2. Detecting Change in the Bering Sea Ecosystem Sergei Rodionov 1, James E. Overland 2, Nicholas A. Bond 1 1 JISAO, University of Washington, Seattle, WA. 2 PMEL, NOAA, Seattle, WA. 1 JISAO, University of Washington, Seattle, WA. 2 PMEL, NOAA, Seattle, WA.

3. -2.5 -2 -1.5 -0.5 0 0.5 1 19001905191019151920192519301910 5% significance level 0 0.1 0.2 0.3 0.4 0.5 0.6 RSI 1910 The SARS Method January PDO Searching for the first regime shift SARS – Sequential Analysis of Regime Shifts RSI – Regime Shift Index l = 10

4. -1.5 -0.5 0 0.5 1 1.5 2 2.5 1900190519101915192019251930 5% significance level 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 RSI 1910 1912 1914 1922 Searching for the next regime shift January PDO l = 10

5. 0.05 p = 0.1 The North Pacific Index (Nov-Mar) 1899-2003 RSI 1924 1948 1977 1924 1948 1977 1958 1989 2003

6. p = 0.05 l = Arctic Oscillation, 1951-2003 10 75 1989 1994 1996 1972 1977

7. www.Beringclimate.noaa.gov

8. 1943 1934 1989 1998 1976 1977 PDOa PDOw PDOs PDOa PDOs PDOa PDOs PDOw ALPI NPI NCAR PNA NPI CPC NPI NCAR PDOw AO EPI PDOs EPI AI NPI CPC Regime Shifts in Climatic Indices p = 10 l = 0.1

9. The 1989 Shift Winter (DJF) Surface Air Temperature

10. l = 10 p = 0.1 1989 Arctic Oscillation, Winter (DJF)

11. 1989 l = 10 p = 0.1 Pacific Decadal Oscillation, Winter (DJF)

12. The North Pacific Index, Winter (NDJFM) l = 10 p = 0.1 1989

13. EPI NPI CPC R = -0.26 Data: 1950-2003 1990 1998 1990 1998 R = -0.70 Data: 1980-2003 l = 10 p = 0.1

14. East Pacific Index (AMJJ) Correlations with SLP 1950-2003 North Pacific Index (AMJJ)

15. Regime Shifts in Atmospheric Indices l = 10 p = 0.1 1929 1938 1997 1969 1959 1977 1989 SLPw SATa BSPI SATw SATa SLPw OWS MIX SATa – Annual surface air temperature, St. Paul. SATw – Winter surface air temperature, St. Paul SLPw – Winter SLP over the Bering Sea BSPI – Bering Sea pressure index OWS – Optimal wind speed for larval feeding, Mooring 2 MIX – Summer wind mixing, Mooring 2

16. 199819891977194719241911 Mean Winter (NDJFM) SLP over the Bering Sea

17. 199819891977194719241940 Mean Winter (DJFM) SAT at St. Paul

18. Regime Shifts in Oceanic Indices 1965 1977 1983 2000 SST Prib SST M2 ICI SST Prib IRI SAT Prib – Winter SST near the Pribilof Islands SAT M2 – Winter SST at Mooring 2 ICI – Ice Cover Index IRI – Ice Retreat Index l = 10 p = 0.1 SST M2 1988

19. Ice Cover Index and Surface Temperature at Mooring 2 ICI Temperature l = 10 p = 0.1 1978 1988 1977

20. Regime Shifts in Biological Indices l = 10 p = 0.1 1977 1966 1981 1984 1992

21. Time Series of Fish Stocks Herring year-class strength Pollock recruitment at age 1 Bristol Bay sockeye salmon runs 1977 1989 1979 1997 1978 1989 1985 2001 l = 10 p = 0.1 l = 5 p = 0.1 l = 10 p = 0.1

22. Conclusions Characteristics of the SARS method: Characteristics of the SARS method: –Automatic detection of regime shifts, –Improved performance at the ends of time series, –Can be tuned up to detect regimes of different scales, –Can handle the incoming data regardless of whether they are presented in the form of anomalies or absolute values, –Works well with the time series containing a trend, –Can be applied to a large set of variables.

23. Conclusions (continued) An application of SARS to the Bering Sea ecosystem demonstrated that An application of SARS to the Bering Sea ecosystem demonstrated that –The shift of 1977 was the strongest one in the last 50 years; –A number of indices experienced a regime shift around 1989 (AO, PDO w, temperature at Mooring 2, herring), 1998 (PDO s, salmon), or both (NPI CPC, EPI, winter SLP, flathead sole); –The regime of 1989-1997 was characterized by a relative winter cooling and reduced cyclonic activity; –Regime shifts in biological indices are not concentrated around certain, dominant years. The RSI values are rather evenly distributed between 1977 and 1992.