See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225139766

The Pacific Decadal Oscillation

Article in Journal of Oceanography · February 2002

DOI: 10.1023/A:1015820616384

CITATIONS READS

1,989 5,641

2 authors:

Nathan J. Mantua Steven Hare

National Oceanic and Atmospheric Administration The Pacific Community
85 PUBLICATIONS 15,814 CITATIONS 104 PUBLICATIONS 21,225 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Steven Hare on 02 June 2014.

The user has requested enhancement of the downloaded file.

Journal of Oceanography, Vol. 58, pp. 35 to 44, 2002

Review

The Pacific Decadal Oscillation

NATHAN J. M ANTUA1 * and S TEVEN R. HARE2
1
University of Washington, Joint Institute for the Study of the Atmosphere and Oceans,
Seattle, WA 98195-4235, U.S.A.
2
International Pacific Halibut Commission, P.O. Box 95009, Seattle, WA 98195-2009, U.S.A.

(Received 19 May 2001; in revised form 16 August 2001; accepted 16 August 2001)

The Pacific Decadal Oscillation (PDO) has been described by some as a long-lived El Keywords:
Niño-like pattern of Pacific climate variability, and by others as a blend of two some- ⋅ Regime shift,
times independent modes having distinct spatial and temporal characteristics of North ⋅ climate impacts,
Pacific sea surface temperature (SST) variability. A growing body of evidence high- ⋅ PDO,
⋅ IPO,
lights a strong tendency for PDO impacts in the Southern Hemisphere, with impor-
⋅ NPO,
tant surface climate anomalies over the mid-latitude South Pacific Ocean, Australia
⋅ fishery oceanogra-
and South America. Several independent studies find evidence for just two full PDO phy.
cycles in the past century: “cool” PDO regimes prevailed from 1890–1924 and again
from 1947–1976, while “warm” PDO regimes dominated from 1925–1946 and from
1977 through (at least) the mid-1990’s. Interdecadal changes in Pacific climate have
widespread impacts on natural systems, including water resources in the Americas
and many marine fisheries in the North Pacific. Tree-ring and Pacific coral based
climate reconstructions suggest that PDO variations—at a range of varying time
scales—can be traced back to at least 1600, although there are important differences
between different proxy reconstructions. While 20th Century PDO fluctuations were
most energetic in two general periodicities—one from 15-to-25 years, and the other
from 50-to-70 years—the mechanisms causing PDO variability remain unclear. To
date, there is little in the way of observational evidence to support a mid-latitude
coupled air-sea interaction for PDO, though there are several well-understood mecha-
nisms that promote multi-year persistence in North Pacific upper ocean temperature
anomalies.

1. Introduction years. This situation, which originated with a strongly

Climate records from around the Pacific Basin con- anomalous winter in 1976–1977, has been termed a “cli-
tain evidence for strong interannual to interdecadal vari- matic regime”, following a regime shift in 1977. The 1977
ability, in special cases with remarkably large-scales change in Pacific climate was first reported by Nitta and
(O(104 km)) of spatial coherence (NRC, 1998). El Niño/ Yamada (1989) and Trenberth (1990), who described a
Southern Oscillation (ENSO) has long been known to be step-like shift in the mean state of winter sea level pres-
the prominent source for hemispheric-scale interannual sure (SLP) in the North Pacific. Miller et al. (1994) pro-
climate variations for the Pacific and the global tropics vided the first detailed depiction of the climatic changes
(Rasmussen and Wallace, 1983). In the last two decades and dubbed the 1976/77 North Pacific event a regime shift.
of the 20th Century, the extratropical Pacific Ocean was Biologists noted dramatic late-1970’s changes in
in an almost continuous El Niño-like state despite the much of the biota around the North Pacific. Ebbesmeyer
absence of tropical El Niño events in a majority of those et al. (1991) quantified the change in 40 “environmen-
tal” (climatic and biological) variables demonstrating a
statistically significant step between 1976 and 1977 in a
* Corresponding author. E-mail: [email protected]. composite of the time series. It was observations on Pa-
edu cific salmon, however, specifically the catch history of
Copyright © The Oceanographic Society of Japan. Pacific salmon going back 70 years, that provided the most

35
tantalizing evidence that a definite link existed between Minobe, 1997). These and other studies also provided
interdecadal changes in North Pacific climate and North evidence that PDO variations had considerable influence
Pacific fisheries. In a series of papers, Francis and Hare on climate-sensitive natural resources in the Pacific and
focused on Alaska salmon production and its link to cli- over parts of North America in the 20th Century.
mate (Francis and Hare, 1994; Hare and Francis, 1995; Subsequent study has revealed several new and im-
Francis and Hare, 1997), arguing that Alaska salmon pro- portant wrinkles to a rapidly growing literature on the
duction was best characterized as alternating regimes, general topic of PDV and on the nature of the PDO. Ac-
where the transition from one regime to another was cumulating evidence suggests that the PDO mode of vari-
abrupt. ability exhibits a robust symmetry in interdecadal climate
The race to describe and understand interdecadal variations of the Northern and Southern Hemispheres (e.g.
changes in the Pacific accelerated through the 1990’s. White and Cayan, 1998; Garreaud and Battisti, 1999;
Latif and Barnett (1996) provided a comparison of the Dettinger et al., 2000), with signature responses in East
low-frequency variability in observations with that in the Asia, North, South and Central America, and Australia.
output from a coupled ocean/atmosphere model simula- Historical records tracking aspects of Pacific marine eco-
tion, and proposed a mechanism for Pacific Decadal Vari- systems suggest a strong association between PDO vari-
ability (PDV) with a near-20 year periodicity. Zhang et ability and Pacific salmon production (Beamish and
al. (1997) offered a series of analyses teasing apart sub- Bouillon, 1993; Beamish et al., 1999; Hare et al., 1999),
tle spatial differences between Pacific climate variability Pacific sea birds (Vandenbosch, 2000), Alaska ground fish
at interannual versus interdecadal time scales. Mantua et and zooplankton production in the central and eastern
al. (1997) capitalized on the maturity of the rapidly evolv- North Pacific (Hollowed et al., 1998; Francis et al., 1998),
ing research, synthesizing and extending research results and Gulf of Alaska marine species assemblages (Anderson
from fishery, climate and hydroclimate studies, and and Piatt, 1999), to name just a few. Careful reconstruc-
labeled the dominant pattern of PDV the Pacific tions of instrumental data have extended the PDO record
(inter)Decadal Oscillation (PDO). Other studies have back to 1854 (Kaplan et al., 2000), and paleoclimate re-
used other names for what we call the PDO, for example: constructions now provide an extended, albeit sometimes
the Interdecadal Pacific Oscillation (IPO) of Power et al. contradictory, view of PDV and PDO behavior back to
(1997, 1999a), and the North Pacific Oscillation (NPO) 1600 (cf. Minobe, 1997; Evans et al., 2000; Linsley et
of Gershunov and Barnett (1998). al., 2000; Biondi et al., 2001; Gedalof and Smith, 2001).
The collective body of research suggested that three Research into the dynamics of PDV has also pro-
main characteristics distinguished PDO from ENSO: first, duced numerous publications, yet at this time mechanisms
20th century PDO “events” persisted for 20-to-30 years, for PDO behavior remain mysterious (see Miller and
while typical ENSO events persisted for 6 to 18 months; Schneider (2000) for a comprehensive review). In spite
second, the climatic fingerprints of the PDO were most of the remaining mysteries, a number of insights into
visible in the extratropics, especially the North Pacific/ mechanisms favoring multi-year persistence of North
North American sector, while secondary signatures ex- Pacific climate anomalies have recently come to light
isted in the tropics, and the opposite was true for ENSO; (Schneider and Miller, 2001; Seager et al., 2001; Deser
and third, the mechanisms causing PDO variability were (Clara Deser, NCAR, Boulder Colorado, personal com-
not known, while causes for ENSO variability were rela- munication); and Barsugli and Battisti, 1998), indicating
tively well-understood (Zhang et al., 1997; Mantua et al., promising prospects for PDV predictability at lead times
1997; NRC, 1998). of one to a few years.
A PDO index developed by Hare (1996) and Zhang Mantua et al. (1997) proposed that the PDO repre-
(1996), also used by Mantua et al. (1997), is the leading sents a special class of PDV defined by a preferred spa-
PC from an un-rotated EOF analysis of monthly, “re- tial pattern with a range of interdecadal time scales of
sidual” North Pacific sea surface temperature (SST) variability. We argue here that the case for a robust PDO
anomalies, poleward of 20°N for the 1900–1993 period mode of PDV is, on balance, strengthened by the results
of record (see lower panel of Fig. 1). “Residuals” are here of recent studies, although many critical questions about
defined as the difference between observed anomalies and the PDO await answers. Whether there is a preferred PDO
the monthly mean global average SST anomaly (see Zhang time scale is critical for several reasons, including the
et al., 1997). A remarkable characteristic of this index is issue of mechanisms and how understanding those mecha-
its tendency for multiyear and multidecadal persistence, nisms should aid the development of a PDO monitoring
with a few instances of abrupt sign changes. Based on a and prediction system. Regardless of PDO predictability,
variety of studies, sign changes beginning in 1925, 1947, we also believe that recognition of PDO variability is
and 1977 have been labeled regime shifts (Hare and important because it clearly demonstrates that “normal”
Francis, 1995; Zhang et al., 1997; Mantua et al., 1997; climate conditions can vary over time periods compara-

36 N. J. Mantua and S. R. Hare

ble to the length of a human’s lifetime, and climate anoma- sistence in the Pacific North America (PNA)
lies that persist for one to a few decades can cause espe- teleconnection pattern described by Wallace and Gutzler
cially large impacts on ecosystems and societies. (1981) (not shown). Because all these patterns were de-
For brevity, we will provide only a select review of rived from linear analyses, climate anomalies associated
PDO research in the remainder of this article, and in do- with cool phases of the PDO are simply opposites of those
ing so will omit many valuable research results. We apolo- for warm PDO phases (not shown).
gize here for those omissions, but hope that our survey Although the PDO mode of variability has been dis-
offers readers a solid foundation for the present state of cussed widely in the literature, the more general quest
PDO research. for understanding PDV is an area of very active research.
One important lesson is clear from the published litera-
2. PDO Characteristics ture: different analyses yield different descriptions of 20th
Century PDV. Some studies find evidence for distinct and
2.1 Spatial patterns independent lobes of North Pacific SST variability em-
Typical sea surface temperature, surface wind, and bedded within the canonical PDO pattern shown in Fig.
sea level pressure anomaly patterns for warm phases of 1. Nakamura et al. (1997) examined low-pass filtered
the PDO are shown in the top panels of Fig. 1. During North Pacific SST data for the 1968–1992 period of record
warm PDO phases sea surface temperatures (SSTs) tend and identified two independent centers of action, one en-
to be anomalously cool in the central North Pacific coin- compassing the subtropical front north of Hawaii and the
cident with anomalously warm SSTs along the west coast other encompassing the subarctic front that defines the
of the Americas. For November-to-March averages, warm Kuroshio/Oyashio Extension. Barlow et al. (2001)
PDO sea level pressure (SLP) anomalies have low pres- analyzed Pacific SSTs for the 1945–1993 period of record
sures over the North Pacific which cause enhanced and identified a different pair of North Pacific SST modes,
counterclockwise winds, and high SLP over the northern each spatially correlated with the canonical PDO pattern
subtropical Pacific which cause enhanced clockwise of Mantua et al. (1997). In contrast, Kaplan et al. (2000)
winds. Anomalously high SLP in the western tropical applied an optimal interpolation scheme to available SST
Pacific and low SLP in the eastern tropical Pacific depict and SLP records for the 1854–1992 period of record, then
a relatively weak negative phase of the Southern Oscilla- recovered the PDO pattern as the second leading mode of
tion (see Trenberth and Shea, 1987). PDO circulation co-variability between the five-year low-pass-filtered glo-
anomalies in the Northern Hemisphere extend through the bal fields (the leading mode of co-variability was a trend
depth of the troposphere, and are well-expressed as per- mode). While these results yield somewhat different pic-

Fig. 1. (top) Anomalous climate conditions associated with warm phases of the Pacific Decadal Oscillation (PDO), and (bottom)
November–March average values of the PDO index. Values shown are °C for sea surface temperature (SST), millibars for sea
level pressure (SLP) and direction and intensity of surface wind stress. The longest wind vectors represent a pseudostress of
10 m 2/s 2. Actual anomaly values for a given year at a given location are obtained by multiplying the climate anomaly by the
associated index value. Adapted and updated from Mantua et al. (1997).

The Pacific Decadal Oscillation 37

tures of past PDV, there remains a wealth of evidence in years back in time. Minobe (1997) used Fritts’ (1991) tree-
support of spatial modes that generally resemble, if not ring based temperature reconstructions to project North
reproduce, the canonical PDO pattern (cf. Tanimoto et American air temperatures back to 1600. The leading EOF
al., 1993; Graham, 1994; Trenberth and Hurrell, 1994; had the same 50–70 peak periodicity as the instrumental
Latif and Barnett, 1994, 1996; Zhang, 1996; Hare, 1996; record from which the PDO was identified. Biondi et al.
Mantua et al., 1997; Minobe, 1997; Nakamura et al., 1997; (2001) used ring-widths from moisture stressed trees in
Enfield and Mestas-Nuñez, 1999; Folland et al., 1999; Southern California and Baja, Mexico, to create a paleo-
Kaplan et al., 2000; Barlow et al., 2001; Tourre et al., PDO time series to 1661; Gedalof and Smith (2001) used
2001). tree-ring chronologies from a coastal transect spanning
northern California to the Gulf of Alaska to reconstruct a
2.2 Temporal scales of variability PDO index to 1600. The PDO index was positively cor-
Research aimed at identifying temporal scales of related with the dominant climate signal in the 20th Cen-
PDV also yield a variety of results, again based on the tury sections of these two dendrochronologies (Fig. 2, see
data examined and the analysis techniques employed. In Table 1). Gedalof and Smith (2001) identified 11 regime
a pair of closely related studies, Minobe (1999, 2000) shifts in the PDO record since 1650 with the most recent
applied Wavelet analysis to indices for boreal winter and occurring in 1976/77. With average duration of a regime
spring North Pacific SST and SLP and found PDO fluc- being 23 years, they suggest that another shift is due
tuations were most energetic at periodicities in the 15-to- around the end of the century. While the two
25 year and 50-to-70 year bands. Chao et al. (2000) ap- dendrochronologies capture much of the interdecadal
plied Singular Spectrum Analysis to a persistence index variability in the instrumental PDO indices of Mantua et
for North Pacific SST variations, and they found evidence al. (1997) and Kaplan et al. (2000) (Fig. 2), they also
for oscillatory variations at 15-to-20 and near 70 year exhibit periods in which they show little, if any corre-
periodicities. Tourre et al. (2001) used a Multi-Taper- spondence with each other. This situation warrants fur-
Method/Singular Value Decomposition (MTM/SVD) ther investigation, and highlights opportunities to narrow
technique to identify coherent patterns of low-frequency the uncertainty of pre-instrumental PDV, perhaps through
20th Century Pacific SST and SLP variations from 30°S multi-proxy reconstructions.
to 60°N. The canonical PDO SST pattern shown in Fig. 1
is clearly reproduced by the spatial patterns of Tourre’s
et al. (2001) Interdecadal mode (which has peak variance Table 1. Correlation coefficients between the time series dis-
played in Fig. 2. Correlations were computed on the com-
at 12-to-25 year periods), and somewhat similar to that
mon period of record (1903–1981). Note that these time
of their Decadal mode (which has peak variance at 9-to-
series are 5-year running averages of the raw data series.
12 year periods) (see Tourre’s et al. (2001); figure 2).
PDO Index Gedalof Biondi Kaplan
2.3 Paleoclimate reconstructions
To better understand the long-term behavior of the Gedalof 0.55
PDO, several studies report on proxy environmental re- Biondi 0.58 0.31
Kaplan 0.77 0.58 0.31
corders of PDO-related climate changes several hundred

Fig. 2. 5-year running average plots of tree-ring based PDO reconstructions of Gedalof and Smith (2001) and Biondi et al.
(2001), along with Kaplan et al.’s (2000) COADS SST index for 1854–1992 and Mantua et al.’s (1997) SST-based PDO
index. Each time series has been normalized with respect to the available period of record, and they are plotted with an offset
for clarity.

38 N. J. Mantua and S. R. Hare

 Results of PDO index reconstructions from outside
North America have also been published. Evans et al.
(2000) examined 15 tree-ring chronologies from mid-lati-
tude North and South America and found 20th Century
coherence in these records closely matched that in the
PDO index. Linsley et al. (2000) examined Sr/Ca vari-
ability in a long lived coral from Rarotonga and found a
strong PDO signal in the extracted coral SST history that
spans the period from 1726 to 1997. These last two proxy
records are of special interest because they substantiate a
robust PDO connection to tropical and southern hemi-
sphere climate (Evans et al., 2001).

3. PDO Impacts

3.1 Surface climate

Many of the climate anomalies associated with PDO
are broadly similar to those connected with ENSO varia-
tions (El Niño and La Niña), though generally not as ex-
treme (Latif and Barnett, 1996; Mantua et al., 1997;
Minobe, 1997). Correlations between the November–
April PDO index and the 0.5 degree gridded surface tem-
perature and precipitation data of Willmott and Matsuura
(2000) (see also Willmott and Robeson, 1995) are shown
in Fig. 3.
The correlations suggest the following patterns of
PDO precipitation anomalies: warm phases of the PDO
coincide with anomalously dry periods in eastern Aus-
tralia, Korea, Japan, the Russian Far East, interior Alaska,
in a zonally elongated belt from the Pacific Northwest to
the Great Lakes, the Ohio Valley, and in much of Central
America and northern South America; warm PDO phases
also tend to coincide with anomalously wet periods in Fig. 3. Correlations between November–April mean precipita-
the coastal Gulf of Alaska, the southwest US and Mexico, tion (top) and temperature (bottom) and the November–
southeast Brazil, south central South America, and west- April mean PDO index shown in Fig. 1. Precipitation and
ern Australia. temperature data are the 0.5 degree grid climatologically
The correlations suggest the following patterns of aided interpolation (CAI) fields produced at the University
November–April PDO temperature anomalies: warm of Delaware by Cort Willmott and collaborators (available
via the internet at http://climate.geog.udel.edu/, also see
phases of the PDO tend to coincide with anomalously
Willmott and Robeson, 1995). Negative correlation coeffi-
warm temperatures in northwestern North America, north- cients are shaded in blues, positive correlation coefficients
ern South America, and northwestern Australia, and are shaded in reds and yellows.
anomalously cool temperatures in eastern China, Korea,
Japan, Kamchatka, and the southeast US and Mexico. It
is notable that Minobe (2000) and Cayan et al. (2001)
find that the most prominent PDO temperature signal in tober-to-September) streamflow anomalies in the Ameri-
North America is in the boreal spring, rather than winter cas, wherein warm PDO (El Niño-like) periods tend to
season. have anomalously wet subtropics but dry tropics and
Independent studies have confirmed PDO signals in midlatitudes in both North and South America. Power et
the Southern Hemisphere. Garreaud and Battisti (1999) al. (1997, 1999a, 1999b) examined interdecadal changes
extended the study of Zhang et al. (1997) to the Southern in eastern Australian climate, finding warm PDO periods
Hemisphere and identified a clear pattern of symmetric to be associated with anomalously warm-dry conditions,
atmospheric circulation changes associated with the PDO. while cool PDO periods are associated with cool-wet con-
Dettinger et al. (2000) found evidence for a symmetric ditions, consistent with the correlation fields displayed
pattern of PDO-related precipitation and water year (Oc- in Fig. 3.

The Pacific Decadal Oscillation 39

4. Marine Ecosystems
In the past few decades a number of studies have
identified compelling evidence for connections between
PDV and variations in Pacific marine ecosystems.
Kawasaki (1991) has documented a remarkable 20th Cen-
tury coherence between interdecadal fluctuations in sar-
dine population off Japan, California, Chile and Peru (see
also Yasuda et al., 1999). Studies linking 20th Century
Pacific salmon catches in eastern Asia and western North
America to variability in the Aleutian Low have been
published by Beamish and co-workers (Beamish, 1993;
Beamish and Bouillon, 1993; Beamish et al., 1999).
The differing regional responses of salmon stocks
along the west coast of North America have been exam-
ined by Adkison et al. (1996) and Peterman et al. (1998).
Their findings indicated that Alaskan stocks showed a
strong uniform response to climate but British Columbia
stocks were mixed. Hare et al. (1999) extended the geo-
graphic scope to include stocks from Washington, Oregon
and California and analyzed catch records from the five
major salmon species. They identified an “inverse pro-
duction regime”, associated with the PDO, where the
warm phase of the PDO favors high production for Alaska
stocks and low production for Washington, Oregon and
California (WOC) stocks. The cool phase of the PDO has
the opposite effect on Alaska and WOC stocks. The es- Fig. 4. A graphical depiction of the “Inverse Production Re-
sence of the results of their analysis is illustrated in Fig. gimes” of Hare et al. (1999). The bars represent loadings
4. from a principal component analysis (PCA) of 30 salmon
The response of groundfish stocks to the PDO has time series for the period 1925–1997. Regional definitions
also been documented in several studies. A strong one are as follows: 1 - Western Alaska, 2 - Central Alaska,
year jump in recruitment coincident with the 1976–77 3 - Southeast Alaska, 4 - British Columbia, 5 - Washington,
regime shift was demonstrated for many commercially 6 - Oregon, 7 - California. Three climate indices were in-
exploited stocks in the Northeast Pacific by Beamish cluded in the PCA: Pacific Decadal Oscillation (PDO), Aleu-
(1993) and for sablefish in particular (McFarlane and tian Low Pressure Index (AL) and the El Niño-Southern
Oscillation (ENSO). The longest bar, Central Alaska pink
Beamish, 1992). Pacific halibut recruitment was shown
salmon, represents a correlation coefficient with a value of
by Clark et al. (1999) to have undergone interdecadal
0.855, and represents the correlation between that time se-
shifts closely matched to the phases of the PDO (Fig. 4). ries and the illustrated temporal component (score) from
Like Alaska salmon, halibut flourish during warm phases the PCA.
of the PDO. Hollowed et al. (1998) assembled recruit-
ment time series for the major exploited groundfish and
pelagic species in Alaska and WOC. They found that,
while a large fraction of the species appeared to respond A number of other studies have shown impacts of
more to ENSO events, several flatfish species (arrowtooth the PDO on other components of the marine and terres-
flounder, Greenland turbot, Pacific halibut) exhibited trial ecosystems of the North Pacific and North America.
PDO-like recruitment histories. In one of the most thor- At the plankton level, primary and secondary productiv-
ough documentations of the changes that have taken place ity responses to the climate shift of 1976–77 have been
in the groundfish complex, Anderson and Piatt (1999) documented by Venrick et al. (1987), Brodeur and Ware
assembled 45 years of small mesh trawl survey records (1992), Brodeur et al. (1996), Roemmich and McGowan
from the Gulf of Alaska. They show that the marine eco- (1995), McGowan et al. (1998), and Mackas et al. (1998).
system underwent a transformation from one dominated At the higher, non-piscivore, trophic levels, Piatt and
by lower trophic level forage species (e.g. capelin, shrimp, Anderson (1995) and Francis et al. (1998) discuss decadal
sand lance) prior to the mid-1970s, to one dominated by changes in marine mammal and pisciverous bird
higher trophic level groundfish (e.g. gadids and flatfish) populations, particularly in response to the climatic re-
since that time. gime shift of 1976–77. More recently, Vandenbosch

40 N. J. Mantua and S. R. Hare

(2000) linked Hawaiian Island and Farallon Island seabird tian Low, the local intensification of the westerlies cool
population variability to phases of the PDO. Finally, the interior North Pacific via enhanced surface heat fluxes
Cayan et al. (2001) document a long term change to an and anomalous Ekman advection of the mean meridional
earlier onset of spring in the western United States—as temperature field; this wind field also displaces the zero
measured by blooming times of lilac and honeysuckle windstress curl line to lower latitudes, thereby generat-
bushes—and variability in this timing also shows a high ing anomalous upwelling Rossby Waves at the latitude of
correlation with the springtime PDO index. In perhaps zero-windstress curl that slowly propagate to the west;
the broadest examinations to date of the widespread cli- these upwelling Rossby waves eventually raise the
matic impacts on the ecosystems of the North Pacific, thermocline in the KOE region one to several years hence;
Hare and Mantua (2000) conducted a principal compo- subsequent deep winter mixing in the KOE region trans-
nent analysis on a matrix of 100 climatic and biological mits the thermocline anomalies to the surface where they
time series. The climatic time series were selected to rep- eventually cool the SST. The end result is that the same
resent the atmosphere and ocean across the North Pacific wind anomalies that generate negative SST anomalies in
while the biological time series ranged across all trophic the interior North Pacific eventually generate SST anoma-
levels. The dominant principal component has the same lies of the same sign in the KOE region one to a few years
time trajectory as the PDO. later. Any persistence in those winds will therefore result
in an amplified persistence in the North Pacific SST field.
5. Dynamics and Predictability
The physical mechanism(s) behind the PDO are not 6. Discussion and Conclusions
currently known. Some climate simulation models pro- We have provided a review of late 20th Century stud-
duce PDO-like oscillations (e.g. Latif and Barnett, 1994), ies of Pacific Decadal Variability (PDV), with particular
although often for different reasons (NRC, 1998). The attention to a special case of PDV known as the Pacific
mechanisms giving rise to PDO will determine whether Decadal Oscillation (PDO). Much controversy now ex-
skillful PDO climate predictions for up to one or more ists over how PDO works, and how it might best be moni-
decades into the future are possible. Even in the absence tored, modeled and predicted. The stakes in PDO science
of a theoretical understanding, PDO climate information are high, as an improved PDO understanding offers even
improves season-to-season and year-to-year climate fore- sharper views of future climate and its attendant impacts
casts for North America because of its strong tendency on resources than those now provided by ENSO science
for multi-season and multi-year persistence. alone.
While causes for the PDO remain unclear, several We believe that the case for a robust PDO mode of
mechanisms promoting persistence in extratropical cli- PDV is, on balance, strengthened by the results of recent
mate have been identified. Alexander et al. (1999, 2001) studies, while acknowledging the fact that many critical
detail simple mixed layer mechanisms that give rise to questions about the PDO remain unanswered. Regardless
the reemergence of subsurface thermal anomalies from of PDO predictability, we also believe that recognition
one winter to the next. Deser (Clara Deser, NCAR, per- of PDO variability is important because it clearly dem-
sonal communication) reports on the ability to reproduce onstrates that “normal” climate conditions can vary over
the observed multiyear autocorrelation structure of North time periods comparable to the length of a human’s life-
Pacific and North Atlantic SSTs with a simple entraining time, and climate anomalies that persist for one to a few
mixed-layer model. Barsugli and Battisti (1998) use a sim- decades can cause especially large impacts on ecosystems
ple model to demonstrate that local air-sea interactions and societies.
yield “differential thermal damping” on atmospheric Recent advances in understanding mechanisms for
anomalies that redden the spectrum of variability and in- persistence and slow changes in extratropical SST anoma-
crease the overall variance over that which would occur lies offer improved confidence for PDV predictability at
in the absence of feedback. lead times of one to a few years. Accurate PDO monitor-
The complimentary results of Seager et al. (in re- ing and prediction may have practical benefits in both
view) and Schneider and Miller (2001 (in press)) offer seasonal and longer term climate forecasts for select re-
strong support for multi-year predictability for important gions. Gershunov and Barnett (1998), for example, ar-
aspects of mid-latitude ocean variability. Both studies gued that combining PDO and ENSO information may
report on the dynamic response of the thermocline in the enhance the skill of empirical North American climate
Kuroshio/Oyashio Extension (KOE) region to the inte- forecasts.
grated wind-stress curl over the previous few years. Their The potential for skillful PDV predictions at lead
hypothesis suggests the following: in a situation with times beyond a few years hinges on the premise that un-
enhanced westerly surface winds over the central North stable coupled ocean-atmosphere interactions and delayed
Pacific, for instance due to an anomalously deep Aleu- negative feedbacks contribute to PDV. Direct observa-

The Pacific Decadal Oscillation 41

tional evidence for these types of interactions, at least fects on sockeye salmon (Oncorhynchus nerka) productiv-
outside the tropical Pacific, is tantalizing yet equivocal ity in Bristol Bay, Alaska and Fraser River, British Colum-
(see NRC, 1998; Miller and Schneider, 2000). Proxy and bia. Fish. Oceanogr., 5, 137–152.
instrumental evidence for robust PDO impacts within and Alexander, M. A., C. Deser and M. S. Timlin (1999): The
reemergence of SST anomalies in the North Pacific Ocean.
around the Pacific Rim, both in the tropics and mid-lati-
J. Climate, 12, No. 8, 2419–2433.
tudes of the Northern and Southern Hemisphere, supports
Alexander, M. A., M. S. Timlin and J. D. Scott (2001): Winter-
the idea that causes for PDO variability originate in the to-winter recurrence of sea surface temperature, salinity and
tropics (Evans et al., 2001). mixed layer depth. Prog. Oceanogr., 49, 41–62.
The present day skill in PDO-related forecasts comes Anderson, P. J. and J. F. Piatt (1999): Community reorganiza-
from persistence. This skill disappears when there is an tion in the Gulf of Alaska following ocean climate regime
unforeseen sign change in the PDO pattern. Such a shift. Mar. Ecol. Prog. Ser., 189, 117–123.
change—a flip from warm to cool PDO phases—may have Barlow, M., S. Nigam and E. H. Berbery (2001): ENSO, Pa-
taken place in 1998, coincident with the demise of the cific decadal variability, and U.S. summertime precipita-
1997/98 El Niño and the beginning of the subsequent La tion, drought, and streamflow. J. Climate, 14, 2105–2128.
Niña episode (Hare and Mantua, 2000; Schwing and Barsugli, J. J. and D. S. Battisti (1998): The basic effects of
atmosphere-ocean thermal coupling on midlatitude variabil-
Moore, 2000). However, because no one is certain how
ity. J. Atmos. Sci., 55, 477–493.
the PDO works, it is not possible to say with confidence
Beamish, R. J. (1993): Climate and exceptional fish produc-
that the 1998 changes in Pacific climate mark the begin- tion off the west coast of North America. Can. J. Fish. Aquat.
ning of a 20-to-30 year long cool phase of the PDO. Sci., 50, 2270–2291.
The lack of PDO understanding presents a barrier to Beamish, R. J. and D. R. Bouillon (1993): Pacific salmon pro-
both real-time monitoring and forecasting PDO regime duction trends in relation to climate. Can. J. Fish. Aquat.
shifts. The research community’s experience with ENSO Sci., 50, 1002–1016.
demonstrated that improved understanding and predic- Beamish, R. J., D. J. Noakes, G. A. McFarlane, L. Klyashtorin,
tions came with the synergy of observational, theoreti- V. V. Ivanov and V. Kurashov (1999): The regime concept
cal, and modeling studies (NRC, 1996). Each of these lines and natural trends in the production of Pacific salmon. Can.
of PDO research have been identified as high priorities J. Fish. Aquat. Sci., 56, 516–526.
Biondi, F., A. Gershunov and D. R. Cayan (2001): North Pa-
by the ongoing International CLIVAR program (see
cific decadal climate variability since 1661. J. Climate, 14,
http://www.clivar.org/). PDO science is relatively new
5–10.
compared to ENSO science, but insights into the PDO Brodeur, R. D. and D. M. Ware (1992): Long-term variability
came at a furious pace in the last decade of the 20th Cen- in zooplankton biomass in the subarctic Pacific Ocean. Fish.
tury. More insights into how PDO works, and how to pre- Oceanogr., 1, 32–38.
dict PDO variations, are sure to follow throughout the Brodeur, R. D., B. W. Frost, S. R. Hare, R. C. Francis and W. J.
first decade of the 21st Century. Ingraham, Jr. (1996): Interannual variations in zooplankton
biomass in the Gulf of Alaska and covariation with Califor-
Acknowledgements nia Current zooplankton. Calif. Coop. Oceanic Fish. Invest.
We thank S. Minobe, R. Allan and an anonymous Rep., 37, 80–99.
reviewer for providing insightful comments and construc- Cayan, D. R., A. A. Kammerdiener, M. D. Dettinger, J. M.
Caprio and D. H. Peterson (2001): Changes in the onset of
tive suggestions for revising an early draft of this article.
spring in the western United States. Bull. Amer. Meteor. Soc.,
We are also indebted to R. Francis, E. Miles, J. M.
82, 399–415.
Wallace, and W. Wooster for lively discussions that have Chao, Y., M. Ghil and J. C. McWilliams (2000): Pacific
contributed to advancing the research into PDV and its interdecadal variability in this Century’s sea surface tem-
impacts on regional climate and natural resources, and T. peratures. Geophys. Res. Lett., 27, 2261–2264.
Mitchell for developing the mapping routines used to cre- Clark, W. G., S. R. Hare, A. M. Parma, P. J. Sullivan and R. J.
ate Fig. 3. This study was prompted by collaborations in Trumble (1999): Decadal changes in growth and recruit-
the University of Washington’s Climate Impacts Group, ment of Pacific halibut (Hippoglossus stenolepis). Can. J.
an interdisciplinary effort within the Joint Institute for Fish. Aquat. Sci., 56, 173–183.
the Study of the Atmosphere and Oceans and School of Dettinger, M. D., D. S. Battisti, R. D. Garreaud, G. J. McCabe,
Marine Affairs, and was funded under NOAA’s coopera- Jr. and C. M. Bitz (2000): Interhemispheric effects of
interannual and decadal ENSO-like climate variations on
tive agreement #NAI17RJ1232 and The Hayes Center.
the Americas. p. 1–16. In Present and Past Interhemispheric
This is JISAO contribution #862.
Climate Linkages in the Americas and their Societal Ef-
fects, ed. by V. Markgraf, Cambridge University Press. Cam-
References bridge, U.K.
Adkison, M. D., R. M. Peterman, M. F. Lapointe, D. M. Gillis Ebbesmeyer, C. C., D. R. Cayan, D. R. Milan, F. H. Nichols, D.
and J. Korman (1996): Alternative models of climatic ef- H. Peterson and K. T. Redmond (1991): 1976 step in the

42 N. J. Mantua and S. R. Hare

 Pacific climate: forty environmental changes between 1968– Hare, S. R. and R. C. Francis (1995): Climate change and salmon
1975 and 1977–1984. p. 115–126. In Proceedings of the production in the Northeast Pacific Ocean. p. 357–372. In
7th Annual Climate (PACLIM) Workshop, April 1990, ed. Climate Change and Northern Fish Populations, ed. by R.
by J. L. Betancourt and V. L. Tharp, California Department J. Beamish, Can. Spec. Publ. Fish. Aquat. Sci. 121.
of Water Resources, Interagency Ecological Studies Pro- Hare, S. R. and N. J. Mantua (2000): Empirical evidence for
gram Technical Report 26. North Pacific regime shifts in 1977 and 1989. Prog.
Enfield, D. B. and A. M. Mestas-Nuñez (1999): Multiscale Oceanogr., 47, 103–146.
variabilities in global sea surface temperatures and their Hare, S. R., N. J. Mantua and R. C. Francis (1999): Inverse
relationships with tropospheric climate patterns. J. Climate, production regimes: Alaskan and West Coast Salmon. Fish-
12, 2719–2733. eries, 24, 6–14.
Evans, M. N., A. Kaplan, M. A. Cane and R. Villalba (2000): Hollowed, A. B., S. R. Hare and W. S. Wooster (1998): Pacific-
Globality and optimality in climate field reconstructions Basin climate variability and patterns of Northeast Pacific
from proxy data. p. 53–72. In Present and Past Interhemi- marine fish production. p. 89–104. In Proceedings of the
spheric Climate Linkages in the Americas and Their Societal 10th ‘Aha Huliko’ a Hawaiian Winter Workshop on Biotic
Effects, ed. by V. Markgraf, Cambridge University Press, Impacts of Extratropical Climate Variability in the Pacific,
Cambridge, U.K. ed. by G. Holloway, P. Muller and D. Henderson. NOAA
Evans, M. N., M. A. Cane, D. P. Schrag, A. Kaplan, B. K. Award No. NA67RJ0154, SOEST Special Publication.
Linsley, R. Villalba and G. M. Wellington (2001): Support Kaplan, A., Y. Kushnir and M. A. Cane (2000): Reduced Space
for tropically-driven Pacific decadal variability based on Optimal Interpolation of historical marine sea level pres-
paleoproxy evidence. Geophys. Res. Lett., 28, 3689–3692. sure: 1854–1992. J. Climate, 13, 2987–3002.
Folland, C. K., D. E. Parker, A. Colman and R. Washington Kawasaki, T. (1991): Long-term variability in the pelagic fish
(1999): Large scale modes of ocean surface temperature populations. p 47–60. In Long-Term Variability of Pelagic
since the late nineteenth century. p. 73–102. In Beyond El Fish Populations and Their Environment, ed. by T.
Niño: Decadal and Interdecadal Climate Variability, ed. by Kawasaki, S. Tanaka, Y. Toba and A. Taniguchi, Pergamon
A. Navarra, Springer-Verlag, Berlin, 374 pp. Press, New York.
Francis, R. C. and S. R. Hare (1994): Decadal-scale regime Latif, M. and T. P. Barnett (1994): Causes of decadal climate
shifts in the large marine ecosystems of the North-east Pa- variability over the north Pacific and North America. Sci-
cific: a case for historical science. Fish. Oceanogr., 3, 279– ence, 266, 634–637.
291. Latif, M. and T. P. Barnett (1996): Decadal climate variability
Francis, R. C. and S. R. Hare (1997): Regime scale climate over the North Pacific and North America: dynamics and
forcing of salmon populations in the Northeast Pacific— predictability. J. Climate, 9, 2407–2423.
some new thoughts and findings. p. 113–128. In Estuarine Linsley, B. K., G. M. Wellington and D. P. Schrag (2000):
and Ocean Survival of Northeastern Pacific Salmon: Pro- Decadal sea surface temperature variability in the subtropi-
ceedings of the Workshop, ed. by R. L. Emmett and M. H. cal South Pacific from 1726 to 1997 A.D. Science, 290,
Schiewe, U.S. Dep. Commer., NOAA Tech. Memo. NMFS- 1145–1148.
NWFSC-29. Mackas, D. L., R. Goldblatt and A. G. Lewis (1998):
Francis, R. C., S. R. Hare, A. B. Hollowed and W. S. Wooster Interdecadal variation in developmental timing of
(1998): Effects of interdecadal climate variability on the Neocalanus plumchrus populations at Ocean Station P in
oceanic ecosystems of the NE Pacific. Fish. Oceanogr., 7, the subarctic North Pacific. Can. J. Fish. Aquat. Sci., 55,
1–21. 1878–1893.
Fritts, H. C. (1991): Reconstructing Large-Scale Climatic Pat- Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace and R. C.
terns from Tree-Ring Data. The University of Arizona Press, Francis (1997): A Pacific interdecadal climate oscillation
Tucson & London, 286 pp. with impacts on salmon production. Bull. Amer. Meteor.
Garreaud, R. D. and D. S. Battisti (1999): Interannual ENSO Soc., 78, 1069–1079.
and interdecadal ENSO-like variability in the Southern McFarlane, G. and R. J. Beamish (1992): Climatic influence
Hemisphere tropospheric circulation. J. Climate, 12, 2113– linking copepod production with strong year-classes in
2123. sablefish, Anaplopoma fimbria. Can. J. Fish. Aquat. Sci.,
Gedalof, Z. and D. J. Smith (2001): Interdecadal climate vari- 49, 743–753.
ability and regime-scale shifts in Pacific North America. McGowan, J. A., D. R. Cayan and L. M. Dorman (1998): Cli-
Geophys. Res. Lett., 28, 1515–1518. mate-ocean variability and ecosystem response in the North-
Gershunov, A. and T. P. Barnett (1998): Interdecadal modula- east Pacific. Science, 281, 210–217.
tion of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79, Miller, A. J. and N. Schneider (2000): Interdecadal climate re-
2715–2726. gime dynamics in the North Pacific Ocean: theories, obser-
Graham, N. E. (1994): Decadal-scale climate variability in the vations and ecosystem impacts. Prog. Oceanogr., 47, 355–
1970s and 1980s: observations and model results. Clim. Dyn. 379.
10, 135–159. Miller, A. J., D. R. Cayan, T. P. Barnett, N. E. Graham and J.
Hare, S. R. (1996): Low frequency climate variability and M. Oberhuber (1994): The 1976–77 climate shift of the
salmon production. Ph.D. Dissertation. Univ. of Washing- Pacific Ocean. Oceanography, 7, 21–26.
ton, Seattle, WA, 306 pp. Minobe, S. (1997): A 50–70 year climatic oscillation over the

The Pacific Decadal Oscillation 43

 North Pacific and North America. Geophys. Res. Lett., 24, Schwing, F. and C. Moore (2000): A year without a summer for
683–686. California, or a harbinger of a climate shift? Eos Trans.,
Minobe, S. (1999): Resonance in bidecadal and pentadecadal AGU, 81, 301, 304–305.
climate oscillations over the North Pacific: Role in climatic Seager, R., Y. Kushnir, N. H. Naik, M. A. Cane and J. Miller
regime shifts. Geophys. Res. Lett., 26, 855–858. (2001): Wind-driven shifts in the latitude of the Kuroshio-
Minobe, S. (2000): Spatio-temporal structure of the Oyashio Extension and generation of SST anomalies on
pentadecadal variability over the North Pacific. Prog. decadal time scales. J. Climate, 14, 4249–4265.
Oceanogr., 47, 381–408. Tanimoto, Y., N. Iwasaka, K. Hanawa and Y. Toba (1993): Char-
Nakamura, H., G. Lin and T. Yamagata (1997): Decadal cli- acteristic variations of sea surface temperatures with mul-
mate variability in the North Pacific during the recent dec- tiple time scales in the North Pacific. J. Climate, 6, 1153–
ades. Bull. Amer. Meteor. Soc., 78, 2215–2225. 1160.
National Research Council (NRC) (1996): Learning to Predict Tourre, Y. M., B. Rajagopalan, Y. Kushnir, M. Barlow and W.
Climate Variations Associated with El Niño and the South- B. White (2001): Patterns of coherent decadal and
ern Oscillation. National Academy Press, Washington, D.C., interdecadal climate signals in the Pacific Basin during the
171 pp. 20th Century. Geophys. Res. Lett., 28, 2069–2072.
National Research Council (NRC) (1998): Decade-to-Century Trenberth, K. E. (1990): Recent observed interdecadal climate
Scale Climate Variability and Change: A Science Strategy. changes in the northern hemisphere. Bull. Amer. Meteor.
National Academy Press, Washington, D.C., 141 pp. Soc., 71, 988–993.
(http://www.nap.edu). Trenberth, K. E. and J. W. Hurrell (1994): Decadal atmosphere-
Nitta, T. and S. Yamada (1989): Recent warming of tropical ocean variations in the Pacific. Clim. Dyn., 9, 303–319.
sea surface temperature and its relationship to the Northern Trenberth, K. E. and D. J. Shea (1987): On the evolution of the
Hemisphere circulation. J. Meteor. Soc. Japan, 67, 375–383. Southern Oscillation. Mon. Wea. Rev., 115, 3078–3096.
Peterman, R. M., B. J. Pyper, M. F. Lapointe, M. D. Adkison Vandenbosch, R. (2000): Effects of ENSO and PDO events on
and C. J. Walters (1998): Patterns of covariation in survival seabird populations as revealed by Christmas bird count
rates of British Columbia and Alaskan sockeye salmon data. Waterbirds, 23, 416–422.
(Oncorhynchus nerka) stocks. Can. J. Fish. Aquat. Sci., 55, Venrick, E. L., J. A. McGowan, D. R. Cayan and T. L. Hayward
2503–2517. (1987): Climate and chlorophyll a: long-term trends in the
Piatt, J. F. and P. Anderson (1995): Response of common murres central north Pacific Ocean. Science, 238, 70–72.
to the Exxon Valdez oil spill and long-term changes in the Wallace, J. M. and D. S. Gutzler (1981): Teleconnections in
Gulf of Alaska marine ecosystem. p. 720–737. In Exxon the geopotential height field during the Northern Hemi-
Valdez Oil Spill Symposium Proceedings, ed. by S. D. Rice, sphere winter. Mon. Wea. Rev., 109, 784–812.
R. B. Spies, D. A. Wolfe and B. A. Wright, AFS Symp. No. White, W. B. and D. R. Cayan (1998): Quasi-periodicity and
18. global symmetries in interdecadal upper ocean temperature
Power, S., F. Tseitkin, S. Torok, B. Lavery, R. Dahni and B. variability. J. Geophys. Res., 103, 21,335–21,354.
McAvaney (1997): Australian temperature, Australian rain- Willmott, C. J. and K. Matsuura (2000): Terrestrial Air Tem-
fall and the Southern Oscillation, 1910–1992: Coherent perature and Precipitation: Monthly and Annual Time Se-
variability and recent changes. Austr. Meteorol. Mag., 47, ries (1950–1996) (WWW URL: http://
85–101. climate.geog.udel.edu/~climate/html_pages/
Power, S., F. Tseitkin, V. Mehta, B. Lavery, S. Torok and N. README.ghcn_ts.html).
Holbrook (1999a): Decadal climate variability in Australia Willmott, C. J. and S. M. Robeson (1995): Climatologically-
during the twentieth century. Int. J. Climatol., 19, 169–184. aided interpolation (CAI) of terrestrial air temperature. Int.
Power, S., T. Casey, C. Folland, A. Colman and V. Mehta J. Climatol., 15, 221–229.
(1999b): Inter-decadal modulation of the impact of ENSO Yasuda, I., H. Sugusaki, Y. Watanabe, S. Minobe and Y. Oo-
on Australia. Clim. Dyn., 15, 319–324. zeki (1999): Interdecadal variations in Japanese sardine and
Rasmussen, E. M. and J. M. Wallace (1983): Meteorological ocean/climate. Fish. Oceanogr., 8, 18–24.
aspects of the El Niño/Southern Oscillation. Science, 222, Zhang, Y. (1996): An observational study of atmosphere-ocean
1195–1202. interaction in the northern oceans on interannual and
Roemmich, D. and J. McGowan (1995): Climatic warming and interdecadal time-scales. Ph.D. Dissertation, Univ. of Wash-
the decline of zooplankton in the California Current. Sci- ington, Seattle, WA, 162 pp.
ence, 267, 1324–1326. Zhang, Y., J. M. Wallace and D. S. Battisti (1997): ENSO-like
Schneider, N. and A. J. Miller (2001): Predicting western North interdecadal variability. J. Climate, 10, 1004–1020.
Pacific ocean climate. J. Climate, 14, 3997–4002.

44 N. J. Mantua and S. R. Hare

View publication stats