Abstract

Recent studies indicate that correlations between ENSO and winter precipitation in the southwestern United States may vary with the phase of the Pacific decadal oscillation (PDO). In-phase relationships between ENSO and the PDO strengthen the impact of ENSO on winter precipitation, while out-of-phase relationships weaken this impact. It has been suggested that this knowledge of PDO phase can improve long lead winter forecasts. However, all of these studies have focused on the impact of the PDO on both El Niño and La Niña only. Years of neutral ENSO have been neglected even though neutral years occur roughly half the time and coincide with highly variable winter precipitation. It is expected that some of this variability may be caused by the phase of the PDO, although the extent of its relationship with ENSO is not well understood. When years of neutral ENSO from 1925 to 1998 are split by PDO phase, it is found that Arizona winter precipitation and its predictability are strongly influenced. Years of neutral ENSO/cold PDO are nearly as dry as years of La Niña/cold PDO and provide improved predictability for dry winters. While years of neutral ENSO/warm PDO are associated with wet winters and have above-average precipitation, it is of a lesser magnitude than years of El Niño/warm PDO. Because years of neutral ENSO and La Niña are similarly dry and together compose nearly three-quarters of the multidecadal cold phase of the PDO, an increased risk of extended drought for Arizona is implied upon the return of cold PDO.

1. Introduction

Knowledge of the pattern of late-summer sea surface temperatures (SSTs) relating to the El Niño–Southern Oscillation (ENSO) has long been an important tool to forecasters generating long-lead forecasts for winter precipitation in the southwestern United States. Recent discoveries of multidecadal variability of the impact of ENSO on winter precipitation due to the phase of the Pacific decadal oscillation (PDO) have the potential to improve these long-lead forecasts even more (Gershunov and Barnett 1998; McCabe and Dettinger 1999, 2002). While the dynamic mechanisms behind the PDO and the extent of its relationship with ENSO are not well understood, Miller and Schneider (2000) offer several possibilities in an in-depth review of the literature. More recently, Gedalof et al. (2002) and Newman et al. (2003) have suggested that ENSO may be the forcing mechanism for the PDO.

Gershunov and Barnett (1998) were the first to show the “modulation effect” of the PDO on ENSO winter precipitation patterns. When both are in the same phase (El Niño/warm PDO; La Niña/cold PDO), the relationship is constructive in nature and creates a stronger, more stable ENSO climate signal with regard to winter precipitation in the western United States. Out-of-phase relationships between ENSO and the PDO, with each in a different phase (El Niño/cold PDO; La Niña/warm PDO), are destructive in nature and weaken this signal. Gutzler et al. (2002) conducted an analysis for the southwestern United States that examined how winter precipitation patterns during high-index ENSO are impacted by PDO phase. Not considered in that study is how precipitation patterns during low index, or periods of neutral ENSO, are impacted by PDO phase. This becomes relevant when it is considered that years of neutral ENSO occur roughly half the time using many definitions for determining ENSO years.

The goal of this study is to quantify and compare the differences in the strength of the modulation effect of the PDO on Arizona winter precipitation during years of neutral ENSO to years of El Niño and La Niña in Arizona. Several statistical techniques will be used to analyze how winter precipitation during neutral-ENSO conditions varies according to PDO phase. Because the second half of this study is similar in scope to that of Gutzler et al. (2002), much of the same data and methodology are selected. Arizona is chosen as the study location in this preliminary paper because of its proven high correlation between winter precipitation and the phase of ENSO (Redmond and Koch 1991).

2. Data and methodology

a. ENSO–PDO time series

SST data from the Niño-3 region (5°N–5°S, 90°– 150°W) are used to represent ENSO. The Japan Meteorological Agency (JMA) determination of El Niño (La Niña) is made when the 5-month running mean of SST anomalies (relative to a base period climatology of 1950–79) in the region are ≥0.5°C (≤0.5°C) for at least 6 consecutive months (Trenberth 1997). This method produces 17 yr of La Niña, 21 yr of El Niño, and 36 yr of neutral ENSO from 1925 to 1998. As pointed out by Trenberth (1997), this method corresponds with what have historically been considered El Niño and La Niña events. The dataset used in this study is the Kaplan extended Niño-3 dataset (Kaplan et al. 1998) and was obtained from the International Research Institute for Climate Prediction (IRI) data library (available online at http://iridl.ldeo.columbia.edu/SOURCES/.Indices/.nino/.EXTENDED/.NINO3/;).

The PDO characterizes low-frequency changes in SSTs in the Pacific Ocean with a phasing of 20–30 yr. The PDO index is the leading principal component or eigenvector of the mean November–March monthly SSTs in the Pacific Ocean north of 20°N latitude (Mantua et al. 1997). Positive values of the index refer to above-normal SSTs along the west coast of North America and along the equator and below-normal SSTs in the central and western North Pacific around 45°N latitude. Negative values of the index refer to the opposite distribution of SSTs in these same areas. When viewed in terms of the modulating effect of the PDO, actual annual or seasonal index values of the PDO are not used. Instead, multidecadal warm and cold phases are used to delineate changes in the PDO (Gershunov and Barnett 1998; Gutzler et al. 2002). According to Mantua and Hare (2002), there have only been three complete phases of the PDO since 1925 (warm 1925–46 and 1977–98, cold 1947–76). While some researchers have speculated that the PDO may have shifted to a cold phase beginning in 1999 (Hare and Mantua 2000; Schwing and Moore 2000), most scientists agree that this may not be known for several years (Mantua and Hare 2002; Chavez et al. 2003). For this reason, the analysis will consist of the years 1925–98, which represent the three complete phases outlined above. The PDO dataset used in this study was obtained from the Joint Institute for the Study of Atmosphere and Oceans (JISAO) at the University of Washington (available online at http://jisao.washington.edu/pdo/PDO.latest).

b. Precipitation data

This study will consider winter in Arizona to occur from December through March (DJFM). While some investigators, including Diaz et al. (2001), have suggested that ENSO impacts are greatest for the December–January–February (DJF) period, correlations between winter precipitation and both antecedent July– August–September (JAS; 0.28–0.17) and concurrent (DJF; 0.39–0.27) Niño-3 SST anomalies are more robust for DJFM compared to DJF. The precipitation time series used in this study was obtained from the National Oceanic and Atmospheric Administration/National Climatic Data Center (NOAA/NCDC). Monthly DJFM precipitation was summed for the seven climate divisions of Arizona for each year from 1925 to 1998. Each climate division represents a simple unweighted average from all representative stations within that division (Guttman and Quayle 1996). Gutzler et al. (2002) deal with inhomogeneity issues regarding time-varying distributions of station data and reconcile this by showing that selected individual stations in the Southwest demonstrate the same characteristics as their respective climate divisions with regard to long-term trends in winter precipitation.

Because arid environments such as Arizona have positively skewed precipitation distributions, a square-root transformation was applied to each climate division to normalize the data into an approximately Gaussian distribution. Normalizing the precipitation data in this way reduced skewness (1.14 to 0.51) and kurtosis (1.55 to −0.09) to the point where no statistically significant deviation from a normal distribution is determined at a 0.99 confidence level.

3. Methodology and results

a. Percent of normal

To quantify differences in the strength of the modulation effect of the PDO during years of neutral ENSO versus that of El Niño and La Niña, each ENSO phase was stratified by PDO phase. The percent of raw normal winter precipitation for these six subcategories of the period of study is shown in Fig. 1. A two-tailed Student's t test for the difference of means between each ENSO phase split by warm and cold PDO using normalized winter precipitation was calculated, and the significance for each ENSO phase is shown by gray shading (Fig. 1). The modulation effects for El Niño (Fig. 1a) and La Niña (Fig. 1c) support results from previous studies (Gershunov and Barnett 1998) and are included only for comparison with the modulation of neutral ENSO (Fig. 1b). Only the modulation of neutral-ENSO years by PDO phase approaches statistical significance, with three climate divisions significant at a α = 0.05 level. Also notable is that the decrease in precipitation associated with La Niña and neutral-ENSO years during the cold PDO are of similar magnitude.

Fig. 1.

Percent of normal DJFM precipitation for each AZ climate division during each ENSO phase split by warm and cold PDO phases from 1925 to 1998 (warm PDO in italics). Gray shading denotes significance (α value) from the difference of means using a two-tailed Student's t test (using normalized DJFM precipitation): α ≤ 0.05, dark shaded; 0.05 < α ≤ 0.10, light shaded; α > 0.10 not shaded

Fig. 1.

Percent of normal DJFM precipitation for each AZ climate division during each ENSO phase split by warm and cold PDO phases from 1925 to 1998 (warm PDO in italics). Gray shading denotes significance (α value) from the difference of means using a two-tailed Student's t test (using normalized DJFM precipitation): α ≤ 0.05, dark shaded; 0.05 < α ≤ 0.10, light shaded; α > 0.10 not shaded

b. Extreme winters

Of importance to watershed managers and other is the frequency of winters with extreme precipitation. The extreme winters are defined when normalized DJFM precipitation is one standard deviation above or below the normalized historical mean for each climate division (Maxwell and Holbrook 2002). Winters of two standard deviations above and below the mean were also analyzed, although N is small for ±2σ results to have significant meaning. Maxwell and Holbrook (2002) show that in-phase relationships between ENSO and the PDO are much more likely to coincide with winters of extreme precipitation than are out-of-phase relationships. Extremely wet winters are associated with El Niño/ warm PDO, while extremely dry winters are associated with La Niña/cold PDO.

In this study, the frequency of extreme winters during years of neutral ENSO is similar for each phase of the PDO. However, whether the extreme winter is wet or dry is strongly determined by PDO phase (Table 1). Dry winters during neutral ENSO/cold PDO occur with roughly the same frequency as during La Niña/cold PDO. Extremely dry winters occur on average every 3.3 neutral ENSO/cold PDO years and every 3.0 La Niña/cold PDO years. The relationship between extremely wet and dry winters and neutral ENSO/warm PDO is not as robust. Extremely wet (dry) winters occur on average every 5.1 (13.4) neutral ENSO/warm PDO years.

Table 1. 

Return interval (yr) of extreme winters of precipitation for selected ENSO–PDO relationships (extreme winters are 1 and 2 σ above and below mean). Return intervals greater than 25 yr are neglected

Return interval (yr) of extreme winters of precipitation for selected ENSO–PDO relationships (extreme winters are 1 and 2 σ above and below mean). Return intervals greater than 25 yr are neglected
Return interval (yr) of extreme winters of precipitation for selected ENSO–PDO relationships (extreme winters are 1 and 2 σ above and below mean). Return intervals greater than 25 yr are neglected

c. Forecasting skill scores

One way to determine the skill of a teleconnection in predicting winter precipitation for a single location over many years is to use a modified Brier score (Gutzler et al. 2002). Because terciles are used, the dataset for this section of the study is limited to 1927–98 to provide terciles of 24 yr. Gutzler et al. (2002) show that in-phase relationships between Niño-3 and PDO improve skill scores, while out-of-phase relationships decrease skill scores. In order to examine how precipitation patterns following periods of late-summer (JAS) neutral ENSO are impacted by PDO phase, middle-tercile (neutral) Niño-3 anomalies were used to predict low-tercile (dry) winters. While operational forecasts include dynamic predictions of ENSO, persistence of late-summer SSTs is a useful predictor of ENSO through winter, as the seasonal lag correlation between JAS SSTs and subsequent DJF SSTs is 0.86 (Gutzler et al. 2002).

Maps showing skill scores for each climate division predicting dry winters following late-summer neutral ENSO conditions for each PDO phase are shown in Fig. 2. While the fact that dry winters are more likely to follow late-summer neutral-ENSO conditions during the cold phase of the PDO is expected (Fig. 2b), the degree to which this occurs is unexpected. These results indicate that on average across Arizona, dry winters follow late-summer neutral-ENSO conditions roughly 70% of the time (5 out of 7 yr). This is in contrast to the results from Gutzler et al. (2002) that determined that dry winters follow late-summer cold-ENSO (La Niña) conditions roughly 45% of the time (5 out of 11 yr).

Fig. 2.

Skill score map for AZ. (a) Map of skill scores, S-, for predictions of dry DJFM precipitation anomalies following neutral JAS Niño-3 anomalies during warm-phase PDO from 1927 to 1998. (b) Same as (a) but during cold-phase PDO. For both (a) and (b), gray shading corresponds with predictive significance: S- ≥ 0.75 not useful (blank); S- < 0.75 useful (shaded)

Fig. 2.

Skill score map for AZ. (a) Map of skill scores, S-, for predictions of dry DJFM precipitation anomalies following neutral JAS Niño-3 anomalies during warm-phase PDO from 1927 to 1998. (b) Same as (a) but during cold-phase PDO. For both (a) and (b), gray shading corresponds with predictive significance: S- ≥ 0.75 not useful (blank); S- < 0.75 useful (shaded)

4. Conclusions

The goal of this study has been to quantify how neutral-ENSO impacts on precipitation vary by phase of the PDO for Arizona. Using two different methods for classifying neutral ENSO, it was found that late-summer neutral-ENSO conditions precede dry winters nearly all of the time during the cold phase of the PDO. Neutral-ENSO years that occur during the cold phase of the PDO are shown to have a similar association with decreased precipitation and the frequency of extreme winters as do La Niña years during cold PDO. The association between wet winters and neutral-ENSO years that occur during the warm phase of the PDO is not robust. It is worth considering that 1999 and 2002, which were not included in the study, both had late-summer neutral-ENSO conditions and rank as the two driest winters in Arizona since 1925. A future project will involve a more in-depth look at this problem covering the western United States.

Acknowledgments

I greatly appreciate the comments provided by three anonymous reviewers in the revision of this manuscript.

REFERENCES

REFERENCES
Chavez
,
F. P.
,
J.
Ryan
,
S. E.
Lluch-Cota
, and
M.
Niquen
,
2003
:
From anchovies to sardines and back: Multidecadal change in the Pacific Ocean.
Science
,
299
,
217
221
.
Diaz
,
H. F.
,
M. P.
Hoerling
, and
J. K.
Eischeid
,
2001
:
ENSO variability, teleconnections and climate change.
Int. J. Climatol
,
21
,
1845
1862
.
Gedalof
,
Z.
,
N. J.
Mantua
, and
D. L.
Peterson
,
2002
:
A multi-century perspective of variability in the Pacific Decadal Oscillation: New insights from tree rings and coral.
Geophys. Res. Lett., 29, 2204, doi:10.1029/2002GL015824
.
Gershunov
,
A.
, and
T. P.
Barnett
,
1998
:
Interdecadal modulation of ENSO teleconnections.
Bull. Amer. Meteor. Soc
,
79
,
2715
2725
.
Guttmann
,
N. B.
, and
R. G.
Quayle
,
1996
:
A historical perspective of U.S. climate divisions.
Bull. Amer. Meteor. Soc
,
77
,
293
303
.
Gutzler
,
D. S.
,
D. M.
Kann
, and
C.
Thornbrugh
,
2002
:
Modulation of ENSO-based long-lead outlooks of Southwestern U.S. winter precipitation by the Pacific decadal oscillation.
Wea. Forecasting
,
17
,
1163
1172
.
Hare
,
S. R.
, and
N. J.
Mantua
,
2000
:
Empirical evidence for North Pacific regime shifts in 1977 and 1989.
Progress in Oceanography, Vol. 47, Pergamon, 103–146
.
Kaplan
,
A.
,
M.
Cane
,
Y.
Kushnir
,
A.
Clement
,
M.
Blumenthal
, and
B.
Rajagopalan
,
1998
:
Analyses of global sea surface temperature 1856–1991.
J. Geophys. Res
,
103
,
18567
18589
.
Mantua
,
N. J.
, and
S. R.
Hare
,
2002
:
The Pacific decadal oscillation.
J. Oceanogr
,
58
,
35
44
.
Mantua
,
N. J.
,
S. R.
Hare
,
Y.
Zhang
,
J. M.
Wallace
, and
R. C.
Francis
,
1997
:
A Pacific interdecadal climate oscillation with impacts on salmon production.
Bull. Amer. Meteor. Soc
,
78
,
1069
1079
.
Maxwell
,
K. D.
, and
V. P.
Holbrook
,
cited 2002
:
Pacific Decadal Oscillation and Arizona Precipitation.
Western Regional Tech. Attachment 02–08. [Available online at http://www.wrh.noaa.gov/wrhq/02TAs/0208/;.]
.
McCabe
,
G. J.
, and
M. D.
Dettinger
,
1999
:
Decadal variations in the strength of ENSO teleconnections with precipitation in the western United States.
Int. J. Climatol
,
19
,
1399
1410
.
McCabe
,
G. J.
, and
M. D.
Dettinger
,
2002
:
Primary modes and predictability of year-to-year snowpack variations in the western United States from teleconnections with Pacific Ocean climate.
J. Hydrometeor
,
3
,
13
25
.
Miller
,
A. J.
, and
N.
Schneider
,
2000
:
Interdecadal climate regime dynamics in the North Pacific Ocean: Theories, observations and ecosystem impacts.
Progress in Oceanography, Vol. 47, Pergamon, 355–379
.
Newman
,
M.
,
G. P.
Compo
, and
M. A.
Alexander
,
2003
:
ENSO-forced variability of the Pacific decadal oscillation.
J. Climate
,
16
,
3853
3857
.
Redmond
,
K. T.
, and
R. W.
Koch
,
1991
:
Surface climate and streamflow variability in the western United States and their relationship to large-scale circulation indices.
Water Resour. Res
,
27
,
2381
2399
.
Schwing
,
F.
, and
C.
Moore
,
2000
:
A year without a summer for California, or a harbinger of a climate shift?
Eos, Trans. Amer. Geophys. Union
,
81
,
301,
.
304
305
.
Trenberth
,
K. E.
,
1997
:
The definition of El Niño.
Bull. Amer. Meteor. Soc
,
78
,
2771
2777
.

Footnotes

Corresponding author address: Gregory B. Goodrich, Department of Geography, Arizona State University, P.O. Box 870104, Tempe, AZ 85287-0104. Email: gregory.goodrich@asu.edu