the relationship of the north american monsoon to tropical ... · the north american monsoon is a...

VOLUME 14 15 DECEMBER 2001J O U R N A L O F C L I M A T E

q 2001 American Meteorological Society 4449

The Relationship of the North American Monsoon to Tropical and North Pacific SeaSurface Temperatures as Revealed by Observational Analyses

CHRISTOPHER L. CASTRO, THOMAS B. MCKEE, AND ROGER A. PIELKE SR.

Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

(Manuscript received 18 April 2000, in final form 27 June 2001)

ABSTRACT

The North American monsoon is a seasonal shift of upper- and low-level pressure and wind patterns thatbrings summertime moisture into the southwest United States and ends the late spring wet period in the GreatPlains. The interannual variability of the North American monsoon is examined using the NCEP–NCAR reanalysis(1948–98). The diurnal and seasonal evolution of 500-mb geopotential height, integrated moisture flux, andintegrated moisture flux convergence are constructed using a 5-day running mean for the months May throughSeptember. All of the years are used to calculate an average daily Z score that removes the diurnal, seasonal,and intraseasonal variability. The 30-day average Z score centered about the date is correlated with Pacific seasurface temperature anomaly (SSTA) indices associated with the El Nino–Southern Oscillation (ENSO) and theNorth Pacific oscillation (NPO). These indices are Nino-3, a North Pacific index, and a Pacific index that combinesthe previous two. Regional time-evolving precipitation indices for the Southwest and Great Plains, which considerthe total number of wet or dry stations in a region, are also correlated with the SSTA indices. The use ofnonnormally distributed point source precipitation data is avoided.

Teleconnections are computed relative to the climatological evolution of the North American monsoon, ratherthan to calendar months, thus more accurately accounting for the climatological changes in the large-scalecirculation. Tropical and North Pacific SSTs are related to the occurrence of the Pacific Transition and EastPacific teleconnection patterns, respectively, in June and July. A high (low) NPO phase and El Nino (La Nina)conditions favor a weaker (stronger) and southward (northward) displaced monsoon ridge. These teleconnectionpatterns affect the timing and large-scale distribution of monsoon moisture. In the Great Plains, the spring wetseason is lengthened (shortened) and early summer rainfall and integrated moisture flux convergence are above(below) average. In the Southwest, monsoon onset is late (early) and early summer rainfall and integratedmoisture flux convergence are below (above) average. Relationships with Pacific SSTA indices decay in thelater part of the monsoon coincident with weakening of the jet stream across the Pacific and strengthening ofthe monsoon ridge over North America. The most coherent summer climate patterns occur over the entire westernUnited States when the Pacific index is substantially high or low, such as during the Midwest flood of 1993and drought of 1988. The Pacific index in spring is a good predictor of early summer height anomalies overthe western United States when the time evolution of the North Pacific SST dipole is considered.

1. Introduction

The western United States is particularly sensitive tointerannual variability of summer climate. In the mostarid regions of the southwest United States, this vari-ability can be larger than the mean summer rainfall itself(Higgins et al. 1998). Climate and weather extremesproduce unique dangers and their effects vary in scope.Severe weather hazards include high winds, hail, light-ning, and tornadoes. Flash flooding arises from indi-vidual localized summer thunderstorms with high rain-fall rates. These events are common in the Southwestbecause of its steep terrain and poor soil moisture hold-ing capacity. Sustained summer flood or drought con-

Corresponding author address: Christopher L. Castro, Departmentof Atmospheric Science, Colorado State University, Fort Collins, CO80523.E-mail: [email protected]

ditions over broad areas, such as the flood of 1993 orthe drought of 1988, result from shifts in the large-scalecirculation pattern (e.g., Trenberth and Branstator 1992;Bell and Janowiak 1995; Trenberth and Guillemot1996). Short- or long-term departures from average con-ditions may adversely affect infrastructure, agriculturalproduction, water supply, and hydroelectric power gen-eration (e.g., Meehl et al. 2000). The sensitivity to theseextreme conditions is likely to increase in the next cen-tury if the current trends of population growth and rapiddevelopment in urban areas continue. There is a criticalneed to understand the causes of interannual variabilityso seasonal forecasts can be improved.

The North American Monsoon System (NAMS) isintimately linked to the large-scale distribution of sum-mer moisture in the western United States. Dry mid-latitude westerly flow persists until the middle of June.Monsoon onset is associated with a shift in circulation

4450 VOLUME 14J O U R N A L O F C L I M A T E

in late June or early July, on average. An extension ofthe Bermuda high retreats west, initiating light easterlyflow at middle and upper levels (above 850 mb) overthe southwest United States (Bryson and Lowry 1955;Adams and Comrie 1997; Higgins et al. 1997b). A di-urnal surface heat low forms over southwest Arizonaand southern California. Two low-level jets (LLJs), theGreat Plains LLJ and the Baja LLJ, are active at nightand early morning to midday, respectively (Bonner1968; Douglas 1995; Higgins et al. 1997a).

There is a rapid increase in precipitation that beginsin southern Mexico in June, advances northward alongthe Sierra Madre mountains, and arrives into the south-west United States by early July (Douglas et al. 1993;Adams and Comrie 1997; Higgins et al. 1997a). Mois-ture for daily NAMS thunderstorms comes from theGulf of Mexico, Gulf of California, and the eastern Pa-cific to varying degrees depending on location (Brysonand Lowry 1955; Reitan 1957; Rasmussen 1967; Hales1972; Brenner 1974; Douglas et al. 1993; Schmitz andMullen 1996; Adams and Comrie 1997; Higgins et al.1997b). Most U.S. NAMS studies focus exclusively onthe Southwest because of its clear monsoon signature.However, large-scale circulation changes associatedwith the monsoon are linked with precipitation changesover other regions of North America, most notably adecrease of precipitation in the central United States(Tang and Rieter 1984; Okabe 1995; Higgins et al.1997b; Barlow et al. 1998). Given this high degree ofinterconnectedness for the North American summertimecirculation, the present study considers summer tele-connections for the United States and Mexico as awhole.

Interannual variability of winter climate in the west-ern United States has a well-established connection tothe El Nino–Southern Oscillation (ENSO). Numerousmodel and observational studies confirm that variationsin the location and strength of tropical convection pro-duce coherent teleconnection patterns in the large-scalecirculation over North America (e.g., Horel and Wallace1981; Livezey and Mo 1987). The recognition of theseENSO-associated teleconnections and their associatedclimate patterns has substantially improved the accuracyof long-range seasonal forecasts for the winter season.El Nino winters are typically dry and warm in the PacificNorthwest and wet in the Southwest. By contrast, LaNina winters are wet in the Pacific Northwest and dryin the Southwest (e.g., Ropelewski and Halpert 1986).

North Pacific SSTs are also important for winter andsummer climate in the western United States. Barlowet al. (2000) provide a brief summary of three basicmodes of North Pacific SSTs. The first mode is typicalof ENSO, with interannual variation in the tropical andextratropical Pacific (e.g., Nitta and Yamada 1989;Zhang et al. 1997). The second mode, referred by var-ious authors as the Pacific decadal oscillation (PDO) orNorth Pacific oscillation (NPO), is associated with aPacific SST dipole that varies on a decadal timescale

(e.g., Mantua et al. 1997; Minobe 1997; Gurshunov andBarnett 1998). A high (low) NPO phase is characterizedby a strong (weak) Aleutian low, a cold (warm) centralNorth Pacific, and a warm (cold) eastern North Pacific.The third mode, or North Pacific mode, is confined tothe far North Pacific and also varies on a decadal time-scale. While some modeling studies suggest that NorthPacific SSTs are primarily a result of ENSO-forced at-mospheric variability (e.g., Lau and Nath 1996), otherssuggest that North Pacific SST variability, particularlythe North Pacific mode, is independent of ENSO andmaintained by self-sustaining mechanisms (Nakamuraet al. 1997). As of yet, it is not clear what the causesof North Pacific SST variability are, and that is not thesubject of this paper. Using long-term sea level pressureand daily precipitation data, Gurshunov and Barnett(1998) showed that the NPO modulates the winterENSO response. ENSO-associated teleconnection pat-terns and climate anomalies are most likely when ElNino (La Nina) corresponds with its favored NPO phase.Destructive combinations of ENSO and NPO tend toweaken the effects of either mode of influence, resultingin a weak and incoherent climate response. The NPOwas in a high phase, on average, from the late 1970sthrough most of the 1990s, coincident with an increasein frequency of El Nino events.

Several conclusions can be drawn from the previousanalyses of North American summertime precipitation(Carleton et al. 1990; Harrington et al. 1992; Bunkerset al. 1996; Gutzler and Preston 1997; Mo et al. 1997;Ting and Wang 1997; Comrie and Glenn 1998; Higginset al. 1998, 1999; Higgins and Shi 2000). The monsoonridge position and strength guide the mid- and upper-level moisture transport. Two configurations of themonsoon ridge exist that can vary in frequency of oc-currence either intraseasonally or interannually. In theridge north configuration, a strong ridge is located overthe Great Plains. Upper-level easterlies carry moisturefrom the Gulf of Mexico south of the ridge. The mon-soon arrives early and is wet in the Southwest. TheGreat Plains are dry on the subsiding branch of theridge. In the ridge south configuration, a trough occursin the western United States and the monsoon ridge isweakened and located over northwest Mexico. West-erly upper-level winds over the western United Statesdirect moisture into the Great Plains. The Southwestis dry and monsoon onset is delayed. The relationshipof the monsoon ridge configurations to summer mois-ture transport, precipitation, and wet and dry summersin the central United States is also well established(Mo et al. 1997).

Summer precipitation in the western United States isrelated to the distribution of Pacific SSTs. The relation-ship is particularly strong in the Great Plains. Bunkerset al. (1996) found statistically significant above (below)average precipitation anomalies over North and SouthDakota during El Nino (La Nina) summers. Ting andWang (1997) correlated sea surface temperature anom-

15 DECEMBER 2001 4451C A S T R O E T A L .

TABLE 1. Frequenly used acronyms and symbols.

Acronymor symbol Meaning

COADSCNPENSOENPEP

Comprehensive Ocean Atmosphere DatasetCentral North Pacific region (268–368N, 1778E–1648W)El Nino–Southern OscillationEastern North Pacific region (358–508N, 1258–1508W)East Pacific pattern

ITCZLLJMFMFC

NAMSNPNPO

Intertropical convergence zoneLow–level jetIntegrated moisture flux (sigma levels 14–28)Integrated moisture flux convergence (sigma levels 14–28)North American Monsoon SystemNorth Pacific sea surface temperature indexNorth Pacific oscillation

PPDOPRPTrr2

Combined Pacific sea surface temperature anomaly indexPacific decadal oscillationRegional precipitation indexPacific Transition patternCorrelation coefficientExplained variance

SPISSTSSTAZx(t)

F500

Standardized precipitation indexSea surface temperatureNormalized sea surface temperature anomalyThirty-day average Z score of reanalysis variable x

centered on the date500-mb geopotential height

alies (SSTAs) in the Pacific with a precipitation indexof Great Plains stations. Using a singular value decom-position technique, they showed that there are twomodes of covariation between Pacific SSTs and centralU.S. precipitation. Their first mode is primarily a trop-ical SST mode and related to precipitation over thenorthern Great Plains (and the Midwest). Their secondmode resembles the NPO and is related to precipitationover the southern Great Plains. As a whole, the GreatPlains precipitation studies suggest that wet (dry) sum-mers in that region are associated with El Nino (La Nina)and high (low) NPO.

An opposite relationship between Pacific SSTs andNAMS precipitation may exist in the Southwest. Carle-ton et al. (1990) noted in wet monsoons that east PacificSSTs off the California coast are cooler than average insummer. In a study using 65 years of monthly precip-itation from stations across Arizona and New Mexico,Harrington et al. (1992) found different patterns in sum-mer precipitation for extremes of the Southern Oscil-lation, with El Nino (La Nina) favoring above-averageJuly precipitation in northeast New Mexico (west-cen-tral Arizona). Higgins et al. (1999) correlated June–September precipitation totals for three monsoon re-gions in the United States and Mexico with the SouthernOscillation Index (SOI). They found a statistically sig-nificant positive (negative) correlation between La Nina(El Nino) and total summer precipitation in southwestMexico. Higgins et al. (1998) showed that negative SSTanomalies in the eastern tropical Pacific during winterand spring are associated with wet monsoons in Arizonaand New Mexico. Higgins and Shi (2000) have relatedNAMS onset and precipitation in the southwest UnitedStates to decade-scale fluctuations in North Pacific SSTsassociated with the NPO.

In this study, the original hypothesis of Gurshunovand Barnett (1998) is extended to the summer season.We propose a large-scale framework for considering thesummer variability: variations in the NAMS caused bythe combination of ENSO and North Pacific variability.The detasets and methodology are discussed in sections2 and 3. Interannual variability of 500-mb geopotentialheight, atmospheric moisture, and a precipitation indexare related to Pacific SSTs in sections 4 and 5. Laggedrelationships with Pacific SSTs and Pacific SST evo-lution are evaluated in section 6. A discussion and sum-mary are presented in sections 7 and 8. Table 1 is a listof the paper’s frequently used acronyms and symbols.

2. Description of data

The updated National Centers for Environmental Pre-diction–National Center for Atmospheric Research(NCEP–NCAR) daily reanalysis is used to evaluate at-mospheric circulation and moisture in the western Unit-ed States. The reanalysis assimilation system is a mod-ified version of the medium-range forecast spectral mod-el with T62 resolution and 28 sigma levels. The NCEP–

NCAR reanalysis incorporates the widest possible arrayof data sources with advanced quality control and mon-itoring systems (Kalnay et al. 1996). The study of in-terannual climate variability is an intended use of long-term reanalyses and the NCEP–NCAR reanalysis hasalready been used for this purpose in previous work onthe NAMS (e.g., Higgins et al. 1998, 1999; Higgins andShi 2000). In this study we utilize the reanalysis specifichumidity (q), surface pressure (ps), and winds (v) forsigma levels 28–14 (below approximately 400 mb), andstandard pressure level geopotential heights (F) andwinds. These reanalysis variables are more reliable, ascompared to model parameterized variables like evap-oration or precipitation, because they are directly influ-enced by the original rawinsonde observations (Kalnayet al. 1996). Four times daily data (0000, 0600, 1200,and 1800 UTC) were obtained from the NCAR archivesfor the months May through September from 1948 to1998.

The Cooperative Summary of the Day (Coop) data,from the National Climate Data Center, has surface ob-servations for all past and present cooperative sitesthroughout the United States. Daily precipitation is re-corded as the precipitation reading 24 h ending at thetime of observation, read to hundredths of an inch. Traceamounts of precipitation are recorded. Missing precip-itation is also recorded or, at some stations, entiremonths are absent from the record. Daily Coop precip-itation data from 1950 to 1995 were obtained for 340Great Plains stations (378–458N, 978–1058W) and 231Southwest stations (318–398N, 1068–1168W). The pre-cipitation regions are shown in Fig. 1a.


FIG. 1. Regions used to define the (a) Great Plains and Southwestprecipitation and (b) Pacific SSTA indices. The ENP (lag) region isused only for lag correlation.

The Comprehensive Ocean Atmosphere Dataset(COADS) contains monthly average SST on a 28 latitudeby 28 longitude grid from 1854 onward. SST observa-tions are from ship reports and other in situ platforms.A quality control procedure eliminates outlying obser-vations if they fall outside a prescribed number of stan-dard deviations about the smoothed median at a specificlocation (Slutz et al. 1985). The normalized monthlySSTA from 1950 to 1997 were computed for two regionsof the North Pacific, the central North Pacific (CNP;268–368N, 1778E–1648W) and the eastern North Pacific(ENP; 358–508N, 1258–1508W). These Pacific regionscorrespond roughly to the SST dipole in the secondmode of Pacific SST found by Ting and Wang (1997)and are correlated (r2 . 0.25) with summer precipitationin the Great Plains. We also calculated the SSTAs fora modified ENP region (408–558N, 1808–1608W) for lagcorrelation analysis. The traditional Nino-3 normalizedSSTA index is used to define ENSO. The Nino indicesare readily accessible from the Climate Prediction Cen-

ter. The Pacific SSTA regions are shown in Fig. 1b. Wealso use the Reynolds and Smith (1994) optimally in-terpolated SST dataset, rather than COADS, to view thetime-evolving spatial patterns of SSTAs.

3. Analysis methods

Though precipitation is the principal variable of con-cern, analysis of precipitation throughout the westernUnited States is not straightforward, due to the generallypoor spatial and temporal resolution afforded by theavailable precipitation stations and the nonnormal sta-tistical characteristics of precipitation (e.g., Cowie andMcKee 1986). We include here an alternate but com-plementary approach: analysis of four-times daily re-analysis integrated moisture flux (MF) and moisture fluxconvergence (MFC). While these variables are also sub-ject to data problems, particularly in Mexico and theGulf of California (e.g., Barlow et al. 1998), the morelarge-scale nature of the moisture field relative to pre-cipitation (e.g., Higgins et al. 1997b; Schmitz and Mul-len 1996) and the spatial and temporal consistency en-couraged by the reanalysis procedures suggests thatmoisture flux may yield more robust results. At locationsthroughout the western United States, the daily statis-tical distribution of MF and MFC is near normal in thesummer season (not shown).

The MF and MFC for each analysis time are com-puted from reanalysis winds (v) and specific humidity(q) on sigma surfaces. These variables are directly re-lated to evaporation (E) and precipitation (P) throughthe water balance equation (Trenberth and Guillemot 1995;Schmitz and Mullen 1996; Higgins et al. 1997a):

s28psMF 5 (qv) ds (1)Egs14

s28psMFC 5 P 2 E 5 2 = · (qv) ds. (2)Egs14

The version of the water balance equation used here forMFC omits local storage of water vapor, surface runoff,and ground storage. The MF and MFC for sigma levels28–14 are computed by a stepwise integration proce-dure. Since the amount of water vapor at 400 mb is twoorders of magnitude less than at the surface, transportof water vapor at upper levels (less than sigma level14) is assumed negligible (e.g., Higgins et al. 1997b).

Five-day running averages (m) and standard devia-tions (s) of the reanalysis variables are constructed foreach day from May through September. Each of the fourdaily reanalysis times is considered separately. Thesefields yield an updated 50-yr reanalysis climatology ofthe height, winds, and moisture fields of the NAMS.This new climatology is similar to more comprehensiveNAMS climatologies by Schmitz and Mullen (1996) andHiggins et al. (1997b) based on shorter lengths of re-cord. Some important climatological features include


TABLE 2. Average summer (Jun/Jul/Aug–JJA) values of PacificSSTA indices (1950–97). Years of ENSO and North Pacific com-posites identified.

Year Nino-3 NP index P index Composite

195019511952195319541955

20.330.87

20.200.57

20.9320.63

1.262.68

21.1421.2321.0721.66

0.923.55

21.3420.6622.0022.30

NPENSONPENSOENSOENSO

19561957195819591960

20.381.270.36

20.3420.03

20.630.002.86

20.3120.90

21.011.263.23

20.6520.94

NPENSONP

NP19611962196319641965

20.2020.04

0.8720.87

1.27

21.520.881.12

21.320.90

21.720.842.00

22.192.17

NPNPENSOENSOENSO

19661967196819691970

0.0120.14

0.320.69

21.17

20.212.00

20.290.46

20.85

20.201.850.021.15

22.02

NP

ENSOENSO

19711972197319741975

20.451.66

20.980.03

20.72

21.390.79

21.9820.93

0.09

21.842.46

22.9620.8920.62

ENSOENSOENSONPENSO

19761977197819791980

0.860.06

20.470.300.27

21.830.26

20.202.130.36

20.960.32

20.682.430.63

ENSO

ENSONP

19811982198319841985

20.331.251.50

20.5020.57

0.530.64

21.1020.9920.30

0.201.890.40

21.4920.87

NPENSOENSOENSOENSO

19861987198819891990

0.201.69

21.6220.10

0.23

0.7821.3323.4720.25

0.08

0.980.35

25.0920.35

0.31

NPENSOENSO

1991199219931994199519961997

0.990.200.48

20.0220.1820.14

2.65

20.202.931.680.450.190.681.36

0.783.132.160.420.010.534.01

ENSONPENSONPNPNPENSO

the evolution of the monsoon ridge, the seasonal evo-lution of Great Plains and Baja LLJs, and the diurnaland seasonal evolution of integrated moisture flux con-vergence. Additional discussion of these aspects ofNAMS climatology computed for this work can befound in Castro et al. (2000).

In the same manner as the precipitation studies, thereanalysis variables (x) of MF, MFC, and F500 are con-verted to Z scores (Zx) for the four reanalysis times. Aspatially varying Z score is defined for each time as

x 2 m(t) x (t)Z 5 . (3)x (t) sx (t)

A time-varying correlation approach is used to relatePacific SSTs, atmospheric circulation, and atmosphericmoisture transport through the course of the summerseason. We define two simple SSTA indices. The NorthPacific (NP) index is

NP 5 ENP 2 CNP. (4)

A Pacific index (P) combines the NP and Nino-3 indices:

P 5 Nino-3 1 ENP 2 CNP. (5)

This new simple P index is proposed as a way to relatethe combination of temporal variability of tropical andNorth Pacific SSTs. High (low) P roughly correspondswith El Nino (La Nina) and high (low) NPO. The Pindex weights the influence of the North Pacific morethan tropical Pacific SSTs, reflecting the magnitude ofcorrelation between Pacific SSTs and Great Plains sum-mer precipitation found by Ting and Wang (1997).

To isolate the contributions to the explained varianceof reanalysis variables from tropical and North PacificSST, two composites of years are constructed. The ap-proach is similar to Gurshunov and Barnett (1998), ex-cept here we compost years by the summer SSTA in-dices and use a more liberal threshold of SSTA to in-crease the sample size. For the ENSO composite, theabsolute summer average value of Nino-3 is greater than0.4. For the North Pacific composite, the summer ab-solute value of Nino-3 is less than 0.4 and the summerabsolute value of the North Pacific index is greater than0.4. The North Pacific composite is revealing in that ityields the relationship between the NAMS and NorthPacific SSTs in ENSO-neutral years. Using these cri-teria, we obtain 25 years for the ENSO subset and 16years for the North Pacific subset (see Table 2). The 30-day average Z score of the reanalysis variable centeredon the date ( x(t)) is correlated with the time-coincidentZvalue of the Nino-3 index for the years in the ENSOcomposite and the NP index for the years of the NorthPacific composite.

To evaluate statistical significance, we use the fieldsignificance test of Livezey and Chen (1983) as describedin Wilks (1995). The effective degrees of freedom (n)for the data at each spatial point is determined as

Nn 5 , (6)

t

where N is the total number of days from 15 Maythrough 15 September (123) multiplied by the numberof years in used in the particular correlation analysis.The t value for each point is computed as

K

t 5 1 1 2 C (iDt)C (iDt)Dt, (7)O 1 2i51

where C1(iDt) is the autocorrelation of x(t) at lag iDt,ZC2(iDt) is the autocorrelation of the particular SSTA


FIG. 2. Average Jun–Aug values of (a) P index, (b) NP index, and(c) Nino-3 index from 1950 to 1997.

index at lag iDt, and Dt is in days. For the total numberof lags (K) we use 61 days. Though it could be argueda greater number of lags is necessary, autocorrelationsof x(t) at 30 days aside of the date approach zero overZthe midlatitude Pacific and North America. The statis-tically significant correlation coefficient (rsig) is then

Xr 5 . (8)sig Ïn 2 3

Here X is the absolute value of the desired confidenceinterval on a normal distribution. To satisfy statisticalsignificance at the 95% level, X 5 1.96.

To evaluate the relationship of Coop precipitation datato Pacific SSTs, we introduce regional time-evolvingprecipitation indices for the Great Plains and the South-west. At an individual station, for each day a 30-dayprecipitation total is computed centered about the date.A missing 30-day total for the station is recorded if thereare more than five missing days in the record. For eachdate the years are ordered highest to lowest accordingto the precipitation totals. The station precipitation totalis considered above (below) average on that date if agiven year ranks in the top (bottom) 20 years. The num-ber of stations in the region (S) recording an above-average (Swet) or below-average (Sdry) precipitation totalare tallied for each date. The regional precipitation indexis then

S (t) 2 S (t)wet dryPR(t) 5 . (9)

S

We correlated the precipitation indices with the threePacific SST indices throughout the entire summer, usingall the years of record.

Considering summer precipitation in such a way hasseveral advantages. The use of point source precipita-tion, which is not normally distributed, is avoided. Aswith the average Z score correlation analyses for re-analysis variables, we consider a time-evolving rela-tionship between Pacific SSTs and precipitation. Theactual station precipitation totals, typically the accu-mulation of small-scale thunderstorm events, are lessimportant than whether it was generally wet or dry overthe area in the 30-day period. The precipitation indicescontain a large sampling of stations within each area,so any trends are likely to mirror those in MFC andreflect real large-scale climate variability.

4. Correlation analyses for ENSO and NorthPacific composites

a. Time series of Pacific SST indices

The average summer Pacific SSTA indices for 1950–97 are shown in Fig. 2 and Table 2. We note, as inMantua et al. (1997), that the NP and Nino-3 indicesthat compose the P index are not independent of eachother. Nino-3 is most related to the NP index in springand early summer, with an average correlation of 0.36

in the period March through July. Near-zero values ofthe P index arise when Pacific SSTs are near average,or when strong tropical and North Pacific SSTs occurin destructive combinations, such as in 1976 and 1983.In the Great Plains, long-term variation in the P indexseems to correspond with sustained dry (low P) or wetperiods (high P). Using the standardized precipitationindex in Colorado, McKee et al. (1999) note the wetperiods since 1950 as 1957–59, 1965–75, and 1979–96,and the dry periods as 1951–57, 1963–65, and 1975–78. An opposite relationship may exist in Arizona,which experienced wet summers in the mid-1950s co-incident with frequent monsoon-related flash floodevents (Carleton et al. 1990; Hirshboeck 1999). Thefollowing time-coincident correlation analyses for theENSO and North Pacific composites are presented as15-day snapshots starting at Julian day 140 (20 May)through Julian day 245 (2 September), respectively.


Note that for the North Pacific composite, a higher valueof correlation is required for statistical significance be-cause of the smaller sample size.

b. Upper-level circulation

In the ENSO composite, there is a time evolution ofcorrelation patterns between the Nino-3 index and

(Fig. 3). A coherent series of atmospheric tele-ZF (t)500

connection patterns across the Pacific Ocean and NorthAmerica occur during the course of the summer season.In late spring (days 140 and 155), the Nino-3 index iscorrelated with a positive phase of the North Pacificpattern, with height deviations in the western and centralNorth Pacific (r 5 20.5) and over the Gulf of Alaska(r 5 0.6). The positive (negative) phase of the NorthPacific pattern, a preferred El Nino (La Nina) responseduring the spring, is associated with a southward (north-ward) shift of the jet stream across the North Pacific(Bell and Janowiak 1995). The spring storm track isdirected toward southern California and the Southwest(Pacific Northwest). Though the positive (negative)phase of the North Pacific pattern in winter and springbrings above- (below-) average precipitation in theSouthwest, it precedes dry (wet) monsoons (Higgins andShi 2000).

The most statistically significant relationship betweenthe Nino-3 index and upper-level circulation in theENSO composite is at monsoon onset in late June andearly July, not monsoon peak in late July and earlyAugust. At onset, Nino-3 is correlated with the negativephase of the Pacific transition (PT) pattern, which oc-curred in association with the flood of 1993 in the mid-western United States (Bell and Janowiak 1995). ThePT pattern appears exclusively in summer and is distinctfrom the ENSO-associated winter teleconnection pat-terns. A stronger than average trough (Nino-3 high) orridge (Nino-3 low) is centered in the northern RockyMountains. The strongest negative correlation with

in North America (r 5 20.7, day 185) duringZF (t)500

the entire summer, significant at the 99% level, is cen-tered over Montana in late June to early July. Like withthe winter Pacific North America pattern, PT is part ofa coherent series of height deviations that extend acrossthe Pacific with centers of action near 258N, 1508E (r5 0.4), the central North Pacific (r 5 20.6), and theGulf of Alaska (r 5 0.3). In late July and August (day200 onward) the correlation between Nino-3 and

becomes statistically insignificant.ZF (t)500

If the hypothesis offered by Gurshunov and Barnett(1998) is true for the summer as well as winter, wewould expect statistically significant patterns in atmo-spheric circulation for the North Pacific composite(when ENSO is neutral). The correlation of the NP indexwith for the North Pacific composite (Fig. 4)ZF (t)500

reveals a different and coherent teleconnection pattern.Note that only days 155–200 are shown in Fig. 4. Asignificant relationship of the NP index to the negative

phase of the east Pacific (EP) pattern emerges in Juneand July. On day 170, there are centers of action in thecentral North Pacific (r 5 20.6), the Gulf of Alaska (r5 0.6), and the northern Great Plains and upper Midwest(r 5 20.6). As with the relationship with Nino-3 andthe PT pattern in the ENSO composite, the relationshipof the NP index to the EP pattern becomes less statis-tically significant through late July and August (notshown). Tropical and North Pacific SSTs are related tothe location of the monsoon ridge in distinct ways. Alow (high) Nino-3 is related to a north (south) displace-ment of the monsoon ridge from its climatological po-sition. A low (high) NP index is related to a northeast(southwest) displacement of the ridge.

The evolution of the these teleconnection patterns islinked to rapid changes in the mean climatological fea-tures associated with NAMS. Figure 5 shows the cli-matological evolution of the 200-mb zonal winds acrossthe Pacific. A strong Pacific jet extending from Japanacross the central Pacific is present from late Maythrough the beginning of July. Winds in the core regionof the jet range from 25 to 35 m s21. As the monsoonridge evolves over North America, the core of the Pacificjet retreats northward and dramatically weakens to near-ly half its spring value (15–20 m s21). The upper-levelmonsoon ridge over the western United States, as seenfrom the climatological evolution of 500-mb geopoten-tial height, attains its maximum strength over the FourCorners region in late July and August (Fig. 6). Coin-cident with the weakening of the jet over the Pacific,the jet over North America strengthens (25 m s21 onday 200). The evolution of the monsoon ridge over thewestern United States also coincides with the maximumwestward extension of the Bermuda high into the south-east United States (not shown).

c. Atmospheric moisture

It can be argued that reliability of NCEP–NCAR re-analysis moisture fields is questionable, especially inlight of the poorly resolved topography over the westernUnited States and limited rawinsonde observations. Pre-vious NAMS climatological studies have also shownthere are substantial differences in moisture transportbetween the European Centre for Medium-RangeWeather Forecasts and NCEP–NCAR reanalyses (Hig-gins et al. 1997b; Schmitz and Mullen 1996). In spiteof these serious concerns, the MF and MFC correlationanalyses are vital to the present investigation of NAMSinterannual variability because they 1) verify conclu-sions of previous work, 2) fit into a consistent frame-work of the teleconnection patterns revealed by analysisof 500-mb geopotential height, and 3) show identicaltrends as precipitation data from actual Coop observingsites, as discussed in the next section. The results of thecorrelation analyses for the moisture fields is shown forday 185 only in Fig. 7, when the teleconnection rela-tionships are most statistically significant.


FIG

.3.

Cor

rela

tion

ofti

me-

coin

cide

ntN

ino-

3in

dex

wit

hfr

omJu

lian

day

140

thro

ugh

Juli

anda

y24

5fo

rye

ars

ofE

NS

Oco

mpo

site

.C

onto

urin

terv

alis

0.1

and

abso

lute

valu

esle

ssZ

F(t

)50

0

than

0.3

are

not

show

n.D

ark

and

ligh

tsh

aded

regi

ons

indi

cate

fiel

dsi

gnifi

canc

eat

the

95%

leve

lfo

rne

gati

vean

dpo

siti

veco

rrel

atio

n,re

spec

tive

ly.


FIG

.3.

(Con

tinu

ed)


FIG

.4.

Sam

eas

Fig

.3

for

corr

elat

ion

ofN

Pin

dex

wit

hfo

rye

ars

ofN

orth

Pac

ific

com

posi

tefo

rJu

lian

days

155–

200.

ZF

(t)

500


FIG. 5. Five-day running mean climatological evolution of 200-mb zonal wind over the Pacific andNorth America from Julian day 140 through Julian day 245. Contour interval is 5 m s21. Values greaterthan 20 m s21 are shaded.

The correlation of MF(t) with Nino-3 and the NP indexZfor the ENSO and North Pacific composites is presentedas a vector in Figs. 7a and 7b. There are two importantfeatures present in both of the composites that yieldsome additional information about the large-scale cir-culation associated with the teleconnection patterns. Thecirculation anomaly in either the northern Great Plains(ENSO composite) or the upper Midwest (North Pacificcomposite) indicates the presence of a surface ridge (LaNina, low NP) or trough (El Nino, high NP). The low-level circulation anomaly would modulate the strengthof the Great Plains LLJ, directing low-level moistureinto eastern Mexico or the central United States, re-spectively (Mo et al. 1997). Though not indicated in theanalysis of 500-mb geopotential height, the moistureflux correlation vectors in Figs. 7a and 7b indicate a

relationship to an anticyclonic (El Nino, high NP) orcyclonic (La Nina, low NP) circulation in the easternPacific west of Baja California (statistically significantin the ENSO composite). A cyclonic circulation in thisarea, associated with easterly waves or tropical cy-clones, is a precursor to the Gulf surge, an episodicenhancement of the Baja LLJ and moisture transportinto the Colorado River valley (Douglas and Real 2000).Carleton et al. (1990) also suggest the SST gradientbetween the Gulf of California and east Pacific wouldbe larger in La Nina conditions, which would alsostrengthen the Gulf surge. The moisture flux analyseshint that the teleconnection patterns associated withtropical and North Pacific SSTs likely affect the low-level moisture transport of the Great Plains and BajaLLJs in the same way. A more detailed reanalysis over


FIG. 6. Same as Fig. 5 for 500-mb height over North America. Contour interval is 25 m.

North America, long-term observational studies of bothlow-level jets, and associated regional atmosphericmodeling are necessary to test these hypotheses.

Similarly, the correlation of MFC(t) with the SSTAZindices for the composites is presented in Figs. 7c and7d. Both composites show that an inverse relationshipbetween atmospheric moisture in the Southwest andGreat Plains exists interannually as well as seasonally.

In the ENSO composite, there is a statistically signifi-cant positive correlation of MFC(t) with Nino-3 (r 5 0.7)Zin the northern Great Plains and the upper Midwest anda negative correlation near the northern end of the Gulfof California (r 5 20.7). In the North Pacific composite,the statistically significant positive correlation with

MFC(t) is centered in the southern Great Plains (r 5 0.7)Zand a negative correlation (r 5 20.5) exists over Ari-


←

FIG. 7. Correlation of Nino-3 and the NP index with reanalysismoisture variables for years of ENSO and North Pacific Pacific com-posites, respectively. The correlation with MF(t) is shown in (a) andZ(b) with a unit vector length of 1 (see scale at bottom of figure). Thecorrelation with MFC(t) is shown in (c) and (d). Contour interval isZ0.1 and absolute values less than 0.3 are not shown. Shading indicatesfield significance at the 95% level.

zona and New Mexico. When the PT and EP telecon-nection patterns are strongest at monsoon onset, MFCis favored in either the Southwest or Great Plains andMidwest. The spatial relationship of tropical and NorthPacific SSTs to Great Plains and Midwest MFC is inagreement with precipitation modes in the central U.S.precipitation found by Ting and Wang (1997), showingthat the tropical and North Pacific associated telecon-nection patterns have distinct relationships with atmo-spheric moisture and precipitation.

5. Utility of the Pacific index

Since tropical and North Pacific SSTs are related todifferent teleconnection patterns during NAMS evolu-tion, the combined Pacific (P) index might be the betteroverall diagnostic measure for summer moisture acrossthe western United States. To test this hypothesis, weran the same correlation analyses as in section 4 usingall the years available in the reanalysis and all threeindices. The largest improvement in explained varianceby the P index over Nino-3 and NP indices is for MF(t) .ZFigure 8 shows, for example, the explained variance of

MF(t) by the SST indices on day 185. The P index ex-Zplains the most variance of MF(t) over the SouthwestZ(Fig. 8c), with increases in New Mexico by approxi-mately 20% over the Nino-3 index and in northwestMexico by approximately 5%–10% over the NP index.

The high and low P index composite summertimeevolution of the dominant daily component of MFC wasconstructed (Fig. 9) using the 15 highest and lowestsummer average P index years shown in Fig. 9 and Table2. These 15-yr averages are compared to the 50-yr MFCclimatology in the same figure. The greatest differencesin MFC between high and low P index years in theGreat Plains and Southwest are at monsoon onset. Dif-ferences in MFC evolution in the Southwest begin inmid-June. If the onset of the monsoon is defined as thepoint where the average dominant daily component ofMFC becomes positive in the Southwest, there is a dif-ference of about 10 days in onset dates between high P(16 July) and low P (5–6 July) years in the Southwest.These dates are slightly later and earlier, respectively,than the climatological onset date of 7 July that Higginset al. (1997b) obtained using area-averaged precipitationdata for their Southwest region. Differences in GreatPlains MFC become apparent by the beginning of June(Fig. 9b). MFC increases (decreases) in the Great Plainsafter this time in the high (low) P index years. The


←

FIG. 8. Correlation and percent explained variance (r2) of MF(t) on Julian day 185 (4 Jul) by (a) Nino 3, (b) NP index, and (c) P index usingZall reanalysis years. Shading intervals at 20%, 30%, and 40%. Vector length is 1 and vector scale is shown at bottom of each panel.

climatological decrease in MFC that occurs near thebeginning of July (Fig. 9b) is delayed until early to mid-July in high P index years. The absolute percentagedifference in MFC from climatology for high and lowP index years is much higher in the Great Plains (50%–100%) than in the Southwest (5%–10%), where mois-ture is less variable, on average. The evolution of MFCabruptly reverts to climatology after the beginning ofAugust, when the teleconnection patterns break down.

The behavior of the correlation of Great Plains andSouthwest precipitation indices with the P index showssome difference from the correlation of the Nino-3 andNP indices with MFC(t) , but results are broadly consis-Ztent (Fig. 10a). The greatest discrepancy is during Mayin the Southwest, when the precipitation index shows astrong positive correlation with the P index (r . 0.5).The positive relationship in spring likely reflects theinfluence of the North Pacific pattern on the winter andspring storm track into the southwest United States.However, as the PT and EP relationships emerge andthe monsoon ridge becomes the dominant control onprecipitation in mid-June the correlation between theGreat Plains and Southwest begins to diverge. The max-imum correlation with the P index occurs near monsoononset, 29 June (day 180) in the Great Plains (r 5 0.64)and 9 July (day 190) in the Southwest (r 5 20.37).Great Plains and Southwest precipitation are inverselyrelated until the end of July. The maximum absolutevalues of r in both regions is largest when the combinedP index is used in the correlation analyses. Beyond lateJuly the correlation with the P index in the Southwestis weakly positive but not statistically significant. Asecondary maximum in correlation appears in the GreatPlains in August that does not appear in MFC, possiblydue to moisture recycling by local evaporation and tran-spiration. The precipitation indices, in general, capturethe same time evolution of the NAMS relationship toPacific SSTs revealed in MFC, particularly at monsoononset. They confirm that a conclusive link exists be-tween reanalysis MFC and point source station precip-itation data considered over broad regions, at least inthe United States.

Extending the analysis with the precipitation indicesthrough the whole year (Fig. 10b) verifies the presenceof established winter ENSO–NPO precipitation rela-tionships in the western United States. In addition tosummer, precipitation indices in the Great Plains andSouthwest are significantly related to Pacific SSTs (r .0.3) in the same way during early spring (March) andmid- to late fall (November). These peaks reflect anintensified (weakened) storm track through the south-west United States in El Nino (La Nina) years. Thoughlocal convective thunderstorms or mesoscale convective

complexes are the dominant mechanism for summerrainfall in the Southwest and the Great Plains, precip-itation in the summer has an equally strong, if not stron-ger, relationship to Pacific SSTs as in winter.

Given this evolution in precipitation relationshipswith Pacific SSTs, one might pose the question of howa persistent ENSO–NPO regime would affect long-termdry or wet periods. In this context, ‘‘long term’’ refersto a period of approximately six months to several years.In the Southwest the sign of the correlation dramaticallyreverses between winter and summer. Dry or wet periodsin the Southwest may be seasonally intense but weakenin the long term. In the Great Plains, however, the signof correlation of precipitation with the P index is nearlyalways positive through the entire year. Persistent Pa-cific SSTs, therefore, would tend to amplify long-termwetness or dryness there. The mid-1950s drought andthe extended wet period since the late 1970s are goodexamples (e.g., McKee et al. 1999).

6. Time-lagged relationships with Pacific SSTAindices and SSTA evolution

The simultaneous correlation analyses, using all thereanalysis years, were repeated with Pacific SST indiceslagged from 1 to 4 months. Only the variable isZF (t)500

considered for day 185, when the simultaneous rela-tionships are strong. The correlation of Nino-3 with

shows a relationship to atmospheric circulationZF (t)500

that is strongest in the simultaneous correlation analysesand statistically significant at the 95% level up to a 4-month lag (Fig. 11a). Though a strong relationship existsbetween the NP index and in the simultaneousZF (t)500

correlation analyses, there is no statistically significantrelationship at just a 2-month lag (not shown). SpringNorth Pacific SSTs in the regions used for the NP indexhave little relationship to subsequent summer atmo-spheric circulation anomalies. A similar result was ob-tained by Ting and Wang (1997) in a lag correlation ofsummer average 500-mb geopotential height with theirNorth Pacific mode of SST related to summer precipi-tation variability in the central United States.

Higgins et al. (1999) showed that an association withtropical Pacific SSTs is important for the NAMS. Priorto wet monsoons, cold SST anomalies appear in theequatorial Pacific cold tongue near the date line in winterand increase in amplitude during the spring. Thesechanges in tropical Pacific SST are associated with aweakened, northward displaced ITCZ and suppressedlocal Hadley circulation. The opposite is true for drymonsoons. The coherence of summer circulation anom-alies with antecedent spring Nino-3 suggests tropicalPacific SSTs may be a good NAMS predictor, though


FIG. 9. Evolution of the dominant daily component of integrated moisture flux convergence (units: mm day 21) for Southwest (0000 UTC)and Great Plains (0600 UTC) from May to Sep. The 50-yr reanalysis climatology is shown in (a) and (b). The 15 composite high P index(solid line with squares) vs low P index (dashed, line with triangles) summers are shown in (c) and (d).

the physical mechanisms relating the two need to beinvestigated.

The weak lag correlation using the NP index is likelyan artifact of the regions used to define it. Other areasin the North Pacific, besides that used for the NP index,may be more related to the NAMS in the spring season.Since NAMS interannual variability is related to theoccurrence of the North Pacific pattern, the regions ofSST associated with this winter and spring teleconnec-tion should have a relationship to subsequent summerconditions. Higgins and Shi (2000) demonstrated thatearly onset monsoons in the Southwest are associatedwith warm SSTs in the subtropical North Pacific (near208N, 1708W) and cold SSTs in the midlatitude centralNorth Pacific (408N, 1808) during winter. This distri-bution of SSTs is consistent with a negative phase ofthe North Pacific pattern, with cold SSTs in the vicinityof large sensible and latent heat fluxes near the north-ward-displaced jet stream. Higgins and Shi also devel-oped a winter North Pacific SST index, similar in prin-

ciple to the NP index in this study, and demonstratedits usefulness in predicting the monsoon onset date.

To further investigate whether the summer PacificSSTAs may be part of a pattern that evolves in time,we use the Reynolds and Smith (1994) data. We com-pute the composite monthly evolution of SSTA, startingin the previous winter, using the 15 highest and 15 low-est P index summers (Figs. 12 and 13). For brevity, notethat the SSTA evolution in Figs. 12 and 13 is shown in2-month averaged time blocks. We use the P index herebecause this index, as discussed earlier, explains thegreatest amount of variance in NAMS atmosphericmoisture and precipitation. Two important features arerevealed that have implications for NAMS predictabil-ity. First, the magnitude of SSTAs in the eastern tropicalPacific increases in time from the previous winter, con-firming the earlier analysis of Higgins et al. (1998) re-lating monsoon onset date to the development of El Ninoor La Nina conditions in the spring. Second, the NorthPacific SST dipole is clearly seen in the summer months,


FIG. 10. Correlation of time-coincident P index with regional pre-cipitation indices for the Great Plains (solid line, dark circles) andSouthwest (dashed line, open circles) for (a) monsoon season (Maythrough Sep) and (b) the entire year.

oriented along a diagonal from approximately 308N,1808 to 508N, 1308W. Again, the location of the dipolein summer corresponds roughly to the North Pacificmode of Ting and Wing (1997). In the antecedent spring,the Pacific SST dipole is still present, but it is meridi-onally oriented along approximately 1708W, corre-sponding with the North Pacific teleconnection patternand consistent with Higgins and Shi (2000).

Using the COADS data, we constructed a modifiedNP index for the spring, shifting ENP box to 408–558N,1808–1608W (see Fig. 2) and repeated the lag correlationanalysis in the same way described earlier. With thechange in regions for the NP index in spring, a statis-tically significant relationship now exists between theatmospheric circulation and North Pacific SSTs at 4-month lag (Fig. 11b). Using the modified NP index, weconstructed a modified P index and repeated the lagcorrelation. The modified P index has the highest andmost statistically significant correlation with atZF (t)500

4-month lag (Fig. 11c). The entire state of the tropical

and extratropical Pacific in spring must be consideredto maximize predictability for subsequent summerweather conditions in the western United States.

7. Discussion

The contrasting summers of 1993 and 1988 serve asarchetypal climatologies associated with high and lowP index, respectively. In the Midwest flood of 1993, astrong trough situated over the northern Rocky Moun-tains and northern Great Plains maintained westerly flowacross all of the western United States in the summer.The jet stream and associated synoptic eddies werestronger than average and south of their mean clima-tological position. The Great Plains and Midwest, to theeast of the trough and north of the jet stream, werefavored for the development of thunderstorms in theform of mesoscale convective complexes (Bell and Jan-owiak 1995; Mo et al. 1997; Trenberth and Guillemot1996). Monsoon rainfall in the Southwest was belowaverage because onset was delayed until early August,the latest on record since 1948 (Okabe 1995; Higginsand Shi 2000). By contrast, in the drought of 1988, amonsoon ridge located over the Great Plains steered thejet stream north into Canada from late spring into earlysummer (Bell and Janowiak 1995; Trenberth 1992). TheGreat Plains and Midwest, on the subsiding branch ofthe ridge, received little precipitation and temperatureswere much above average. There was enhanced easterlyflow on the southern side of the ridge. The monsoonbegan early in late June and summer rainfall was aboveaverage in the Southwest (Higgins and Shi 2000).Though this circulation pattern broke down in mid-Julyand rainfall returned to normal in the central UnitedStates, the late summer rains were insufficient to breakthe hydrologic drought (Trenberth and Guillemot 1996).

Modeling studies of 1988 and 1993 suggest the dis-tribution of tropical and midlatitude Pacific SSTs, andassociated diabatic heating patterns, produced time-co-incident teleconnection responses. Trenberth and Bran-stator (1992) argued that latitudinal variation in theITCZ changed the distribution of tropical Pacific heat-ing, in agreement with observations of dry and wet mon-soon years (e.g., Higgins et al. 1998). Using a linearizedbaroclinic model, they showed that these heating pat-terns provide the sources and sinks for quasi-stationaryRossby waves that propagate into the extratropics. Liuet al. (1998) took a similar modeling approach with astationary wave model linearized about the mean sum-mer climate. However, they found that the summertimecirculation pattern over North America was relativelyinsensitive to diabatic heating in the tropical Pacific.The greater effect on the circulation over North Americawas from diabatic heating associated with an upper-leveltrough or ridge off the west coast of the continent, dia-batic heating in the western Pacific, and vorticity forcingby transient eddies.

Newman and Sardeshmukh (1998) demonstrated that


FIG

.11

.C

orre

lati

onof

wit

h(a

)N

ino-

3,(b

)m

odifi

edN

Pin

dex,

and

(c)

mod

ified

Pin

dex

at4-

mon

thla

gon

Juli

anda

y18

5.C

onto

urin

terv

alis

0.1

and

abso

lute

valu

esle

ssth

an0.

2Z

F(t

)50

0

are

not

show

n.D

ark

and

ligh

tsh

aded

regi

ons

indi

cate

fiel

dsi

gnifi

canc

eat

95%

leve

lfo

rne

gati

vean

dpo

siti

veco

rrel

atio

n,re

spec

tive

ly.


FIG

.12

.Tw

o-m

onth

aver

aged

SS

TA

evol

utio

nfo

r15

com

posi

tehi

ghes

tP

inde

xsu

mm

ers

begi

nnin

gth

epr

evio

usJa

nuar

y.C

onto

urin

terv

alis

0.2.

Abs

olut

eva

lues

grea

ter

than

plus

and

min

us0.

4ar

esh

aded

ligh

tan

dda

rkre

spec

tive

ly.


FIG

.13

.S

ame

asF

ig.

12fo

r15

low

est

Pin

dex

sum

mer

s.


FIG. 14. Idealized relationship of monsoon ridge position and midlevel moisture transport toPacific SSTs at monsoon onset.

height anomalies over western North America have atime-varying sensitivity to Rossby wave forcing in thePacific. As the westerlies weaken, the forcing regionshifts from the east and central Pacific in winter to thewest Pacific in spring. Height anomalies over NorthAmerica are most sensitive to forcing in the west Pacificin late spring as the Asian monsoon is intensifying, butthis sensitivity diminishes in July and August. Chen andNewman (1998) also concluded that Rossby wave forc-ing from southeast Asia and the west Pacific were crit-ical to the development of an anticyclone over centralNorth America in the drought of 1988.

The spatial and temporal evolution of the telecon-nection patterns in the present study is consistent withNewman and Sardeshmukh (1998). Tropical and NorthPacific SSTs are each associated with distinct summercirculation responses over North America in the formof the PT and EP teleconnection patterns. The PT pattern(Fig. 3) appears to be related to forcing in the tropicalPacific that moves westward and northward with timefrom late spring to early summer. The EP pattern (Fig.4) may also be related to forcing in the west Pacific or,more likely, off the west coast of North America as Liuet al. (1998) suggest in their modeling study.


The PT and EP patterns have their maximum rela-tionship to their respective Pacific SSTA indices at themonsoon onset. The weakening of the teleconnectionpatterns in mid- to late summer coincides with weak-ening of the Pacific jet, strengthening of the monsoonridge, and the westward extension of the Bermuda high.These events always occur because their physical mech-anisms are independent of Pacific SSTs. The Pacific jetweakens because baroclinicity in the Northern Hemi-sphere is weakest in mid-summer. The monsoon ridgeover Mexico and the western United States evolves asa result of sensible heating from elevated heat sourcesand latent heating from diurnal thunderstorms (e.g., Bar-low et al. 1998). Pacific SSTs appear to play a role inmodulating the timing of these events in a climatologicaltransition period in early summer. In short, teleconnec-tions originating in the Pacific likely affect how theNAMS begins, but now how it ends.

Some modeling studies suggest that the North PacificSST dipole develops as a result of time-coincident forc-ing by ENSO-related tropical convection (e.g., Alex-ander 1992; Lau and Nath 1996; Deser and Timlin1997). These studies would suggest that North PacificSSTs provide little predictive value beyond a synoptictimescale of several weeks. However, the presence ofa statistically significant circulation relationship in theNorth Pacific composite suggests that the North Pacificaffects the summertime circulation over North Amer-ica, independent of ENSO. Moreover, the evolution ofthe NAMS is related to a seasonally evolving patternof Pacific SSTAs. Accounting for the evolution of theNorth Pacific SST dipole and the development ofENSO conditions, the relationships between PacificSSTAs and summer atmospheric circulation over NorthAmerica are persistent from the antecedent spring. Theentire time-evolving state of the Pacific, as expressedin the P index, should be considered in either the time-coincident or time-lagged relationships to maximizethe explained variance of atmospheric circulation, at-mospheric moisture, and regional precipitation. Futurework should aim to further understand the physicalmechanisms behind the seasonal cycle of interannualPacific SSTA variability and its implications forNAMS predictability.

Modeling studies confirm that land surface condi-tions, which have not been investigated here, may alsoaffect the development of monsoons (e.g., Anthes andKuo 1986; Dirmeyer 1994; Meehl 1994; Sud et al.1995). Gutzler and Preston (1997) have suggested thatsnow cover affects the surface energy budget of theRocky Mountains and Colorado plateau in late spring,and, hence, the evolution of the monsoon ridge. Inagreement with this hypothesis, Higgins et al. (1998)noted that wet (dry) monsoons in Arizona–New Mexicotend to follow winters with wet (dry) conditions in thePacific Northwest and dry (wet) conditions in the South-west. We suspect, however, that Pacific SST forcing isthe dominant factor in the winter and summer climate

of the southwest United States. The relationship be-tween snowfall and NAMS precipitation is likely ob-served as a coincidence of the changes in large-scaleteleconnection patterns associated with ENSO–NPOand precipitation in the Southwest between winter andsummer.

Central and southern Mexico are not the focus of thisstudy, but this area deserves mention because NAMSprecipitation is most pronounced there. Statistically sig-nificant relationships have been found to exist betweensummer precipitation and tropical Pacific SSTs in thisregion (Higgins et al. 1999). In both ENSO and NorthPacific composites, there are slight negative relation-ships between MFC(t) and Nino-3 and NP indices. How-Zever, these relationships are not as statistically signifi-cant nor as persistent as the relationships in the south-west United States or Great Plains. Long-term reanalysisatmospheric moisture may be worse over Mexico be-cause of the paucity of continuous rawinsonde records.The interannual variability of the NAM in Mexico isprobably less sensitive to the midlatitude teleconnectionpatterns associated with remote Pacific SSTs. Summerprecipitation in Mexico is more likely tied to SSTs inthe adjacent east Pacific, which in turn vary with ENSO.Colder (warmer) SSTs along the western Mexican coastwould enhance (diminish) the ocean–land temperaturegradient, hence the strength of the monsoon (Higginset al. 1999). The effect of east Pacific tropical systemsand their relationship to ENSO should also be consid-ered (Reyes and Mejıa-Trejo 1991; Douglas 2000).Mexican rainfall may also be influenced by factors in-dependent of Pacific SSTs, such as the quasi-biennialoscillation (Douglas 2000). Because of the differencein climatological evolution and possible difference ofcontrols on interannual variability, the NAMS in Mexicoshould be treated separately from the NAMS in the Unit-ed States.

If the evolution of the NAMS in the United States isprimarily driven by remote Pacific SST forcing in theearly part of the monsoon, what happens during pre-monsoon and late monsoon periods? We speculate thatat these times the land surface factors may become moreimportant for the large-scale circulation. In 1988 and1993, antecedent soil moisture influenced surface latentand sensible heat fluxes, which may have provided pos-itive or negative feedback, respectively, to the droughtor flood conditions (Giorgi et al. 1996; Trenberth andGuillemot 1996). Future investigation of the NAMSshould explore the physical linkages between remotePacific SST forcing, land surface processes, and NAMSevolution. Regional atmospheric models are well suitedto such a task.

The fact that ENSO–NPO relationships with precip-itation in the Southwest dramatically change throughthe seasons is a complicating factor in forecasting long-term drought or wet conditions in that region. Extendedwetness and dryness in the Great Plains is very sensitiveto variability in Pacific SSTs since ENSO–NPO is re-


lated to precipitation in the same way throughout theyear. The record or historically wet and dry periods inColorado found by McKee et al. (1999) support thisconclusion. The weather conditions in early summermust be a critical factor in determining the annual soilmoisture surplus or deficit. Though there have been verydry summers like 1988 or 1983, the Great Plains havenot experienced a long-term multiyear drought in thepast two decades, corresponding with a period of sus-tained El Nino–high NPO conditions. The Great Plainscould expect a greater frequency of long-term droughtshould these conditions change in the future.

8. Summary

In this study, a NAMS reanalysis climatology of 500-mb geopotential height, integrated moisture flux, andintegrated moisture flux convergence was computed us-ing the 50-yr NCEP–NCAR daily reanalysis. A time-varying correlation approach was used to relate dailydeviations from the climatology to three Pacific SSTAindices that capture the variability of tropical and NorthPacific SSTs related to ENSO and NPO. New time-evolving precipitation indices, of Southwest and GreatPlains stations, were also related to these Pacific SSTAindices.

Tropical and North Pacific SSTs are related to theoccurrence of both Pacific Transition and East Pacificteleconnection patterns, respectively, over North Amer-ica in summer. High (low) P index years are character-ized by the negative (positive) phases of these patterns.A trough (ridge) is located over the northern RockyMountains and central Great Plains. These relationshipsare strongest at monsoon onset in late June and earlyJuly, when the teleconnection patterns control the large-scale distribution of moisture across the western UnitedStates (Fig. 14). In the Great Plains, because the springwet season is lengthened (shortened), early summerMFC and rainfall are above (below) average. In theSouthwest, monsoon onset is late (early), and early sum-mer MFC and rainfall are below (above) average. TheGreat Plains and Baja LLJs are likely related to PacificSSTs, respectively, by the presence of a surface ridgeor trough in the central United States and changes intropical easterly wave activity. These relationships de-cay in the later part of the monsoon coincident withweakening of the jet stream across the Pacific andstrengthening of the monsoon ridge over North Amer-ica. These events always occur because their physicalmechanisms are independent of Pacific SSTs.

Because tropical and North Pacific SSTs are associ-ated with distinct teleconnection patterns, they are re-lated to the deviation of the monsoon ridge from itsclimatological average position in different ways. Trop-ical Pacific SSTs are related to the variation in ridgeposition to the north and south, and North Pacific SSTsto variation in ridge position to the northeast and south-west. These ridge configuration are consistent with those

of Carleton et al. (1990) associated with interannualvariability of NAMS precipitation in the Southwest. Themost coherent summer climate patterns over the entirewestern United States occur when Pacific SSTs are ina substantially high or low P index configuration, in-dicating that there is a constructive interference ofENSO and NPO in the summer as well as winter. TheP index is the better diagnostic for summer climate inthe western United States and the better predictor whenthe seasonally evolving state of Pacific SSTAs is con-sidered.

Acknowledgments. This research was funded by theSignificant Opportunities in Atmospheric Research andScience Program (SOARS) through the University Cor-poration for Atmospheric Research and the NationalScience Foundation, the Cooperative Institute for Re-search in the Atmosphere through NOAA, and NASAGrant NAG8-1511. The authors would like to acknowl-edge the comments from two anonymous reviewers thatimproved the manuscript.

REFERENCES

Adams, D. K., and A. C. Comrie, 1997: The North American mon-soon. Bull. Amer. Meteor. Soc., 78, 2197–2213.

Alexander, M. A., 1992: Midlatitude atmosphere-ocean interactionduring El Nino. Part I: The North Pacific Ocean. J. Climate, 5,944–958.

Anthes, R. A., and Y. H. Kuo, 1986: The influence of soil moistureon circulations over North America on short time scales. NamiasSymposium, Scripps Institution of Oceanography Reference Se-ries, Vol. 86-17, J. O. Roads, Ed., 132–147.

Barlow, M., S. Nigam, and E. H. Berbery, 1998: Evolution of theNorth American monsoon system. J. Climate, 11, 2238–2257.

——, ——, and ——, 2000: ENSO, Pacific decadal variability, andU.S. summertime precipitation, drought, and streamflow. J. Cli-mate, 14, 2105–2128.

Bell, G. D., and J. E. Janowiak, 1995: Atmospheric circulation as-sociated with the Midwest floods of 1993. Bull. Amer. Meteor.Soc., 76, 681–695.

Bonner, W. D., 1968: Climatology of the low level jet. Mon. Wea.Rev., 96, 833–850.

Brenner, I. S., 1974: A surge of maritime tropical air—Gulf of Cal-ifornia to the southwestern United States. Mon. Wea. Rev., 102,375–389.

Bryson, R. A., and W. P. Lowry, 1955: The synoptic climatology ofthe Arizona summer precipitation singularity. Bull. Amer. Me-teor. Soc., 36, 329–339.

Bunkers, M. J., J. R. Miller, and A. T. DeGaetano, 1996: An ex-amination of El Nino–La Nina related precipitation anomaliesacross the northern plains. J. Climate, 9, 147–160.

Carleton, A. M., D. A. Carpenter, and P. J. Weser, 1990: Mechanismsof interannual variability of the southwest United States summerrainfall maximum. J. Climate, 3, 999–1015.

Castro, C. L., T. B. McKee, and R. A. Pielke Sr., 2000: The rela-tionship of the North American monsoon to tropical and NorthPacific sea surface temperatures as revealed by observationalanalyses. Climatology Rep. 00-1, Department of AtmosphericScience, Colorado State University, 95 pp.

Chen, P., and M. Newman, 1998: Rossby wave propagation and rapiddevelopment of upper level anomalous anticyclones during the1988 U.S. drought. J. Climate, 11, 2491–2504.

Comrie, A. C., and E. C. Glenn, 1998: Principal components-basedregionalization of precipitation regimes across the southwest


United States and northern Mexico, with an application tomonsoon precipitation variability. Climate Res., 10, 201–215.

Cowie, J. R., and T. B. McKee, 1986: Colorado precipitation eventand variability analysis. Climatology Rep. 86-3, Department ofAtmospheric Science, Colorado State University, 102 pp.

Deser, C., and M. S. Timlin, 1997: Atmosphere–ocean interaction onweekly timescales in the North Atlantic and Pacific. J. Climate,10, 393–408.

Dirmeyer, P. A., 1994: Vegetation stress as a feedback mechanism inmidlatitude drought. J. Climate, 7, 1463–1487.

Douglas, A., 2000: The influence of eastern North Pacific storms onsummer rainfall in Mexico. Second Southwest Weather Symp.,Tucson, AZ.

Douglas, M. W., 1995: The summertime low-level jet over the Gulfof California. Mon. Wea. Rev., 123, 2334–2347.

——, and J. C. Real, 2000: The mean fields and synoptic variabilityof moisture fluxes over the Gulf of California and surroundingregions during the summer. Postprints, Second Southwest Weath-er Symp., Tucson, AZ National Weather Service, The Universityof Arizona, and COMET.

——, R. A. Maddox, K. Howard, and S. Reyes, 1993: The Mexicanmonsoon. J. Climate, 6, 1665–1677.

Giorgi, F., L. O. Mearns, C. Shields, and L. Mayer, 1996: A regionalmodel study of the importance of local versus remote controlsof the 1988 drought and 1993 flood over the central UnitedStates. J. Climate, 9, 1150–1162.

Gurshunov, A., and T. B. Barnett, 1998: Interdecadal modulation ofENSO teleconnections. Bull. Amer. Meteor. Soc., 79, 2715–2725.

Gutzler, D. S., and J. W. Preston, 1997: Evidence for a relationshipbetween spring snow cover in North American and summer rain-fall in New Mexico. Geophys. Res. Lett., 24, 2207–2210.

Hales, J. E., 1972: Surges of maritime tropical air northward overthe Gulf of California. Mon. Wea. Rev., 100, 298–306.

Harrington, J. A., R. S. Cerveny, and R. C. Balling, 1992: Impact ofthe Southern Oscillation on the North American monsoon. Phys.Geogr., 13, 318–330.

Higgins, R. W., and W. Shi, 2000: Dominant factors responsible forinterannual variability of the summer monsoon in the south-western United States. J. Climate, 13, 759–775.

——, Y. Yao, E. S. Yarosh, J. E. Janowiak, and K. C. Mo, 1997a:Influence of the Great Plains low-level jet on summertime pre-cipitation and moisture transport over the central United States.J. Climate, 10, 481–507.

——, ——, and X. L. Wang, 1997b: Influence of the North Americanmonsoon system on the U.S. summer precipitation regime. J.Climate, 10, 2600–2622.

——, K. C. Mo, and Y. Yao, 1998: Interannual variability of the U.S.summer precipitation regime with emphasis on the southwesternmonsoon. J. Climate, 11, 2582–2606.

——, Y. Chen, and A. V. Douglas, 1999: Interannual variability ofthe North American warm season precipitation regime. J. Cli-mate, 12, 653–680.

Hirschboeck, K. K., 1999: Climate diagnostics of flooding in Arizonaand implications for climatological forecasts of hydrologic ex-tremes. Proc. 24th Annual Climate Diagnostics and PredictionWorkshop, Tucson, AZ, NOAA/Climate Prediction Center andthe Institute for the Study of Planet Earth and Institute for At-mospheric Physics, The University of Arizona, 367–370.

Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmosphericphenomena associated with the Southern Oscillation. Mon. Wea.Rev., 109, 813–829.

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Re-analysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.

Lau, N. C., and M. J. Nath, 1996: The role of the ‘‘atmosphericbridge’’ in linking tropical Pacific ENSO events to extratropicalSST anomalies. J. Climate, 9, 2036–2037.

Liu, A. Z., M. Ting, and H. Wang, 1998: Maintenance of circulationanomalies during the 1988 drought and 1993 floods over theUnited States. J. Atmos. Sci., 55, 2810–2832.

Livezey, R. E., and W. Y. Chen, 1983: Statistical field significanceand its determination by Monte Carlo techniques. Mon. Wea.Rev., 111, 46–59.

——, and K. C. Mo, 1987: Tropical–extratropical teleconnectionsduring the Northern Hemisphere winter. Part II: Relationshipsbetween monthly mean Northern Hemisphere circulation pat-terns and proxies for tropical convection. Mon. Wea. Rev., 115,3115–3132.

Mantua, N. J., S. R. Hare, U. Zhang, J. M. Wallace, and R. C. Francis,1997: A Pacific interdecadal climate oscillation with impacts onsalmon production. Bull. Amer. Meteor. Soc., 78, 1069–1079.

McKee, T. B., N. J. Doesken, and J. Kleist, 1999: Historical dry andwet periods in Colorado. Climatology Rep. 99-1, Part A: Tech-nical Report, Department of Atmospheric Science, ColoradoState University, 121 pp.

Meehl, G. A., 1994: Influence of the land surface in the Asian summermonsoon: External conditions versus internal feedbacks. J. Cli-mate, 7, 1033–1049.

——, and Coauthors, 2000: An introduction to trends in extremeweather and climate events: Observations, socioeconomic im-pacts, terrestrial ecological impacts, and model projections. Bull.Amer. Meteor. Soc., 81, 413–416.

Minobe, S., 1997: A 50–70 year climatic oscillation over the NorthPacific and North America. Geophys. Res. Lett., 24, 683–686.

Mo, K. C., J. N. Paegle, and R. W. Higgins, 1997: Atmosphericprocesses associated with summer floods and droughts in thecentral United States. J. Climate, 10, 3028–3046.

Nakamura, H., G. Lin, and T. Yamagata, 1997: Decadal climate var-iability in the North Pacific during the recent decades. Bull.Amer. Meteor. Soc., 78, 2215–2225.

Newman, M., and P. D. Sardeshmukh, 1998: The impact of the annualcycle on the North Pacific/North American response to remotelow-frequency forcing. J. Atmos. Sci., 55, 1336–1353.

Nitta, T., and S. Yamada, 1989: Recent warming of tropical sea sur-face temperature and its relationship to the Northern Hemispherecirculation. J. Meteor. Soc. Japan, 67, 375–383.

Okabe, I. T., 1995: The North American monsoon. Ph.D. dissertation,University of British Columbia, 146 pp.

Rasmussen, E. M., 1967: Atmospheric water vapor transport and thewater balance of North America. Part I: Characteristics of thewater vapor flux field. Mon. Wea. Rev., 95, 403–427.

Reitan, C. H., 1957: The role of precipitable water in Arizona’s sum-mer rains. Tech. Rep. on the Meteorology and Climatology ofArid Regions 2, Institute of Atmospheric Physics, The Universityof Arizona, 19 pp.

Reyes, S., and A. Mejıa-Trejo, 1991: Tropical perturbations in theeastern Pacific and the precipitation field over north-westernMexico in relation to the ENSO phenomenon. Int. J. Climatol.,11, 515–528.

Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surfacetemperature analyses using optimum interpolation. J. Climate,7, 929–948.

Ropelewski, C. F., and M. S. Halpert, 1986: North American precip-itation and temperature patterns associated with the El Nino/Southern Oscillation (ENSO). Mon. Wea. Rev., 114, 2352–2362.

Schmitz, J. T., and S. Mullen, 1996: Water vapor transport associatedwith the summertime North American monsoon as depicted byECMWF analyses. J. Climate, 9, 1621–1634.

Slutz, R. J., S. J. Lubker, J. D. Hiscox, S. D. Woodruff, R. L. Jenne,D. H. Joseph, P. M. Stehrer, and J. D. Elms, 1985: ComprehensiveOcean Atmosphere Data Set; Release 1. Climate Research Pro-gram, NOAA/Environmental Research Laboratories, Boulder,CO, 268 pp.

Sud, Y. C., K. M. Lau, G. K. Walker, and J. H. Kim, 1995: Under-standing biosphere-precipitation relationships: Theory, modelsimulations, and logical inferences. Mausam, 46, 1–14.

Tang, M., and E. R. Reiter, 1984: Plateau monsoons of the NorthernHemisphere: A comparison between North America and Tibet.Mon. Wea. Rev., 112, 617–637.

Ting, M., and H. Wang, 1997: Summertime U.S. precipitation vari-


ability and its relation to Pacific sea surface temperature. J. Cli-mate, 10, 1853–1873.

Trenberth, K. E., and G. W. Branstator, 1992: Issues in establishingcauses of the 1988 drought over North America. J. Climate, 5,159–172.

——, and C. J. Guillemot, 1995: Evaluation of the global atmo-spheric moisture budget at seen from analyses. J. Climate, 8,2255–2272.

——, and ——, 1996: Physical processes involved in the 1988drought and 1993 floods in North America. J. Climate, 9, 1288–1298.

Wilks, D. S., 1995: Statistical Methods in the Atmospheric Sciences.Academic Press, 467 pp.

Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like in-terdecadal variability: 1900–93. J. Climate, 10, 1004–1020.

the relationship of the north american monsoon to tropical ... · the north american monsoon is a...

Documents