Empirical evidence for North Pacific regime shifts in 1977 and 1989
Introduction
In the early 1990s, a wide array of evidence began accumulating that a major climate event had transpired in the mid-1970s, which had widespread consequences for the biota of the North Pacific Ocean and Bering Sea. The event was eventually termed a ‘regime shift’ and several studies have documented the climatic (Graham, 1994; Miller, Cayan, Barnett, Graham & Oberhuber, 1994) and ecosystem (Francis & Hare, 1994, Hare & Francis, 1995; Francis, Hare, Hollowed & Wooster, 1998; McGowan, Cayan & Dorman, 1998) changes that took place. A particularly influential study was that of Ebbesmeyer, Cayan, McLain, Nichols, Peterson and Redmond (1991) who assembled 40 environmental time series and demonstrated that a statistically significant ‘step’ change occurred in a composite of the time series in the winter of 1976–77.
There is no common definition of a regime shift, but certain aspects are generally agreed upon. A regime implies a characteristic behavior of a natural phenomenon (sea level pressure, recruitment, etc.) over time. A shift suggests an abrupt change, in relation to the duration of a regime, from one characteristic behavior to another. Although climate variability occurs across a broad spectrum of spatial and temporal scales, in the context of the current discussion a regime spans a decade or more, whereas a shift occurs within a year or so. There is also an important distinction between regime shift and random walk type variability. In a random walk model of climate variability, which involves gradual random change over time, there is no mean (average) state and, in the absence of negative feedback, variance increases with time.
Recognition of the 1976–77 regime shift has opened the question of whether that event was unique or merely the latest in a sequence of regime shifts in the historical record. Based on analyses of temperature, pressure, tree ring and salmon catch records, several researchers have hypothesized that earlier climate shifts in the North Pacific occurred in the early 1920s and mid 1940s (Kondo, 1988; Mantua, Hare, Zhang, Wallace & Francis, 1997; Zhang, Wallace & Battisti, 1997; Minobe, 1997; Ingraham, Ebbesmeyer & Hinrichsen, 1998). Mantua et al. (1997) coined the term ‘Pacific Decadal Oscillation’ (PDO) to describe this interdecadal climate variability. They described the PDO as a long-lived El Niño–Southern Oscillation (ENSO) like pattern of Pacific climate variability. As seen with ENSO, extremes in the PDO pattern are marked by widespread variations in Pacific Basin and North American climate. Viewed from another perspective, extremes in the (tropical) ENSO cycle often influence North Pacific climate in PDO-like ways. The exceptional El Niño of 1997–1998 is a clear case in point, wherein changes in tropical rainfall and atmospheric circulation ‘forced’ strong and persistent climate anomalies over the North Pacific (Barnston et al., 1999). Two main characteristics distinguish the PDO from ENSO. First, typical PDO ‘events’ have shown remarkable persistence relative to that attributed to ENSO events — in this century, major PDO regimes have persisted for 20 to 30 years. Second, the climatic fingerprints of the PDO are most visible in the North Pacific/North American sector, while secondary signatures exist in the tropics — the opposite is true for ENSO.
It is of great interest to determine whether or not the Pacific has undergone a further climatic regime shift since the 1976–77 event, both for practical and for theoretical reasons. Within the field of fisheries, recognition of the impact of the 1976–77 regime shift has influenced management decisions. For example, certain salmon runs are now managed and optimal catch levels computed under the assumption that data collected prior to the mid 1970s is no longer relevant to modeling the dynamics of present-day salmon runs (Beverly Cross, Alaska Department of Fish and Game, personal communication). Similarly, investigations of the optimal harvest rate for the Pacific halibut fishery use regime shift models as one expression of the recruitment process (Clark, Hare, Parma, Sullivan & Trumble, 1999). To a large extent, however, most fisheries management is based on models that either 1) ignore environmental influences on population dynamics processes; 2) assume them to be random and without trend; or 3) subsume influences into an additive error term. Several high profile fisheries management failures (e.g., North Atlantic cod) have highlighted the fact that our current understanding and modeling of commercial fish populations are inadequate. We believe that only by adopting a more holistic view, including the incorporation of environmental forcing, will we increase our understanding of fish population dynamics and so be able to manage them so as to optimize the tradeoffs better between harvest and sustainability.
In this study we evaluate empirical evidence for North Pacific regime shifts in the 1965–1997 period, with a focus on 1989 as a year of special interest. As noted, many studies offer convincing evidence for the occurrence of an important North Pacific regime shift in 1977. Subsequent studies (published in the mid-to-late 1990s) have suggested that another regime shift occurred in the winter of 1988–89 (Polovina, Mitchum, Graham, Craig, DeMartini & Flint, 1994; Mackas, 1995; Sugimoto & Tadokoro, 1998, Watanabe & Nitta, 1999; Overland, Adams & Bond, 1999; Beamish, Noakes, McFarlane, Klyashtorin, Ivanov & Kurashov, 1999; Brodeur, Mills, Overland, Walters & Schumacher, 1999; Welch, Ward, Smith & Eveson, 2000), whereas others have suggested that the post-1977 regime persisted through (at least) 1997 (Mantua et al., 1997, McGowan et al., 1998, Ingraham et al., 1998). Assessments of some of the Pacific climate indicators that were sensitive to the 1977 regime shift have contributed to the controversy over the existence of a 1989 regime shift. For example, Beamish et al. (1999) used four Northern Hemisphere climate indices to form a single ‘regime index’ to examine the late 1980s–early 1990s period for evidence of a regime shift, and their results were equivocal. Our approach follows that of Ebbesmeyer et al. (1991), in which a diverse array of physical and biological data are examined to determine the statistical significance of regime changes.
Section snippets
Data and methods
For this study we assembled 100 physical and biological time series. We attempted to select a broadly representative set of environmental indicators. However, availability (or lack thereof) inevitably shaped the selection process. Additionally, several time series were chosen because they have been used by other researchers as indicators of decadal scale climate/ecosystem change. Our region of interest is the North Pacific Ocean and Bering Sea, i.e., regions that have demonstrated previous
Principal component analysis
We used principal component analysis (PCA) to isolate objectively the most important patterns of common variability in the 100 physical and biological time series. Eigenvalue analysis indicates that only the first two principal components are meaningful. Error bars for higher order PCs (computed using the formula of North, Bell, Cahalan & Moeng, 1982) overlap indicating that the patterns are potentially mixed and non-interpretable (Fig. 2). Scores for the first two PCs are illustrated in Fig. 3
Discussion
We have employed an objective statistical method — Principal Component Analysis (PCA) — to identify temporally coherent changes in indices for parts of the large marine ecosystems of the North Pacific and Bering Sea. PCA requires no a priori assumptions about specific years in the historical record. As a second metric for the significance of regime shift changes, we have employed Ebbesmeyer et al.'s (1991) method, which does require an a priori specification of regime years. The results from
Acknowledgements
This work was motivated by discussions and presentations held during a series of informal meetings — on the subject of climate impacts on marine resources — held in Seattle between 1997 and 1999. We wish to thank the 20–30 scientists involved in those meetings, many of whom contributed time series used in our analysis. These time series, along with meeting summaries, are available at http://www.iphc.washington.edu/Staff/hare/html/decadal/post1977/post1977.html. We especially wish to acknowledge
References (52)
- et al.
Have there been recent changes in climate? Ask the fish
Progress in Oceanography
(2000) Spatio-temporal structure of the pentadecadal variability over the North Pacific
Progress in Oceanography
(2000)- et al.
Community reorganization in the Gulf of Alaska following ocean climate regime shift
Marine Ecology Progress Series
(1999) - et al.
Classification, seasonality and persistence of low-frequency atmospheric circulation patterns
Monthly Weather Review
(1987) - Barnston, A. G., & co-authors. (1999). NCEP forecasts of the El Niño of 1997–98 and its US impacts. Bulletin of the...
- et al.
Pacific salmon production trends in relation to climate
Canadian Journal of Fisheries and Aquatic Science
(1993) - et al.
The regime concept and natural trends in the production of Pacific salmon
Canadian Journal of Fisheries and Aquatic Science
(1999) Atmospheric teleconnections from the equatorial Pacific
Monthly Weather Review
(1969)- et al.
Interannual variations in zooplankton biomass in the Gulf of Alaska and covariation with California Current zooplankton
California Cooperative Oceanic Fisheries Investigation Report
(1996) - et al.
Evidence for a substantial increase in gelatinous zooplankton in the Bering Sea with possible links to climate change
Fisheries Oceanography
(1999)