1. Introduction
The North American summer monsoon is a complex system and consists of several subregions of coherent rainfall variability in the Southwest and northern Mexico (Comrie and Glenn 1998). In addition, the distinct monsoon circulation is part of a large-scale circulation in central and western North America and the eastern North Pacific. The monsoon is established in northern Mexico and the Southwest in late June and early July and persists through September (e.g., Higgins et al. 1998, 1999; Ropelewski et al. 2005). During the monsoon season the atmospheric circulation in the region favors frequent and intense rainfall development, making the season a wet period in the year. Because of an essential role of this wet period in the development of vegetations and sustaining the ecosystems and socioeconomics of the region, many studies have examined the monsoon for its onset, breaks, and interannual variations (e.g., Higgins et al. 1999; Yu and Wallace 2000). Some aspects of these processes are reviewed in Vera et al. (2006). The interannual variation of the monsoon was found to be related to the antecedent winter precipitation and the following spring land surface water and energy flux anomalies in both the northwestern and southwestern United States (Gutzler and Preston 1997; Higgins and Shi 2000; Hu and Feng 2002; Zhu et al. 2005). These relationships have further been found varying considerably at multidecadal time scales (Hu and Feng 2002, 2004; Zhu et al. 2005): they were strong in the decades of 1961–90, yet disappeared in the decades of 1931–60 and from 1990 onward [see Fig. 12d in Hu and Feng (2002)]. This variation indicates the presence of different regimes of the North American summer monsoon circulation and alternation of such regimes at multidecadal time scales.
In a recent modeling study, Hu and Feng (2007) showed two different regimes of atmospheric circulation in western and central North America during its summer season. Additionally, their analyses of model simulations depicted systematic changes of the regimes from the decades of 1961–90 to 1991–2005. In 1961–90, the average geopotential height has an anomaly pattern very different from that in 1991–2005. The geopotential height anomalies in 1961–90 were such that they engaged easterly and southeasterly winds from the Gulf of Mexico with the moisture anomalies in the southwestern United States and northern Mexico in the monsoon season. In the regime of 1991–2005, the geopotential height anomalies engaged a different moisture flow from the Gulf of California with the circulation in the monsoon region to influence the monsoon rainfall variation. These modeling results show that different low-level flows and moisture transport processes developed and affected the monsoon rainfall variations in the two different regimes, and are supported by observations.
While these modeling results support the notion that different large-scale circulation regimes are present and responsible for the observed change of major sources and moisture flows affecting the monsoon rainfall, they raise the following questions: 1) Could similar alternation of the circulation regimes have occurred in the decades prior to 1960? If so, the alternation could be an internal variation of the large-scale circulation in central and western North America, and may continue in future decades and influence the summer monsoon and precipitation. 2) What are the major circulation anomalies facilitating the coherent monsoon circulation and precipitation regimes? 3) What may have been the possible causes for such anomalies and the alternation of the monsoon regimes? These questions and some related issues are addressed in this study using available observational data.
The data used will be described in the next section (section 2) and different large-scale circulation and monsoon regimes in Mexico and the central and western United States are articulated in section 3. It will be shown that the alternation of these regimes also occurred in the decades before 1960, a result suggesting a persistent oscillation (though irregular) of the North American summer monsoon regimes in the twentieth century. After defining the regimes we examine their specifics and evaluate monsoon processes in the different regimes. Results of these evaluations also are presented in section 3. In section 4, a hypothesis for the cause of the multidecadal variations of the monsoon regimes being the Atlantic Multidecadal Oscillation (AMO) is proposed and tested. A summary of this study is given in section 5.
2. Data
Daily precipitation data in gridded format with resolution of 1.0 × 1.0 degrees of latitude and longitude in the United States and Mexico were obtained from the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC; Higgins et al. 2000). In addition, the monthly global gridded precipitation with resolution of 0.5 × 0.5 degrees of latitude and longitude was obtained from the Climatic Research Unit (CRU) at the University of East Anglia (New et al. 2000). Data from CRU cover the period 1901–2002 and data from the NOAA CPC are from 1948 to 2005. Because the CPC data in Mexico are not reliable before 1960 (Gutzler 2004), they are used for the period of 1961–2005. The much longer records in the CRU dataset are suitable for analysis of the multidecadal variations in the monsoon rainfall and regime change. Before applying the CRU data, we evaluated them by comparing them to the data from NOAA CPC. Time series of total precipitation from these two datasets are shown in Fig. 1 for July–September (JAS) averaged in the west Mexican monsoon region (24°–30°N, 107°–112°W, which is defined as the core region of the North American monsoon; Gutzler 2004) and the area of Arizona and New Mexico (AZNM; 32°–36°N, 107°–112°W).
Figure 1a shows that there is a quite a large discrepancy between the two datasets for the years after 1987 in western Mexico. The two datasets are fairly consistent in AZNM (Fig. 1b) and are consistent in the years before 1987 in western Mexico (Fig. 1a). Investigations of this discrepancy in CRU data revealed the source, that is, there was a sharp decrease in the number of stations in western Mexico after 1987 (see the histogram in Fig. 1). Apparently, the remaining scattered stations taking precipitation observations after 1987 were unable to capture the warm season rainfall in this region of complex terrain (Gochis et al. 2007b), resulting in the discrepancy after 1987. An extended investigation of changes in the number of stations for the entire data period from 1901 to 2002 showed a similar lack of an adequate number of stations in western Mexico in the years before 1925 (Fig. 1). Thus, the CRU precipitation data from 1925 to 1987 are used in conjunction with the NOAA CPC data in this study to examine variations of the North American summer monsoon and precipitation.
In addition to precipitation, monthly data of atmospheric geopotential heights and winds were obtained from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis from 1948 to 2005 (Kalnay et al. 1996). Global sea level pressure (SLP) and sea surface temperature (SST) data were obtained from the Hadley Centre Sea Level Pressure (HadSLP2; Allan and Ansell 2006) and the Global Sea Ice and SST (HadISST1) datasets (Rayner et al. 2003), respectively, at the Met Office Hadley Center for the period from 1870 to 2005.
3. Variations of the North American summer monsoon regimes
Because the North American monsoon is a complex system and has areas of subregional coherence it is necessary to consider relationships of the submonsoons and their connections with circulations in the surrounding regions when we depict the monsoon regimes and variations. From this perception, we examined monsoon relationships and used them to describe the North American monsoon regimes and variations. Results from these examinations are shown in Fig. 2, which shows the correlations of the July–September rainfall in the west Mexico (WM) monsoon region and rainfall at each grid point in the land areas of the domain.
In Fig. 2a, the correlation pattern indicates a strong and coherent relationship (statistically significant at 95% confidence level) between rainfall in the west Mexican monsoon region and in the Southwest, strongest in Arizona, for the period from 1961 to 1990. Meanwhile, rainfall in the west Mexican monsoon region has little connection with rainfall in east Mexico. Another important feature in Fig. 2a is the simultaneous but negative correlation of the west Mexican monsoon rainfall and rainfall in the central United States (CUS).
The correlation pattern for precipitation in west Mexico and the central and western United States changed in the period of 1991–2005. In this new pattern shown in Fig. 2b, rainfall in west Mexico varies consistently with rainfall in parts of east Mexico, but has a negative and insignificant correlation with rainfall in the Southwest. A rather striking feature in this change is the sharp reverse of the correlation from west Mexico to the Southwest, suggesting strong and different local circulations influencing rainfall in these two regions during this regime. Another noteworthy feature in Fig. 2b is that relative to rainfall variations in west Mexico, the rainfall in the Southwest still has a negative, though insignificant, relationship with rainfall in the central United States.
Extending the analysis to the decades before 1960 using the CRU data, we found that the rainfall correlation pattern for the decades of 1931–60 (Fig. 2c) is similar to that for 1991–2005, except for a few details (partially because of the different resolution of the two datasets and relatively limited number of stations in the early observation systems in Mexico; see Fig. 1). The spatial coherence of rainfall in Mexico and the Southwest and central United States in the different decades and its change between the epochs, shown in Figs. 2a–c, depicts a multidecadal variation/alternation of two different regimes of the North American summer monsoon and precipitation in the twentieth century.
From Fig. 2, we hypothesize that the multidecadal variation of the monsoon regimes is caused by changes in circulation such that it organized the rainfall in different subregions and created special rainfall patterns in the North American monsoon region. One regime corresponded to circulations that favored simultaneous rainfall development in west Mexico and the Southwest, but discouraged rainfall in east Mexico and also suppressed rainfall in the central United States, or vice versa. The other regime corresponded to different circulations that had similar influences on rainfall in west and east Mexico but opposite effects on rainfall in the Southwest, or vice versa. In this pattern, rainfall variation in the Southwest was utterly different from that in west Mexico. These two regimes have alternated their dominance of the North American summer monsoon and precipitation variations in the last century.
To understand these two circulation regimes and their alternation at multidecadal time scale, we show in Figs. 3 and 4 the composites of the circulation anomalies in wet and dry years in 1961–90 and in 1991–2005, respectively. The wet (dry) is defined as the top (bottom) 25% of rainfall anomalies in the west Mexican monsoon region relative to its mean July–September rainfall of each regime. Years with July–September rainfall with the top 25% anomaly for the regime were averaged to obtain the circulation for the wet years, and years with the rainfall in the bottom 25% anomaly for the regime were averaged to obtain the circulation for the dry years. Figures 3 and 4 show the 500-hPa geopotential height anomalies and the 850-hPa wind anomalies for the different regimes. These anomalies also were calculated relative to the mean of each regime. They describe the circulation features corresponding to the above- or below-average rainfall of the monsoon months in those regimes. These different circulation features compose the circulation regimes with the previously described coherent correlation patterns of rainfall in the monsoon region.
Contrasting the circulation anomalies between the wet and dry monsoon years in the regime of 1961–90 (Fig. 3a versus 3b), we find a reversal of the circulation anomaly pattern. Specifically, positive geopotential height or anticyclonic circulation anomalies are shown in the western two-thirds of the United States and Canada for wet monsoon years (Fig. 3a). Weak negative geopotential or cyclonic anomalies in the west half of the United States correspond to dry years (Fig. 3b).
In accordance with this reversal in geopotential heights and circulation anomalies, the low-level winds show changes from anticyclonic anomalies in wet years (referring to rainfall anomaly in west Mexico) to cyclonic anomalies in dry years. Subsequently, in the wet years (Fig. 3a), the Southwest has primarily easterly wind anomalies coming from the east and Southeast. In dry years, the wind has westerly anomalies from the cool ocean surfaces to the Southwest. In west Mexico, the wet years have circulation anomalies supporting onshore flows at 850 hPa with strong low-level convergence. In dry monsoon years (Fig. 3b), flow anomalies are found to exhibit northwesterly, alongshore flow from comparatively cool ocean areas resulting in only weak convergence in west Mexico. East Mexico is under both the influences of weak cyclonic anomalies and flows from the inland areas and also has less rainfall.
These circulation anomalies support the positive correlation of rainfall in west Mexico and the Southwest and negative correlation between rainfall in west Mexico and the central United States in 1961–90 (Fig. 2a). Particularly, because the anomalous winds are primarily from the north in the central United States in wet monsoon years for west Mexico and reversed to southerly flows from the Gulf of Mexico to the central United States in dry years, these two regions have a strong negative relationship in their warm season rainfall in this regime. It is noteworthy that the low-level wind anomalies in this regime suggest an out-of-phase relationship between the jet from the Gulf of Mexico into the central United States and the jet from the Gulf of California influencing northwest Mexico and the Southwest.
Very different from these anomaly patterns in the regime of 1961–90, the circulation anomalies in the regime of 1991–2005 show strong negative geopotential anomalies over the west two-thirds of North America for wet monsoon years in west Mexico (Fig. 4a) and a reversed pattern for dry years (Fig. 4b). Another important feature in geopotential heights is the anomalies centered in the eastern midlatitude North Pacific. Positive geopotential anomalies are shown in the wet monsoon years and negative anomalies in dry years.
The associated wind anomalies show that in the wet years (Fig. 4a) west Mexico has strong southerly flows concentrated along the west slopes of the Sierra Madre Occidental and along the eastern slopes of the Sierra Madre Oriental. As also supported by the modeling results of Hu and Feng (2007), the strong southerly low-level jet along the west slope of the mountains from the warm waters in the Gulf of California may have contributed to wet monsoons in west Mexico, although the mechanism for rainfall development remains unclear (Mo and Juang 2003). Meanwhile, the strong southerly low-level jet from the Gulf of the Mexico on the east slope fueled the rainfall in northeast Mexico and also contributed to above-average rainfall in the central United States. This anomalous circulation pattern partially explains the positive correlation of rainfall anomalies in west and east Mexico as well as the positive correlation between west Mexico and the central United States in this regime (Fig. 2b).
When northern Mexico and the central United States were experiencing similar rainfall anomalies, a different situation was present in the Southwest. The geopotential anomalies in the western United States and off the west coasts show a deformation pattern (Fig. 4a) that pushes air from the mid- and high-latitude cool water areas into the Southwest. The relatively cool, dry, and stable air bears less chance for rainfall development, resulting in dry conditions in the Southwest. (The reason for this is that the air mass advected off the North Pacific characteristically has a strong capping inversion at the levels at the top of the marine boundary layer and at around 600–400 hPa. These inversions are critical to inhibiting convection in the Mediterranean climate of California and Baja California.) Furthermore, these northerly flows in the deformation collide with southerly flows from the Gulf of California in west Mexico, where increased low-level convergence may have favored increased rainfall. This configuration of flow anomalies and their potential effects on rainfall could have partially caused the negative correlation of the rainfall in the Southwest and northwest Mexico and the sharp change of the correlation from the latter to the former region in this regime (Fig. 2b).
These anomalies reversed in the dry years (again referring to rainfall in west Mexico; Fig. 4b). Along the east and south fringe of the positive geopotential anomaly in the west United States, the associated anticyclonic wind anomalies favor northerly flows from the central United States down to east Mexico and easterlies from there to west Mexico. The relatively dry winds from inland discourage rainfall in these three areas, causing below-average rainfall (Fig. 2b). In the Southwest the reversed anomalous deformation flow field draws southerly flows and also low-level convergence into AZNM, favoring above-average rainfall, which complements the negative correlation of the rainfall in the Southwest and west and east Mexico as well as the central United States in this regime.
A set of anomaly patterns of 500-hPa geopotential heights and low-level winds similar to that for 1991–2005 is shown for 1948–60 in Figs. 5a and 5b. (Note: because there were no adequate data for years before 1948, the composites of 1948–60 are used to represent in an approximate sense the circulation anomalies in the period from 1931 to 1960.) Comparisons of Figs. 4 and 5 suggest that the composites in these two periods have similar circulation patterns although details vary in some areas (discussed next). In Fig. 5a, the wet monsoon years in west Mexico possessed anomalous southerly flows into west Mexico while a similar anomalous deformation field along the West Coast brought northerly flows and cool air into the Southwest. These anomaly patterns of geopotential and winds are slightly different from those in Fig. 4a, however. For example, the positive geopotential anomalies off the West Coast and in the eastern North Pacific are more east–west oriented, compared to the same signed anomaly in Fig. 4a, blocking further extension of the northerly flow to the Southwest. Potentially because of these differences, the Southwest does not show strong negative correlation between its rainfall and rainfall in west Mexico in this period (see Fig. 2c). Nonetheless, the rather weak positive correlation between rainfall of the Southwest and west Mexico in this period is conceptually consistent with that in the period 1991–2005 (Figs. 2c versus 2b), supporting a similar regime in these two different periods.
To summarize, these distinctive circulation anomaly patterns corresponding to specific rainfall anomaly distributions in the southwest and the central United States and Mexico from one period to the next (i.e., 1931–60 to 1961–90, and then 1991–2005) have characterized a variation (alternation) of two circulation and precipitation regimes in the North American monsoon region.
While most of these results are consistent in many aspects with findings in some previous studies, they help comprehend and also improve understanding of the earlier results. For example, Castro et al. (2001) indicated that anticyclonic anomalies in the western United States with upstream cyclonic anomalies over the eastern North Pacific favored regional circulation anomalies that enhanced the monsoon rainfall in the Southwest. Largely because their study period covers 1948–98, this result was extracted statistically and summarized in a schematic (see their Fig. 14). This result becomes clearer and more direct for the regime of 1961–90 (Figs. 3a and 3b), lending strong support to the hypothesis that the eastward shift of the Rossby wave forcing from SST anomalies in the North Pacific during the warm season could be an important factor in that regime, in addition to the land memory effect (Hu and Feng 2002, 2004), influencing the regional circulation and rainfall in the western United States (Newman and Sardeshmukh 1998).
It should be indicated that a negative correlation between the summer rainfall in the Southwest and central United States, similar to Fig. 2b, also was found in Mo and Juang (2003) for their study period from 1991 to 2000 (also see Berbery and Fox-Rabinovitz 2003). However, their suggested inversed relationship of intensity of the two low-level jets from the Gulf of Mexico and from the Gulf of California, as a plausible explanation of the negative correlation in the rainfall, is not supported by the result of this study. In fact, Fig. 4 shows an enhanced low-level jet from the Gulf of California when the jet flow from the Gulf of Mexico was strong. The positive relationship of the two jets and the negative rainfall correlation between the Southwest and the central United States in this regime indicate that other processes, primarily the regional circulation and the associated deformation field discussed earlier, may have played a more important role in the inversed relationship of the rainfall. The positive correlation of the two jets in this regime shows a change from the negative relationship in the previous regime and also depicts a varying relationship of the jets as part of the multidecadal variation of the North American monsoon regime. While identifying the multidecadal variations of these circulation regimes, these results challenge our understanding of possible mechanisms for such variations/alternations.
4. An interpretation of the regime change
The question of what may have attributed to the development and alternation of these two regimes is addressed in this section by examining the relationships of regime changes and variations of the planetary circulations in the Northern Hemisphere (NH).
Following the postulation that multidecadal variations in regional circulations should be reflections of planetary-scale circulation changes, we examined the multidecadal variations in the NH. Two prominent oscillations at such time scales are the Pacific decadal oscillation (PDO; Mantua et al. 1997; Zhang et al. 1997) and the Atlantic multidecadal oscillation (AMO; Kerr 2000; Enfield et al. 2001; Sutton and Hodson 2005).
The PDO has been shown to be a modulator of precipitation as well as streamflow in western Mexico (e.g., Brito-Castillo et al. 2003). Some recent studies suggest, however, that the effect of PDO on the western Mexico monsoon is somewhat transient. For example, Gochis et al. (2007a) showed that the streamflow in the southern part of west Mexico is modestly correlated to the Niño-3.4 SST in PDO warm phase, whereas the streamflow in the northern part of west Mexico has a stronger relationship with Niño-1 and Niño-2 SST in PDO cold phase. Meanwhile, Gutzler (2004) found that the rainfall in the North American monsoon region is weakly correlated to the Niño-3.4 SST in both PDO warm (1978–98) and cold (1951–76) phases. In addition to these transient effects, Zhang et al. (1997) showed that the PDO changed from warm phase to cold phase around 1948 and returned to warm phase in the late 1970s. These phase changes of PDO are rather different from the variation phase of monsoon regimes previously described, further suggesting a less consistent role of PDO in the regime changes of the North American summer monsoon and precipitation.
The influence of AMO on North American summer rainfall was first shown in Enfield et al. (2001), who defined the AMO index and indicated that in the warm phase of AMO deficit summer rainfall frequently occurred in the central United States west of the Continental Divide, the Southwest, and the Mississippi River basin (also see Rogers and Coleman 2003). Meanwhile, McCabe et al. (2004) show that 28% of the variance of droughts in the contiguous United States since 1900 was explained by AMO forcing (24% by the Pacific decadal oscillation, 22% by the other SST forcing, and the rest by the other minor sources). Additionally, the PC loading and AMO correlations in the results of McCabe et al. (2004) suggest strong and persistent droughts in the Southwest and central United States and Mexico during the warm phases of AMO in the twentieth century, consistent with the result in Enfield et al. (2001). From these results, McCabe et al. (2004) speculated that severe droughts similar to that persisted in the 1930s and 1950s in the central United States may repeat in future warm phases of AMO. They further noted that “long-term predictability of drought frequency may reside in the multidecadal behavior of the North Atlantic Ocean.”
Recently, Sutton and Hodson (2005, 2007) examined the multidecadal time-scale variations in the North American summer climate and provided strong evidence showing that the SST variations associated with AMO have played an important role. Both their observational composites and model simulation ensembles show consistent and significant variations in summer season precipitation, surface temperature, and pressure corresponding to the forcing of AMO. Several studies of Wang et al. (2006, 2008) have further revealed the specific roles of the tropical Atlantic warm pool, which is an integral part of AMO, in causing the summer season circulation anomalies and associated precipitation anomalies in North America.
In addition to these findings of the AMO effect on North American warm season climate variations, it also has been found that AMO strongly influences the rainfall in the Indian summer monsoon region and the Sahel Desert area (Zhang and Delworth 2006; Knight et al. 2006). A causal link of AMO and the rainfall anomalies in those regions was suggested to be a northward shift of the intertropical convergence zone (ITCZ) in the warm phase of AMO and the subsequently enhanced southwesterly winds in the lower to midtroposphere and associated strong rainfall in those monsoon regions. These results along with the mounting evidence showing the influence of AMO on Atlantic hurricane frequency and intensity (Goldenberg et al. 2001; Kerr 2005; Elsner 2006), although the influence has been complicated by anthropogenic effects (Mann and Emanuel 2006; Nyberg et al. 2007), show that AMO has strong coherent influences on the planetary-scale circulations, storms, and rainfall. They indicate that AMO is an essential process in the multidecadal climate variations in NH.
These previous studies support the postulation that AMO plays a key role in the North American monsoon regime changes and also provides a context for exploring such a role. Our examinations of the variations of the AMO index and the monsoon regimes have shown that they have very similar phases. The AMO index, defined by Enfield et al. (2001) as the averaged sea surface temperature anomalies in the North Atlantic region (0°–60°N, 75°–7.5°W), after the long-term trend of the SST from the Hadley Centre SST (HadSST1; Rayner et al. 2003) has been removed, has similar phase with the monsoon regime variations in the twentieth century. As shown in Fig. 6a, AMO changed from cold to warm phase in the early 1930s, back to cold phase in the 1960s, and to warm phase again in the early 1990s. These changes concurred with the regime change of the North American summer monsoon (thin line in Fig. 6a). Although inadequate rainfall data in Mexico limited a comparison to the early years before 1930, Figs. 6b and 6c clearly indicate an in-phase relationship between the North American monsoon and precipitation regime variation and AMO. From Figs. 6b and 6c, in the recent cold phase of AMO (1961–90), the monsoon regime is characterized by the strong positive correlation of monsoon rainfall in west Mexico and the Southwest (Fig. 6c) and the strong negative relationship between west Mexico and the central United States (Figs. 6b, 2). These coherent changes of the AMO and the monsoon regimes indicate a causal effect of the AMO in the variation of the monsoon and precipitation regimes.
A plausible physical connection in this causal link between the AMO and the monsoon regime change is the North Atlantic subtropical high pressure system (NASH). During the multidecadal variation of AMO the air mass/pressure in the North Atlantic subtropical high pressure system varies. This variation is reflected in SLP between the warm and cold phases of AMO. The SLP variations reflect changes of the strength and geographical coverage, or the amplitude and spatial extension, of NASH. Because the southerly and southeasterly branches in the west section of NASH bring moisture from the warm water regions in the south (e.g., the Gulf of Mexico and Gulf of California) to northeastern Mexico and the Southwest as well as central United States, variations in the southerly flows associated with changes in NASH can substantially affect warm season rainfall variations in northern Mexico and the central United States and Southwest.
To explore this hypothesis, we examined the SLP change in different phases of AMO. Figures 7a and 7b show the SLP anomalies along with 850-hPa wind anomalies in the warm and cold phases of AMO in the last century, respectively. The anomalies are departures from their means of 1900–2005, shown in Fig. 7c. Examining Figs. 7a and 7b we find negative anomalies of SLP from the subtropics to the midlatitudes in the western NH during the warm phase of AMO. The largest anomalies are found in the west North American continent and west Europe. Positive anomalies of SLP are shown in Fig. 7b during the cold phase. A statistical test further confirms that these large SLP changes are significant (Fig. 7d). Associated with the SLP changes are low-level wind shifts between the warm and cold phase, influencing the monsoon rainfall. These results indicate strong influences of AMO on warm season SLP, atmospheric circulations, and rainfall in central and western North America. These influences have recently been examined in Wang et al. (2008) and are suggested to be primarily responses of the atmosphere to the westward-propagating Rossby waves initiated by the tropical Atlantic warm pool anomalies (also see Gill 1980).
It is interesting to note that the center pressure (and location) of NASH remains fairly steady during AMO, as shown in Fig. 7d by the insignificant changes in the central North Atlantic. In contrast, the large changes in SLP and winds in central and western North America and western Europe indicate that the circulation changes associated with AMO are most significant in the regions of strong meridional exchange (or the “tail” and “head” regions of NASH; see Fig. 7c). As shown in Wang and Lee (2007), the summer season NASH expands or contracts depending on the SST in different AMO phases. In the cold phase, the NASH could extend well over the Caribbean warm waters and engage with large moisture sources (Wang et al. 2006). The strengthened anticyclonic circulation in NASH and the low-level jets from the south (e.g., the Gulf of Mexico) bring moisture into Mexico, the central and western United States, as well as the regions east of the Rocky Mountains, influencing summer rainfall in these regions. In the warm phase, the NASH contracts and the weakened moisture flow to the United States and Mexico negatively affects their summer rainfall. In these alternations, the spatial coverage and intensity of the NASH change substantially and enhance or weaken the north–south exchanges of moisture and momentum in western and central North America, resulting in the development and change of the North American monsoon and precipitation regimes.
In Fig. 7a, the negative anomalies in the west North American continent indicate weakening and contraction of NASH in the warm phase of AMO. The opposite SLP anomalies in the region during the cold phase show expansion and enhancement of NASH (Fig. 7b). These changes of the SLP and wind anomaly patterns during AMO set up different circulation backgrounds, whereby the observed specific regional circulation anomalies can develop and result in the observed rainfall patterns in the North American monsoon region in the different AMO phases. For example, in the recent warm phase of AMO, North America was favored with negative pressure anomalies with a strong center in the north-central United States (Fig. 7a). This pressure anomaly pattern favored development of the regional circulation anomalies shown in Fig. 4, with broad and large negative anomalies in North America with an anomaly center in the west-central United States. Although mechanisms developing and connecting the various anomalies remain to be investigated, these coherent anomalies across North America characterize the particular circulation and precipitation regime in the region.
The difference of these regimes also is shown in their mean precipitation. Figure 8a shows the difference of July–September mean precipitation in the United States and northern Mexico between the two regimes, 1931–60 versus 1961–90. Less precipitation is observed in the regime of 1931–60 than 1961–90. These differences are consistent with the hypothesis that the warm phase of the AMO corresponds to less summer precipitation in North America. A similar pattern of less precipitation as in Fig. 8a also emerged in Fig. 8b from a further analysis of the regression of July–September precipitation in North America with the AMO index. These results lend further support to the critical role of AMO in multidecadal variations in the North American monsoon circulation and precipitation regimes, although regional interaction processes and how they may contribute to or affect the development of precipitation in different regimes need further investigation.
5. Summary and concluding remarks
This study shows two regimes of July–September circulation and precipitation in the western and central United States and northern Mexico. These regimes are described by different yet temporally and spatially coherent variations of precipitation in the region. They have alternated at multidecadal time scales in the twentieth century. In one regime (represented in 1961–90), the Southwest and northwestern Mexico share the same summer precipitation pattern, which is negatively correlated with the rainfall pattern in the central United States. In the other regime (occurring from 1931–60 and also from 1991 onward), northwestern and northeastern Mexico share a similar summer rainfall pattern, which has a positive correlation with the rainfall in the central United States but a reversed relationship with rainfall in the Southwest.
Analyses of the summer circulation in North America showed that in the regime from 1961 to 1990 the circulation affected the development of monsoon rainfall in west Mexico and the Southwest differently from its effect on rainfall in the central United States. A major feature is the anticyclonic circulation anomaly centered in the north and north-central United States. The anomalous northerly flows in the eastern flank of the center discourage rainfall development in the central United States and also suppress the low-level jet from the Gulf of Mexico into the central United States. In the meantime, the easterly and southeasterly flows in the south fringe of the anticyclonic anomaly bring moisture from regions near the Gulf of Mexico to the Southwest and northern Mexico to increase their rainfall. These flow anomalies also enhanced the low-level jet from the Gulf of California to increase rainfall in northwest Mexico as well as the Southwest. (There is therefore a negative relationship of intensity of the southerly jet from the Gulf of Mexico to the central United States and the jet from the Gulf of California to the Southwest in this regime.) A reversed circulation anomaly pattern results in reversed rainfall distribution, supporting coherent precipitation patterns in central and western North America for this regime.
In the recent regime from 1991 onward, the circulation anomaly pattern is quite different from the previous one, featuring a strong anticyclonic anomaly center in the midlatitude North Pacific off the coast of California. The anomalous flow patterns show a weak deformation field centered along northwest Mexico and the Southwest. Because of the different properties of air masses from the north and the south in the deformation flows, the rainfall anomalies in northwest Mexico are nearly opposite to those in the Southwest. In the meantime, the broad cyclonic anomalies to the east of the anticyclonic anomaly, with a center in the northern Great Plains, encourage strong moist air from the regions near the Gulf of Mexico into the central United States, causing more rainfall. Both the jets from the Gulf of Mexico and the Gulf of California are strong in wet years, suggesting a positive relationship between their variations in this regime. The jet from the Gulf of California converges with the northerly flow along the southern tier of the Southwest, contributing primarily to the rainfall in northern Mexico. More rainfall in northern Mexico and the central United States is in contrast to the dry condition in the Southwest in this regime.
Alternations of these patterns at multidecadal time scale are found to be quite cohesive with the AMO, whereas the influences of the North Pacific are somewhat more ambiguous and transitory. In the twentieth century, the warm (cold) phases of AMO corresponded to average lower (higher) surface pressure anomalies in North America with a strong anomaly center in the central and western United States. These pressure anomalies support the regional circulation anomalies previously described, which have contributed to the spatial precipitation patterns in the different monsoon regimes. These coherent variations between AMO and the summer circulation and precipitation regimes indicate that the AMO must have played an essential role in the multidecadal time scale alternation of the North American monsoon and precipitation regimes, further supporting the notion that the long-term predictability of precipitation anomalies in North America may reside in the multidecadal behavior of the North Atlantic Ocean (McCabe et al. 2004). Results of this study have expanded upon those of McCabe et al. (2004), Enfield et al. (2001), and Wang et al. (2006, 2008) and have specified the AMO effect on the circulation and summer precipitation regime change in North America.
Further detailed explanations of the specific processes that may have achieved the effects of AMO in the North American monsoon region remain to be seen, and the interactions between the AMO and circulation anomalies centered in the North Pacific also need to be provided. Modeling studies such as those in Zhang and Delworth (2006) and Knight et al. (2006) may bring new insight on the causal links of the AMO and the North American monsoon regime changes and precipitation variability. From these links, the predictability for variations of circulations and rainfall in North America could be developed and further integrated in drought and water resources planning and mitigation at advanced decadal and longer time scales.
Acknowledgments
We thank Dr. Dave Gochis, two anonymous reviewers, and the editor, Dr. S. P. Xie, for their comments, which have led to improvement of this manuscript. This research has been supported by the USDA Cooperative Research Project NEB-40-040.
REFERENCES
Allan, R. and T. Ansell, 2006: A new globally complete monthly historical gridded mean sea level pressure dataset (HadSLP2): 1850–2004. J. Climate, 19:5816–5842.
Berbery, E. H. and M. S. Fox-Rabinovitz, 2003: Multiscale diagnosis of the North American monsoon system using a variable-resolution GCM. J. Climate, 16:1929–1947.
Brito-Castillo, L., A. V. Douglas, A. Leyva-Contreras, and D. Lluch-Belda, 2003: The effect of large-scale circulation on precipitation and streamflow in the Gulf of California watershed. Int. J. Climatol., 23:751–768.
Castro, C. L., T. B. McKee, and R. A. Pielke Sr., 2001: The relationship of the North American monsoon to tropical and North Pacific sea surface temperatures as revealed by observational analyses. J. Climate, 14:4449–4473.
Comrie, A. C. and E. C. Glenn, 1998: Principal components-based regionalization of precipitation regimes across the southwest United States and northern Mexico, with an application to monsoon precipitation variability. Climate Res., 10:201–215.
Elsner, B., 2006: Evidence in support of the climate change–Atlantic hurricane hypothesis. Geophys. Res. Lett., 33.L16705, doi:10.1029/2006GL026869.
Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U.S. Geophys. Res. Lett., 28:2077–2080.
Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106:447–462.
Gochis, D. J., L. Brito-Castillo, and W. J. Shuttleworth, 2007a: Correlations between sea-surface temperatures and warm season streamflow in northwest Mexico. Int. J. Climatol., 27:883–901.
Gochis, D. J., C. J. Watts, J. Garatuza-Payan, and J. Cesar-Rodriguez, 2007b: Spatial and temporal patterns of precipitation intensity as observed by the NAME Event Rain Gauge Network from 2002 to 2004. J. Climate, 20:1734–1750.
Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nuñez, and W. M. Gray, 2001: The recent increase in Atlantic hurricane activity: Causes and implications. Science, 293:474–479.
Gutzler, D. S., 2004: An index of interannual precipitation variability in the core of the North American monsoon region. J. Climate, 17:4473–4480.
Gutzler, D. S. and J. W. Preston, 1997: Evidence for a relationship between spring snow cover and summer rainfall in New Mexico. Geophys. Res. Lett., 24:2207–2210.
Higgins, R. W. and W. Shi, 2000: Dominant factors responsible for interannual variability of the summer monsoon in the southwestern United States. J. Climate, 13:759–776.
Higgins, R. W., K. C. Mo, and Y. Yao, 1998: Interannual variability of the U.S. summer precipitation regime with emphasis on the southwestern monsoon. J. Climate, 11:2582–2606.
Higgins, R. W., Y. Chen, and A. V. Douglas, 1999: Interannual variability of the North American warm season precipitation regime. J. Climate, 12:653–680.
Higgins, R. W., W. Shi, E. Yarosh, and R. Joyce, 2000: Improved United States precipitation quality control system and analysis. NCEP/Climate Prediction Center ATLAS 7, NCEP/NWS/NOAA, 40 pp.
Hu, Q. and S. Feng, 2002: Interannual rainfall variations in the North American summer monsoon region: 1900–98. J. Climate, 15:1189–1202.
Hu, Q. and S. Feng, 2004: Why has the land memory changed? J. Climate, 17:3236–3243.
Hu, Q. and S. Feng, 2007: Decadal variation of the southwest U.S. summer monsoon circulation and rainfall in a regional model. J. Climate, 20:4702–4716.
Kalnay, E., Coauthors 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77:437–471.
Kerr, R. A., 2000: A North Atlantic climate pacemaker for the centuries. Science, 288:1984–1986.
Kerr, R. A., 2005: Atlantic climate pacemaker for millennia past, decades hence? Science, 309:41–42.
Knight, J. R., C. K. Folland, and A. A. Scaife, 2006: Climate impacts of the Atlantic Multidecadal Oscillation. Geophys. Res. Lett., 33.L17706, doi:10.1029/2006GL026242.
Mann, M. and K. Emanuel, 2006: Atlantic hurricane trends linked to climate change. Eos, Trans. Amer. Geophys. Union, 87:233.
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.
McCabe, G. J., M. A. Palecki, and J. L. Betancourt, 2004: Pacific and Atlantic Ocean influences on multidecadal drought frequency in the United States. Proc. Natl. Acad. Sci. USA, 101:4136–4141.
Mo, K. C. and H. M. H. Juang, 2003: Influence of sea surface temperature anomalies in the Gulf of California on North American monsoon rainfall. J. Geophys. Res., 108.4112, doi:10.1029/2002JD002403.
New, M., M. Hulme, and P. D. 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.
Newman, M. and P. D. Sardeshmukh, 1998: The impact of the annual cycle on the North Pacific/North American response to remote low-frequency forcing. J. Atmos. Sci., 55:1336–1353.
Nyberg, J., B. A. Malmgren, A. Winter, M. R. Jury, K. H. Kilbourne, and T. M. Quinn, 2007: Low Atlantic hurricane activity in the 1970s and 1980s compared to the past 270 years. Nature, 447:698–701.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108.4407, doi:10.1029/2002JD002670.
Rogers, J. C. and J. S. M. Coleman, 2003: Interactions between the Atlantic Multidecadal Oscillation, El Niño/La Niña, and the PNA in winter Mississippi Valley stream flow. Geophys. Res. Lett., 30.1518, doi:10.1029/2003GL017216.
Ropelewski, C. F., D. S. Gutzler, R. W. Higgins, and C. R. Mechoso, 2005: The North American monsoon system. The Global Monsoon System: Research and forecast, WMO Tech. Doc. 1266, 207–218.
Sutton, R. T. and D. L. R. Hodson, 2005: Atlantic Ocean forcing of North American and European summer climate. Science, 309:115–118.
Sutton, R. T. and D. L. R. Hodson, 2007: Climate response to basin-scale warming and cooling of the North Atlantic Ocean. J. Climate, 20:891–907.
Vera, C., Coauthors 2006: Toward a unified view of the American monsoon systems. J. Climate, 19:4977–5000.
Wang, C. and S-K. Lee, 2007: Atlantic warm pool, Caribbean low-level jet, and their potential impact on Atlantic hurricanes. Geophys. Res. Lett., 34.L02703, doi:10.1029/2006GL028579.
Wang, C., D. B. Enfield, S-K. Lee, and C. W. Landsea, 2006: Influences of the Atlantic warm pool on Western Hemisphere summer rainfall and Atlantic hurricanes. J. Climate, 19:3011–3028.
Wang, C., S-K. Lee, and D. B. Enfield, 2008: Climate response to anomalously large and small Atlantic warm pools during the summer. J. Climate, 21:2437–2450.
Yu, B. and J. M. Wallace, 2000: The principal mode of interannual variability of the North American monsoon system. J. Climate, 13:2794–2800.
Zhang, R. and T. L. Delworth, 2006: Impact of Atlantic multidecadal oscillations on India/Sahel rainfall and Atlantic hurricanes. Geophys. Res. Lett., 33.L17712, doi:10.1029/2006GL026267.
Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like interdecadal variability: 1900–93. J. Climate, 10:1004–1020.
Zhu, C., D. P. Lettenmaier, and T. Cavazos, 2005: Role of antecedent land surface conditions on North American monsoon rainfall variability. J. Climate, 18:3104–3121.
Time series of total rainfall for JAS in (a) west Mexico and (b) AZNM (the geographical areas are shown by the dashed-line boxes in Fig. 2a). The solid lines are from NOAA CPC data and the dashed lines are from CRU data. The correlation coefficient of the two series is 0.77 for the years before 1987 for west Mexico and is 0.24 for the years after 1987. The correlation coefficient is 0.89 for 1960–2002 in AZNM. The histograms in the lower sections of (a), (b) show the number of observation stations in the CRU dataset in the corresponding regions in each year.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
Correlations between JAS rainfall in the west Mexico monsoon region and rainfall of individual grids across North America during (a) 1961–90, (b) 1991–2005, and (c) 1931–60. Here (a), (b) are based on NOAA CPC data and (c) is based on CRU data. Shading indicates correlations significant at the 95% confidence level. The dashed-line boxes mark the regions of west Mexico, the AZNM areas, and the central United States defined in this study.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
Composites of 500-hPa geopotential heights (contour line; units: gpm) and 850-hPa wind (arrows; units: m s−1) anomalies for (a) wet and (b) dry west Mexico monsoon years in the regime of 1961–90. The anomalies are based on the climatic means of the regime. (c) Difference between the wet and dry years. Shadings indicate that the differences of 500-hPa geopotential heights between wet and dry years are significant at the 95% confidence level.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
Same as Fig. 3, but for 1991–2005.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
(a) Thick line shows the time series of the AMO index and thin line shows the monsoon regime variation measured by the land memory (Hu and Feng 2002); (b) 21-point moving correlations of JAS rainfall between WM and CUS; and (c) same as (b) but between WM and AZNM. Thin lines in (b) and (c) are based on CRU data and thick lines are based on NOAA CPC data. Dotted lines show the 95% confidence level.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
Composite JAS SLP (contour lines) and 850-hPa wind anomalies (arrows) for AMO (a) warm and (b) cold phase. (c) Mean JAS SLP for 1900–2005. (d) Differences in JAS SLP anomalies between AMO warm and cold phase. Shading shows the significant (>95% confidence level) SLP changes between the warm and cold phases. The wind anomalies are based on data from 1948–60 and 1991–2005 for warm phase and 1961–90 for cold phase. The SLP anomalies are based on data from 1931–60 and 1991–2005 for warm phase and 1900–30 and 1961–90 for cold phase.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
(a) JAS rainfall differences (in percentage changes from the 1961–90 climatic mean) between 1931–60 and 1961–90. Shading shows rainfall changes between the two periods significant at the 95% confidence level. (b) Regression between JAS AMO and JAS rainfall in North America. Shading indicates significant correlations at the 95% confidence level. Data used are from CRU. Most of North America is wetter in the AMO cold phase than the warm phase.
Citation: Journal of Climate 21, 11; 10.1175/2007JCLI2005.1
