which the subarctic domain has a varying influence (Fulton and LeBrasseur 1985) Strong El Niño- La Nina events induce major changes in continental shelf ecosystems. Some of these changes alter ecosystem capacity to support “robust” salmon populations.

34. (Brama japonica) (Mola mola) (Scomber japonicus) (Sarda chiliensis) (Pelicanus occidentalis) Exotic Species Sightings off the BC Coast During 1983 (J. Fulton, P.B.S.)

35. Predator Abundance and Climate VariabilityIn the WCVI Study Area (Ware and McFarlane, 1995 ) Predator Abundance and Climate VariabilityIn the WCVI Study Area (Ware and McFarlane, 1995 ) Fig. 9. Relationship between average June-July sea surface temperature and swept-volume hake biomass for 13 survey years between 1968-91 (p = 0.003; Appendix 1). Fig. 9. Relationship between average June-July sea surface temperature and swept-volume hake biomass for 13 survey years between 1968-91 (p = 0.003; Appendix 1).

36. High northward transport Two-State Model of El Nino – La Nina Mediated Production Responses of Barkley Sd. Sockeye (Hyatt et al. 1989) Low northward transport Zooplankton: low biomass, small size Dominant piscivores coevolved with sockeye Zooplankton: high biomass and large size Dominant piscivores “foreign” to sockeye Average to low mortality for juvenile sockeye Rapid growth by juvenile sockeye Above average mortality for juvenile sockeye Slow growth by juvenile sockeye Average to above average adult returns Below average adult returns Sockeye Fry Recruitment: increases decreases

37. Salmon have complex life histories that integrate events in freshwater & marine ecosystems. Simple associations don’t reveal freshwater or marine causation (i.e. mechanisms for regime changes in climate & salmon production identified in many recent studies remain untested for freshwater vs marine origins not to mention their specific causes) Offshore Marine Zone Coastal Marine Zone Nearshore Freshwater Zone

38. Conclusions • life history events & production variations of specific salmon populations do co-vary with short (ENSO) & longer term (PDO) climate regime changes. • given complex life histories, generalizations about when, where and how salmon populations will respond to variations in climate regimes will be slow to emerge (e.g. conflicting inferences in existing studies). • research into causal mechanisms will provide new information tools and models to improve fisheries management (e.g. Barkley Sd sockeye return forecasts, Okanagan sockeye migration delay prediction, extinction prospects). Salmon are great indicators of many things but are not simple indicators of hardly anything !

39. Possible Adaptation Responses by Managers • eliminate harvest to compensate for decreased stock productivity (most harvest already eliminated; requires renegotiation of Salmon Treaty provisions). • alter water management rules to favour fish production benefits to compensate for decreased stock productivity (holding back scour inducing flows in Okanagan Lake may induce millions of $$ of flood losses in Kelowna; requires renegotiation of Canada-BC Okanagan Basin Water Agreement). • engineer new water-control facilities to reduce water temp, adult migration and egg hatch delays (capital costs in excess of $100 million). • allocate more stored water in summer-fall to control temp-O2 “squeeze” and compensate for mortality events in Osoyoos Lake (limited summer water worth millions of $$ to users e.g. wineries who won’t give up willingly). • ignore stock declines and risk triggering SARA-COSEWIC listing in Canada and ESA listing in the U.S. (would trigger challenges in courts by water and fisheries resource stakeholders in both countries; could provoke constitutional challenge by First Nations groups in Canada).