1. Introduction
In the past several years, the influence of the North Pacific on the climate variability of the United States has received much attention, mainly due to the identification of the so-called Pacific decadal oscillation (PDO) and its possible long-term modulation of El Niño impacts on the climate of the nation (Trenberth and Hurrell 1994; Kawamura 1994; Deser and Blackmon 1995; Zhang et al. 1997; Gershunov and Barnett 1998; Mantua et al. 1997, and others). However, it is worth noting that the importance of the North Pacific in regulating extratropical climate on a wide range of time scales had been known for a long time. In numerous papers dating back to the 1960–70s, Namias and collaborators (e.g., Namias 1959, 1976, and others) observed that SST and atmospheric circulations over the North Pacific had a strong influence on weather and climate over North America. Yet, atmospheric general circulation model studies of the impact of the North Pacific on extratropical climate were generally inconclusive, indicating that North Pacific SST anomalies either have minimal impacts on extratropical circulation or the responses are very sensitive to the basic flow (Pitcher et al. 1988; Ting 1991; Kushnir and Lau 1992; Peng et al. 1997; Alexander et al. 2002). Moreover, ocean model experiments have shown that North Pacific SST anomalies are largely forced by the atmosphere (e.g., Cayan et al. 1995). On the other hand, more recent coupled model studies have shown that air–sea interaction over the North Pacific may amplify midlatitude climate variability induced by El Niño (Lau and Nath 1996, 2001). Given these divergent results, it appears that the role of the North Pacific on climate is still very much an open question.
Up to now, most studies of North Pacific variability and its influence on extratropical climate have been focused on the cold seasons, when the subtropical jet stream and storm tracks are most intense. Few studies have been devoted to the role of the North Pacific on climate during the warm seasons. During the widespread drought over the continental United States in the summer of 1988, conventional wisdom espoused that the drought was caused by the cold SST in the tropical eastern Pacific associated with La Niña (Trenberth et al. 1990 and many others). In contradiction, Namias (1991) postulated that the drought was more strongly linked to North Pacific atmospheric and oceanic anomalies, which were well developed in the spring of 1988. Namias's conjecture was essentially confirmed by subsequent general circulation model experiments that showed that the 1988 drought over the United States was dependent more on the initial atmospheric conditions in spring, rather than on the tropical SST forcing (Mo et al. 1991 and others).
Occurrences of extensive and warmer-than-normal water in the North Pacific accompanying drought conditions over subtropical East Asia during boreal summer have been noted by several recent studies (Park and Schubert 1997; Kawamura 1994). While in some cases an ENSO connection was implicated, the extensive SST and atmospheric circulation anomalies suggested that other amplifying factors may be at work (Overland et al. 2001). Recently, Lau et al. (2000) pointed out that the North Pacific subtropical high and associated extratropical SST may be as important, if not more so, than ENSO in affecting the variability of the East Asian monsoon. Lau and Weng (2000, 2002) identified summertime teleconnection patterns linking precipitation anomalies over the North American continent to those associated with the East Asian monsoon, via variability of the East Asian jet stream and North Pacific SST. Their results are in agreement with studies that showed that rainfall variability in the Midwest and in the southwestern United States may be related to North Pacific SST anomalies (Ting and Wang 1997; Mo et al. 1991), and that rainfall systems in diverse geographic regions may be subject to similar large-scale forcings (Kodama 1992, 1993). These results also corroborate previous findings that anomalous diabatic heating in the tropical western Pacific may excite transpacific wave trains affecting climate variability over the continental United States on interannual and intraseasonal time scales (Nitta 1987; Lau and Peng 1992; Chen and Newman 1998; Mo 2000). Using the canonical ensemble correlation methodology, Lau et al. (2002) have found that a substantial portion of the potential predictability of summertime seasonal precipitation over the U.S. northern Great Plains and the Midwest may be derived from SST variability in the North Pacific during the spring and early summer.
The aforementioned studies have suggested that the North Pacific is an important source of summertime climate variability, in ways that may be distinct from El Niño. In this regard, previous Atmospheric Modeling Intercomparison Project (AMIP) type GCM experiments on extratropical SST impacts may be questionable because of the neglect of the physical processes involved in extratropical air–sea interactions. These considerations have provided strong motivation for a reexamination of the North Pacific in regulating and/or in being driven by the climate variability in the Tropics and the extratropics. Since the North Pacific SST possesses diverse spatial and temporal scales and its climate impacts are strongly dependent on the seasons and on geographic locations, it is necessary to study the myriad scales of variability separately, not only for the boreal winter as in many previous studies, but for all seasons of the year. In this paper, we will provide observational evidence of the presence of North Pacific interannual SST climate modes in boreal summer that are either independent of, or cannot be explained by, El Niño forcings alone. We will discuss their role in regulating warm season climate variability over Eurasia and North America. The present approach is largely limited to associations of large-scale features. Suggestions of causality and/or mechanisms need to be verified with model experiments.
2. Data and analysis procedure
Data used for this study include the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) 500-mb geopotential height field, 850- and 200-mb wind reanalysis, NCEP global SST (Reynolds and Smith 1994), satellite–gauge blended rainfall from the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997), and 1° × 1° global land-based objectively analyzed temperature and rainfall from the Climate Research Unit (CRU), East Anglia University, United Kingdom (New et al. 2000). Station pentad temperature and rainfall data for selected regions in East Asia and North America are also used for analysis of extreme events. The common period for all data is 1955–98, which is used to define the climatology, except for the rainfall data for which the climatology is based on the period 1979–98. The anomaly field is defined as the deviations from the monthly climatology. To focus on interannual variability, the anomaly fields are high passed with an infinite impulse response filter to retain periodicities of less than 9 yr (Hamming 1983). Empirical orthogonal function (EOF) analyses are carried out based on the seasonal mean data to identify the dominant modes of North Pacific SST variability for the summer months June–July–August (JJA; hereafter seasons will be abbreviated by using the first letters of the corresponding months). A North Pacific index (NPI) is then defined using the principal component of the most dominant mode. Based on NPI, correlation, regression, and composite analyses are carried out using the seasonal mean circulation and rainfall anomaly data. Probability distribution functions constructed from the pentad data are used to examine the odds of extreme events with respect to the different phases of the NPI. The significance of serial correlation coefficients is estimated using a Student's t test with degree of freedom estimates following Livezey and Chen (1983).
3. Modes of extratropical SST interannual variability
The relative magnitude and seasonality of the interannual SST variability in the Tropics and extratropics can be gleaned from Fig. 1, which shows spatial patterns of standard deviations of the filtered seasonal mean SST anomalies for the four seasons. Large SST variability associated with El Niño is evident in the tropical central and eastern Pacific for all seasons. The El Niño SST signal is mostly confined near the equatorial eastern Pacific off the coast of South America during spring and summer (Figs. 1a,b), but expands into the central Pacific during the fall and winter. In the extratropics, the largest SST variability is found in the North Pacific and in the North Atlantic along 40°N and is most pronounced during summer and fall, but less so in winter and spring.
Figures 2a and 2b show the spatial patterns of the first two most dominant EOF modes of anomalous SST over the North Pacific for JJA, which account for 25% and 19% of the interannual variance, respectively. The first mode (Fig. 2a) features a mostly warm or cold North Pacific, depending on the signs of the corresponding principal component (PC, not shown). The SST anomaly is centered near 40°N, with a meridional span of 20°, extending from the coast of East Asia across the entire North Pacific. Near the west coast of North America, the SST anomalies split into two cold centers, one over the Gulf of Alaska and another along the subtropics off the west coast of California. The second SST EOF (Fig. 2b) shows major cold anomalies in the northern (>40°N) and eastern (east of 150°W) portions of the ocean basin. The cold center in the western North Pacific appears to be coupled to a warm anomaly in the central and western subtropical Pacific in the form of a meridional dipole SST anomaly. These two SST patterns resemble those obtained by Lau and Weng (2002) from regressions against the principal modes of summertime seasonal precipitation anomalies over the continental United States. Combined EOF calculations using both the North Pacific and North Atlantic show that the North Pacific SST patterns (Figs. 2d,e) are almost identical to those based on the North Pacific alone. The first combined EOF (Fig. 2d) shows that the North Pacific and North Atlantic SSTs vary synchronously with similar SST patterns, that is, a warm North Pacific occurring with a warm North Atlantic and vice versa. The structure of the second EOF shows some similarity in the two ocean basins characterized by a northwest–southeast-oriented dipole SST anomaly, separated by 40°N. SST patterns similar to those shown in Fig. 2 can be obtained from regression against EOFs of the North Atlantic alone (not shown). The results suggest that the North Pacific SST patterns shown in Fig. 2 may be components of an extratropical climate mode linking SST variability in the North Pacific and the North Atlantic via an “atmospheric bridge” across North America (see discussion in next section).
The lagged correlations of the principal component of EOF1 (PC1) with Niño-3.4 SSTs are shown in Fig. 2c. PC1 is significantly correlated with Niño-3.4 (5°S–5°N, 120°–170°W) SST variability (>95% confidence level) during the antecedent spring and early summer. Yet it is not significantly correlated with Niño-3.4 SST in the previous or following winters, when the El Niño signals are strongest. For EOF2, no significant correlation with Niño-3.4 SST exists at any lag. The lagged correlation pattern obtained from using combined North Pacific and North Atlantic data yields the same results (Fig. 2f).
The evolution of coherent SST patterns with PC1 is shown in Fig. 3. Here it can be seen that the North Pacific summertime SST anomalies have a global association, including coherent variations in the tropical Pacific and to a lesser degree in the Atlantic. Notice that the correlation with the tropical eastern Pacific is close to zero everywhere in the previous winter (Fig. 3a), suggesting there is no memory derived from El Niño, which peaks in boreal winter. An organized SST anomaly in the North Pacific emerges first in the previous spring, evolves through summer and fall, and dissipates in the following winter. The SST evolution from spring through summer and fall does not resemble that associated with El Niño. However, the precursory signal in SST should not be interpreted necessarily as the ocean forcing the atmosphere, because it could have arisen simply from the slow time scale in the ocean heat content. It appears that even though the North Pacific pattern may be influenced by SST variability in the tropical Pacific in the previous spring, as the season progresses through the summer to winter, the North Pacific SST establishes its own blend of variability and becomes increasingly decoupled from the influence of tropical SST.
4. Circulation patterns associated with EOF1
The atmospheric circulation associated with the SST EOF1 are examined via composite analyses with respect to maximum and minimum (based on 1-σ threshold) phases of the PC1, which is defined as the North Pacific index (NPI) in the following analyses. As a result of the composite criterion, eight events are selected for each phase and the composite is obtained as the difference (maximum minus minimum) between the means of the two categories. If two successive years meet the 1-σ threshold, only the year with the larger amplitude is selected.
a. The atmospheric bridge
Figures 4a and 4b show the composite JJA 200-mb zonal wind and transients (defined as the root-mean square of the deviation from the 11-day running mean for each summer) with respect to the NPI. From Fig. 4, the North Pacific SST variability is associated with a nearly circumglobal variation of the subtropical jet stream, spanning northern Africa, through northern India, and Eurasia to the North Pacific. Positive SST in the North Pacific is associated with a weakened East Asian jet stream near 35°–40°N, and increased zonal flow to the north. Increased westerly flow is found south of the climatological jet over the central Pacific and Southeast Asia. Concomitant with the poleward shift of the jet is the northward shift of the storm track (Fig. 4b). The transients overlying the warm water (∼40°N) are considerably weaker, while immediately to the north they are enhanced. The shifts in storm track centers are evident in the north–south dipole anomalies found over northern China, the North Pacific, northwestern North America, and the North Atlantic.
Figure 5a shows the 850-mb wind and 500-mb geopotential height pattern associated with NPI. The most pronounced features are zonally oriented low-level anticyclones occupying the entire North Pacific and the North Atlantic, respectively. The anticyclones are associated with increased 500-mb geopotential height, which forms a somewhat amorphous ring around the entire Northern Hemisphere around 40°N. Here, the Pacific signal appears as a part of a global teleconnection pattern including large anomalies not only over the extratropical oceans but also over the polar land regions, as well as the Tropics. The easterlies in the western and southern flanks of the North Pacific anticyclone impinge on central East Asia just south of Japan and Korea. The easterlies appear to be coupled to a cyclonic circulation over the subtropical western Pacific south of the Pacific anticyclone. The Atlantic anticyclone appears to have the similar setup relative to eastern North America as the Pacific anticyclone does to East Asia. The geopotential height anomalies in the Tropics have opposite signs from the midlatitudes, but the same sign as the anomalies in the polar regions (>60°N), except over northern Siberia. Overall, the global pattern suggests the presence of quasi-zonally symmetric structures in the global atmosphere–ocean, interrupted by land–ocean thermal contrasts in the Northern Hemisphere (cf. Schubert et al. 2002).
To ensure the robustness of the teleconnection pattern shown above, we have also used NCEP–NCAR wind reanalyses and CMAP rain for the last 21 years (1979–99) to repeat the previous analysis. The CMAP rainfall data are only available for the shorter period. Because of the shorter data record, only four cases are selected for the composite shown in Fig. 5b. The basic large-scale features, especially the contemporaneous anticyclones over the North Pacific and the North Atlantic, are similar to those shown in Fig. 5a. The anticyclone–cyclone coupling between the North Pacific and the subtropical western Pacific is even more pronounced. Also evident is the presence of a dipole anomaly with enhanced rainfall in the eastern Indian Ocean and the Maritime Continent, and reduced rainfall over the equatorial central and eastern Pacific. Over East Asia, below normal rainfall is found in Japan, southern Korea, and the central Pacific over the warm water, and above normal rainfall is found in northern China and a region just inland of central China. The rainfall and circulation patterns suggest a westward shift of the western Pacific anticyclone, and a weakening of the mei-yu (baiu or changma) rainbelt. The former is a well-known large-scale control of the climate of East Asia (Lau et al. 2000; Wang et al. 2001). The overall concentration of rainfall anomalies in the Indo-Pacific region suggests that this mode is associated with a redistribution of heat sources and sinks within the Asian monsoon region. The anomalies in the Asian monsoon region occur simultaneously with reduced rainfall over the North Pacific Ocean, the Pacific Northwest, western Canada, and the U.S. Midwest, and above normal rainfall over eastern and northeastern North America. The anomalous rainfall patterns in the subtropics and midlatitudes generally coincide with the shifts of the storm tracks as shown in Fig. 4b, suggesting the reduction in rainfall over the North Pacific warm pool may be due to the weakening and the displacement of the storm tracks by the large-scale anticyclones. Since both the North Pacific and North Atlantic anticyclones overlie warm SSTs, it is likely that the extratropical SST anomaly is initially forced by the atmosphere, with the ocean warming up via increased surface solar radiation and reduced surface evaporation from diminished transients, and warm advection (Lau and Nath 2001). It is also possible that the northward displacement of the storm track may lead to increased cloudiness, increased evaporation, and therefore cooling of the ocean to the north (Norris and Weaver 2001). Once the large-scale SST anomaly pattern is set up, it is possible that the altered SST gradients will feedback on the transient and in turn reinforce the atmospheric circulation. More likely, the SST distribution in the North Pacific and the North Atlantic may act in concert with heating over Eurasia or North America to amplify preexisting atmospheric anomalies and anchor them to preferred geographic locations.
b. Associations with the Asian monsoon
In this section, we discuss the association of the aforementioned circulation pattern with the Asian monsoon. To emphasize the distinction between the East Asian and the South Asian monsoons, various regional monsoon indices have been developed in a number of recent studies (Goswami et al. 1999; Wang and Fan 1999; Lau et al. 2000). Based on the observation that the rainfall variability over the East Asia continent and the tropical western Pacific is well correlated with the north–south migration of the East Asian subtropical jet, Lau et al. (2000) used the 200-mb zonal wind difference between the midlatitude region (40°–50°N, 110°–150°E) and the subtropics (25°–35°N, 110°–150°E) to define an index for the East Asian monsoon. This index was referred to as the Regional Monsoon Index 2 (RM2). Wang and Fang (1999) defined the Monsoon Circulation Index 1 (MCI1) for the Indian monsoon as the westerly vertical wind shear in the region (5°–20°N, 40°–80°E), based on the maximum signal of the Rossby wave response to heating over the Bay of Bengal. These two indices are representative of the variability of the East Asian and South Asian regional monsoons, respectively. To compare the associated teleconnection patterns, Fig. 6 shows the pattern of linear regression of 850-mb wind and 500-mb geopotential height against NPI, RM2, and MCI1, respectively. Regions with geopotential height anomalies whose correlation exceeds the 95% significant level are shaded. Only significant wind vectors are shown. As expected, the regression pattern for NPI (Fig. 6a) is similar to the composite pattern shown in Fig. 5, with the most significant anticyclonic features found over the North Pacific and the North Atlantic. As noted before, significant wind features are noted in the equatorial central Pacific and over India. The similarity of NPI to the RM2 pattern (Fig. 6b) is rather striking, indicating that the NPI mode has a strong association with the East Asian monsoon through the fluctuation of the East Asian jet. Notice the lack of a tropical signal in RM2, in contrast to the NPI. The teleconnection pattern generated from MCI1 (heating over the Bay of Bengal) bears some resemblance to that of NPI especially in the Tropics, with anomalies in the subtropics and midlatitudes generally of the same sign, but with a considerably weaker signal compared to RM2. The above results suggest that the NPI pattern is strongly linked to the variability of the heat sources and sinks in the Indo-Pacific region, with the East Asian monsoon contributing to a large part of the extratropical signal and the South Asian monsoon accounting for much of the tropical signal.
5. Intercontinental climate teleconnection
a. Temperature anomalies
To show the extent of the climate regulation by the North Pacific SST anomaly on the overlying atmosphere, Fig. 7 shows the vertical cross section of the composites of the tropospheric temperature averaged along 40°–45°N with respect to NPI during JJA. It can be seen that the warm North Pacific SST anomaly is associated with a positive tropospheric temperature anomaly that extends from the surface to 200 mb, with an anomaly of opposite sign in the lower stratosphere (∼100 mb). The maximum tropospheric signal is about 1.5°C. The temperature anomaly tilts westward with height, and appears to be connected to a major warm anomaly in the upper troposphere over Eurasia (10°–80°E). Also noticeable is the above normal tropospheric temperature anomalies along the west and east coasts of North America and a relative cold dome over the central part of the continent. All the aforementioned tropospheric temperature anomalies appear to be connected to comparable temperature anomalies, but with the opposite sign in the lower stratosphere, indicating the presence of deep vertical modes. The temperature cross section associated with NPI may be identified with summertime stationary wave patterns driven by differential continental and oceanic thermal heating in the Tropics and midlatitudes, and/or modulated by the interaction of the subtropical jet stream with orographic forcings generated by the Tibetan Plateau and the Rockies (Ting 1994; Chen 1993).
The horizontal distribution of CRU surface temperature anomalies associated with NPI shows widespread warming over Eurasia and North America, except near eastern Siberia and Alaska, where below normal temperatures are found (Fig. 8). There is good correspondence between surface warm (cold) anomalies and high (low) 500-mb geopotential height anomalies (see Fig. 5). The colder polar regions and warmer midlatitudes may also be related to the polar shift of the storm tracks noted in the discussion of Fig. 4. Increased cloudiness, and cold advection associated with high latitude storms, will lead to lower surface temperatures, while clear-sky conditions associated with anticyclone development will increase surface heating, and produce higher SSTs and air temperatures. The regions of positive surface temperature anomalies seem to coincide broadly with regions of positive 500-mb height deviations (see Fig. 5a), indicating that these are vertically extended hydrostatic anomalies. Interestingly, the land regions of the Tropics, for example, India, Indonesia, and northern South America, show below normal temperatures, and they all tend to coincide with regions of above normal rainfall (see also Fig. 5b). Over Asia, the warmest regions are concentrated over northeastern China and Japan. Another major warm area is found over northern Siberia and central Asia. Over North America, the warmest region is the Pacific Northwest and western Canada, as well as the northeastern region. The U.S. Midwest and Southeast are slightly cooler than normal. The surface temperature signal is consistent with those evident in the vertical cross section shown in Fig. 7.
b. Continental rainfall anomalies
To better delineate the relationship between the rainfall and circulation over East Asia and North America associated with NPI, composites of the land-based rainfall anomalies from the CRU rainfall data with respect to NPI for the two continents and for the entire data record (eight cases) are shown in Fig. 9. Here, the overall rainfall pattern is broadly similar to that from the CMAP data for 1979–99. The maximum phase of the NPI (warm North Pacific) is associated with below normal rainfall over Japan, South Korea, and the eastern portion of the Yangtze River valley (YRV), and above normal rainfall over much of continental East Asia and South Asia, except the eastern Bay of Bengal near Myanmar, where rainfall is below normal (Fig. 9a). The dry condition along the coastal region is associated with anomalous low-level easterlies extending from the North Pacific anticyclone across Japan and South Korea to the east coast of China. The anomalous easterlies oppose the climatological westerly flow over the region, reduce local evaporation, and block off the low-level southwesterly monsoon flow and, hence, deprive the region of warm, moist air from the South China Sea and the East China Sea. As suggested earlier, the rainfall anomalies may be associated with a westward expansion of the subtropical high toward the continental regions of East Asia, resulting in a weakened mei-yu rainbelt. In conjunction with the enhanced rainfall over the land regions of Asia and South Asia, anomalous low-level monsoon westerlies are found over the India Ocean region and across the Indian subcontinent, signaling a strengthening of the South Asian monsoon. During the maximum phase of NPI, the monsoon trough, indicated by the line of cyclones over northeastern India, the northern South China Sea, and the subtropical western Pacific, is active at 20°N. At higher latitude (>40°N), the anticyclone appears to be coupled to a low over the Sea of Okhotsk, which appears to be a part of a wave train spanning northern Eurasia. The feature at high latitudes is consistent with recent results of Samel and Liang (2003), who show that blocking events in the Sea of Okhotsk may be important in triggering onset of heavy rain over the eastern YRV region.
The counterpart of the positive NPI teleconnection over North America (Fig. 9b) features drier condition over the Pacific Northwest and western Canada, the northern Great Plains, and the Midwest, while enhanced rainfall is found over the northeast United States, the Atlantic coast, the southwest United States, and Mexico. In the Pacific Northwest, the anticyclonic flow during the positive phase of the NPI brings cold dry air from the Gulf of Alaska, which would suppress rainfall over the region. The influence of the North Pacific anticyclones appears to extend north of the Rockies and comes down from central Canada via the northern Great Plains to the Gulf coast. The northerly low-level flow along the Great Plains (∼100°W) opposes the climatological low-level jet, thereby inhibiting warm moist air from the Gulf coast from reaching the Midwest and upper Great Plains, and suppresses rainfall there (Hu and Feng 2001). The wet conditions along the Atlantic coast stem from the combined effects of the cyclone over southeast United States and the Atlantic anticyclone, which regulate warm moist air from the subtropical western Atlantic to the east coast of North America.
Comparing the circulation and rainfall patterns over Asia and over North America, some similarities can be discerned in the relative distribution of the cyclonic and anticyclonic circulation anomalies, as labeled in Fig. 9. The major contrast is that the Pacific anticyclone extends farther inland over the Asian sector compared to North America. As a result, regions to the north of 30°N in East Asia are shielded from tropical influences. In contrast, the east coast of the United States is under tropical influence from the cyclonic circulation in the southeast United States. Over the Indo-Pacific region, the monsoon trough is found near 20°N, with well-defined centers of cyclonic activity, compared to that over the North America sector. The equatorward meridional flow over central North America stemming from the cyclonic circulation over the southeastern United States occurs simultaneously with the North Pacific anticyclone. No corresponding feature is found over central Asia. Perhaps the feature that corresponds best to the southeast cyclone in North America is the cyclonic flow over the South China Sea. However, the South China Sea cyclone is mostly confined to tropical influence. Another notable contrast is that the wave train pattern in the high latitudes linking the Pacific anticyclone appears to be more pronounced than that for the Atlantic anticyclone. In that regard, blocking (or lack thereof) over the Okhotsk Sea may have more of an impact on summertime rainfall anomalies over Japan, Korea, and northeastern China, compared to a similar but weaker feature over the Baffin Bay on rainfall over for the Atlantic coast of North America.
c. Temperature and rainfall distributions
During the summer of 1993, the North Pacific SST was abnormally cold, corresponding to a minimum NPI phase, and a disastrous flood occurred over the U.S. Midwest, in conjunction with above normal rainfall over the Pacific Northwest, South Korea, and Japan (Lau and Weng 2002). In contrast, extreme hot and dry conditions over Japan, Korea, and northeastern China during the summer of 1994 were associated with extensive warm SSTs over the North Pacific, and drier than normal conditions along the west coast and in the Midwest of the United States (Park and Schubert 1997). These two years corresponded to opposite extremes of the NPI. Obviously, summertime rainfall extremes over regions in East Asia and in North America can also occur independent of or in the absence of substantial North Pacific SST anomalies. While the North Pacific SSTs certainly do not dictate the occurrence of individual extreme events in either continents, they may have an effect in shifting the distribution of surface temperature and rainfall over specific regions of the two continents, resulting in changes in the statistics of extreme events. To test this idea, we have computed the cumulative probability distribution of precipitation and temperature based on pentad station data for select regions in the two continents.
Shown in Fig. 10 are the cumulative probability distributions with respect to the normalized (by standard deviation) temperature and rainfall anomalies for Japan/South Korea and northwestern North America, for the maximum and minimum phases of NPI respectively. Over Japan/South Korea, the separation in the temperature distribution between the two phases is very pronounced. For maximum NPI, the cumulative probability of below normal temperature is 40% compared with 80% for the minimum NPI. Equivalently, there is a 60% chance of above normal temperature for maximum NPI, compared to 20% chance for minimum NPI. In the extreme category, the probability of the temperature being one standard deviation or more above normal is approximately 30% for positive NPI, but less than 3% for minimum NPI. The shift in the rainfall distribution is more subtle, indicating only a slight increase in probability of below normal rainfall during maximum NPI (75%) from the minimum NPI phase (70%). Likewise, over the Pacific Northwest, the probability of above normal temperature increases from 30% to 42%, and the probability of below normal rainfall drops from 80% to 70% in the maximum relative to the minimum NPI. The statistics suggest warmer/drier weather on both sides of the Pacific along 40°N during the maximum NPI, and colder/wetter conditions for the minimum NPI.
Next, we test the null hypothesis that the distributions are not distinguishable between the positive and the negative phases of NPI using the chi-square test. The results are shown in Table 1. It can be seen that the probability of exceeding the chi-square function by chance (last column of Table 1) is practically zero for the surface temperature in Japan/Korea. This means that one can safely reject the hypothesis that the temperature difference between the NPI's positive and negative phases arises by chance. For temperature over northwestern North America, the probability of chance occurrence is 22%. For rainfall, the high probabilities indicate that there are not enough cases to reject the null hypothesis. To buttress the above results, we apply the Kolmogorov–Smirnov test to determine if the distributions during the maximum and minimum phases of NPI are significantly different from the total distribution. Following a procedure similar to Sardeshmukh et al. (2000), we determine that the maximum difference is 0.11 for the 95% significant level for a sample of 144 (eight cases with 18 pentads). Table 2 shows that for Japan/Korea, the temperature in either the positive (warm) or the negative (cold) NPI phase is significantly different from that of the total distribution. For northwestern North America, only the cold phase seem to be well separated. In general agreement with the chi-square test, the pentad rainfall distributions are not significantly distinct.
6. Concluding discussions
Using NCEP–NCAR reanalysis wind, sea surface temperature, and rainfall data, we have identified two summertime interannual climate modes in the North Pacific Ocean that are distinct from El Niño. The first mode features a warm or cold anomaly covering most of the North Pacific around 30°–50°N, and the second mode is a dipolelike anomaly between the northern and southern regions of the North Pacific. Both modes are linked to coherent SST variations in the North Atlantic suggesting the presence of an atmospheric bridge linking the two extratropical oceans. While the first mode may have some linkage to El Niño during the antecedent spring, its subsequent evolution suggests that it has indigenous signals in the North Pacific that are quite distinct from those of El Niño. The second mode is largely independent of El Niño.
Results from composite and regression analyses show that the positive (negative) phase of the NPI is associated with a weakening (strengthening) of the subtropical jet stream and northward (southward) shift of storm tracks associated with a large-scale anticyclone (cyclone) over the warm (cold) water of the North Pacific and North Atlantic. Over the western North Pacific, the large-scale structure resembles that associated with the fluctuation of the East Asian summer monsoon. The North Atlantic anomalous anticyclone appears to influence the summertime climate of the east coast of North America and the U.S. Midwest region. The vertical scale of the temperature anomalies over the North Pacific extends over the entire troposphere, including anomalies of the opposite sign in the lower stratosphere.
The NPI pattern has significant impacts on summertime climate over the entire Northern Hemisphere. During the maximum phase of the NPI (warm North Pacific), the surface temperatures are above normal over Japan/Korea, northeastern China and much of the Eurasian land region, and the North America continent, while below normal temperatures are found over the polar regions including eastern Siberia and Alaska, and the monsoon regions of the Indo-Pacific, central America, and northern South America. For maximum NPI, below normal rainfall is found over Japan/Korea, and regions along the east coast of East Asia, in conjunction with a rainfall deficit in western Canada, the northwest United States, and the northern Great Plains, but a wetter Atlantic coast. The reverse is found for the minimum NPI phase. Statistical analyses of pentad temperature distributions confirm a shift toward warmer climate, and more frequent occurrence of extreme hot weather over Japan/Korea and northwestern North America, but colder climate, and more frequent occurrences of extreme cold spells, over eastern Siberia and Alaska. However, there is not enough data to conclude a shift in the distribution of extreme rainfall events in pentad rainfall distributions.
The presence of the summertime interannual climate modes associated with precursory and contemporaneous extratropical SST variation may provide an explanation of the enhanced potential predictability of the seasonal rainfall over North America from the North Pacific SST noted by Lau et al. (2002). The key question is What cause and sustain the large-scale atmosphere and ocean anomalies in the North Pacific? The anticyclonic (cyclonic) circulation over warm (cold) water suggests that to a first-order approximation the North Pacific SST is likely to be forced by the atmosphere. However, the coherence and the longevity of the atmospheric and oceanic anomalies suggest that air–sea interaction may be needed to amplify and sustain the anomalies. Figure 11 depicts a plausible extratropical air–sea feedback mechanism that can be inferred from the present results. An anticyclone over the North Pacific is associated with suppressed rainfall and cloudiness, as well as a less disturbed atmosphere with reduced surface wind fluctuations. These conditions will cause the ocean underlying the anticyclone to warm by increased solar radiation and reduced evaporation. Likewise, the increased cloudiness and storminess associated with the northward shift of the jet stream and the enhancement of the storm track will cause cooling of the ocean to the north. This creates a zone of enhanced meridional SST gradient at the northern edge of the anticyclone, which increases the baroclinicity of the atmosphere. The increased baroclinic zone will further increase (decrease) in jet stream strength and transient eddies to the north (south), reinforcing the north–south SST gradient. As a result, the anticyclone grows and sustains itself by a positive feedback via surface fluxes and transient–mean flow interactions. The above scenario is consistent with the observation of enhanced stratocumulus cloud decks over the cold water of the northern North Pacific and the reduced cloudiness over the warm water to the south (Norris and Weaver 2001).
Our results demonstrate that air–sea interaction associated with the extratropical oceans during the boreal summer is fundamentally different from that in the Tropics, involving the interplay of surface fluxes, atmospheric mean circulation and transient eddies, and possibly radiation effects of stratocumulous cloud decks. The interaction occurs at the time when the influence of the tropical ocean is weakest. In addition to the extratropical air–sea interaction, it is also possible that the North Pacific SST anomalies alter the surface temperature contrasts between land and oceans during spring and summer, which in turn modulate the atmospheric stationary waves. Moreover, positive feedbacks between these climate modes and land surface processes may lead to a regional amplification over the continental regions of East Asia and North America, causing severe droughts and floods over these regions.
Acknowledgments
This work is supported by the Global Modeling and Analysis Program, NASA Office of Earth Sciences. Partial support is also provided to the second and fourth authors by the Climate Environment System Research Center at Seoul National University, through the Korea Science and Engineering Foundation and Brain Korea 21 Program.
REFERENCES
Alexander, M. A., I. Bladé, M. Newman, J. R. Lanzante, N. C. Lau, and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15 , 2205–2231.
Cayan, D. R., A. J. Miller, T. P. Barnett, N. E. Grahm, J. N. Ritchie, and J. M. Oberhuber, 1995: Seasonal-to-interannual fluctuations in surface temperature over the Pacific: Effects of monthly winds and heat fluxes. Natural Climate Variability on Decadal-to-Century Time Scales, D. G. Martinson et al., Eds., National Academy Press, 133–150.
Chen, P., and M. Newman, 1998: Rossby wave propagation and the rapid development of upper-level anomalous anticyclones during the 1988 U.S. drought. J. Climate, 11 , 2491–2504.
Chen, T. C., 1993: Interannual variation of summertime stationary eddies. J. Climate, 6 , 2263–2277.
Deser, C., and M. L. Blackmon, 1995: On the relationship between tropical and North Pacific sea surface temperature variations. J. Climate, 8 , 1677–1680.
Gershunov, A., and T. P. Barnett, 1998: Interdecadal modulation of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79 , 2715–2726.
Goswami, B. N., V. Krishmurthy, and H. Annamalai, 1999: A broad scale circulation index for the interannual variability of the Indian summer monsoon. Quart. J. Roy. Meteor. Soc., 125 , 611–633.
Hamming, R. U., 1983: Digital Filters. 2d ed. Prentice Hall, 257 pp.
Hu, Q., and S. Feng, 2001: Climatic role of the southerly flow from the Gulf of Mexico in interannual variations in summer rainfall in the central United States. J. Climate, 14 , 3156–3170.
Kawamura, R., 1994: A rotated EOF analysis of global sea surface temperature variability with interannual and interdecadal scales. J. Phys. Oceanogr., 24 , 707–715.
Kodama, Y., 1992: Large-scale common features of subtropical precipitation zones (the Baiu Frontal Zone, the SPCZ and the SACZ). Part I: Characteristics of subtropical frontal zones. J. Meteor. Soc. Japan, 70 , 813–835.
Kodama, Y., 1993: Large-scale common features of subtropical convergence zones (the Baiu Frontal Zone, the SPCZ and the SACZ). Part II: Conditions of the circulations for generating the STCZs. J. Meteor. Soc. Japan, 70 , 581–610.
Kushnir, Y., and N. C. Lau, 1992: The general circulation model response to a North Pacific SST anomaly: Dependence on time scale and pattern polarity. J. Climate, 5 , 271–283.
Lau, K-M., and L. Peng, 1992: Dynamics of atmospheric teleconnection during the northern summer. J. Climate, 5 , 140–158.
Lau, K-M., and H. Weng, 2000: Remote forcing of US summertime droughts and floods by the Asian monsoon? GEWEX News, Vol. 10, No. 2, 5–6.
Lau, K-M., and H. Weng, 2002: Recurrent teleconnection patterns linking summertime precipitation variability over East Asia and North America. J. Meteor. Soc. Japan, 80 , 1129–1147.
Lau, K-M., K-M. Kim, and S. Yang, 2000: Dynamical and boundary forcing characteristics of regional components of the Asian summer monsoon. J. Climate, 13 , 2461–2482.
Lau, K-M., K-M. Kim, and S. Shen, 2002: Potential predictability of seasonal precipitation over the United States from canonical ensemble correlation predictions. Geophys. Res. Lett.,29, 1097, doi:10.1029/2001GL014263.
Lau, N. C., and M. J. Nath, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate, 9 , 2036–2057.
Lau, N. C., and M. J. Nath, 2001: Impact of ENSO on SST variability in the North Pacific and North Atlantic: Seasonal dependence and role of extratropical air–sea interaction. J. Climate, 14 , 2846–2866.
Livezey, R. E., and W. Y. Chen, 1983: Statistical field significance and its determination by Monte Carlo techniques. Mon. Wea. Rev., 111 , 46–59.
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.
Mo, K. C., 2000: Intraseasonal modulation of summer precipitation over North America. Mon. Wea. Rev., 128 , 1490–1505.
Mo, K. C., J. R. Zimmerman, E. Kalnay, and M. Kanamitsu, 1991: A GCM study of the 1988 United States drought. Mon. Wea. Rev., 119 , 1512–1532.
Namias, J., 1959: Recent seasonal interactions between North Pacific waters and the overlying atmospheric circulation. J. Geophys. Res., 64 , 631–646.
Namias, J., 1976: Seasonal forecasting experiments using North Pacific air–sea interactions. Bull. Amer. Meteor. Soc., 57 , 163–165.
Namias, J., 1991: Spring and summer 1988 drought over the contiguous United States—Causes and prediction. J. Climate, 4 , 54–65.
New, M., M. Hulme, and P. Jones, 2000: Representing twentieth-century space–time climate variability. Part II: Development of 1901–96 monthly grids of terrestrial surface climate. J. Climate, 13 , 2217–2238.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65 , 373–390.
Norris, J. R., and C. P. Weaver, 2001: Improved techniques for evaluating GCM cloudiness applied to the NCAR CCM3. J. Climate, 14 , 2540–2550.
Overland, J. E., N. A. Bond, and J. M. Adams, 2001: North Pacific atmospheric and SST anomalies in 1997: Links to ENSO? Fish. Oceanogr., 10 , 69–80.
Park, C-K., and S. D. Schubert, 1997: On the nature of the 1994 East Asian summer drought. J. Climate, 10 , 1056–1070.
Peng, S., W. A. Robinson, and M. P. Hoerling, 1997: The modeled atmospheric response to midlatitude SST anomalies and its dependence on background circulation states. J. Climate, 10 , 971–987.
Pitcher, E. J., M. L. Blackmon, G. T. Bates, and S. Munoz, 1988: The effect of North Pacific sea surface temperature anomalies on the January climate of a general circulation model. J. Atmos. Sci., 45 , 173–188.
Reynolds, R., and T. M. Smith, 1994: Improved global sea surface temperature analysis using optimum interpolation. J. Climate, 7 , 929–948.
Samel, A. N., and X-Z. Liang, 2003: Understanding relationship between the 1998 Yangtze River flood and northeast Eurasian blocking. Climate Res., 23 , 149–158.
Sardeshmukh, P. D., G. P. Compo, and C. Penland, 2000: Changes of probability associated with El Niño. J. Climate, 13 , 4268–4286.
Schubert, S. D., M. J. Suarez, P. J. Pegion, M. A. Kistler, and A. Kumar, 2002: Predictability of zonal means during boreal summer. J. Climate, 15 , 420–434.
Ting, M., 1991: The stationary wave response to a midlatitude SST anomaly in an idealized GCM. J. Atmos. Sci., 48 , 1249–1275.
Ting, M., 1994: Maintenance of northern summer stationary waves in a GCM. J. Atmos. Sci., 51 , 3286–3308.
Ting, M., and H. Wang, 1997: Summertime U.S. precipitation variability and its relation to Pacific sea surface temperature. J. Climate, 10 , 1853–1873.
Trenberth, K., and J. W. Hurrell, 1994: Decadal atmospheric–ocean variations in the Pacific. Climate Dyn., 9 , 303–319.
Trenberth, K., G. Branstator, and P. Arkin, 1990: Origins of the 1988 North American drought. Science, 242 , 1640–1645.
Wang, B., and Z. Fan, 1999: Choice of South Asian summer monsoon indices. Bull. Amer. Meteor. Soc., 80 , 629–638.
Wang, B., R. Wu, and K. M. Lau, 2001: Interannual variability of the Asian summer monsoon: Contrasts between the Indian and the western North Pacific–East Asian monsoon. J. Climate, 14 , 4073–4090.
Xie, P., and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78 , 2539–2558.
Zhang, Y., J. M. Wallace, and D. A. Battisti, 1997: ENSO-like interdecadal variability; 1900–1993. J. Climate, 10 , 1004–1020.
Standard deviation of interannual variability (<9 yr) of seasonal mean SST for (a) spring, (b) summer, (c) fall, and (d) winter. Units are °C
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Spatial patterns of dominant modes of North Pacific SST for (a) EOF1, (b) EOF2, and for North Pacific and North Atlantic SST for (d) EOF 1 and (e) EOF2. Unit used is nondimensional, and contour interval is 0.02. Typical magnitude of SST variation (obtained by multiplying EOF loadings with time coefficients) associated with this mode is of the order of 0.5°–1°C. The lagged correlation of the first and second principal components, denoted by solid and dashed lines (c) for the North Pacific only and (f) for the North Pacific and the North Atlantic. The correlation coefficients at the 95% and 99% significance levels are 0.34 and 0.42, respectively
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Lagged correlation SST pattern associated with mode-1 PC for (a) DJF (−2 seasons), (b) MAM (−1 season), (c) JJA (0 season), (d) SON (1 season), (e) DJF (2 seasons), and (f) MAM (3 seasons). Correlation coefficients over the 95% (99%) significance level are lightly (darkly) shaded
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Mode-1 composite of (a) 200-mb zonal wind anomalies and (b) 200-mb transients. Units are m s−1. Areas with correlations exceeding 95% (99%) significance are shaded
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Mode-1 composite of (a) 500-mb height and 850-mb wind for 1950–99, and (b) CMAP rainfall and 850-mb wind for 1979–99. Units for geopotential height are m, for rainfall are mm day−1, and for wind are m s−1
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Linear regression of 850-mb wind (m s−1) and 500-mb geopoential height (m) against (a) NPI, (b) RM2, and (c) MCI1. Regions exceeding the 95% confidence level are shaded
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Longitude–height cross section of tropospheric temperature anomaly averaged across 40°–45°N with respect to NPI during JJA. Areas exceeding 1 σ are shaded
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Mode-1 composite of CRU surface temperature anomalies. Areas with correlations exceeding 95% significance level are stippled. Units are °C
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Mode-1 composite of CRU rainfall and 850-mb wind anomalies over (a) the Asian–Pacific region and (b) North America. Units for rainfall are mm day−1 and for wind are m s−1. Letters A and C denotes the centers of the anticylone and the cyclone referred to in the text
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Cumulative probability distribution of temperature and rainfall (a), (b) for Japan/South Korea region and (c), (d) for northwestern North America. Temperature and rainfall anomalies are normalized by standard deviation
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Schematic showing a plausible extratropical air–sea interaction and transient-mean flow feedback mechanism during boreal summer. Symbols E and SW stands for evaporation and surface shortwave radiation, respectively
Citation: Journal of Climate 17, 4; 10.1175/1520-0442(2004)017<0819:TNPAAR>2.0.CO;2
Chi-square values and probability based on binned pentad temperature and rainfall distributions associated with the positive and negative phases of NPI
Maximum difference in the Kolmogorov–Smirnov test for pentad temperature and rainfall distributions associated with positive and negative phases of NPI. Differences exceeding 95% significance are set in boldface
