1. Introduction
Most studies of regional precipitation anomalies associated with regional teleconnection patterns such as the Pacific–North America (PNA) pattern or North Atlantic Oscillation (NAO) tend to focus on the winter. Few studies have examined interregional patterns of precipitation and the linkages to global and regional circulation patterns in seasons other than winter (Kingston et al. 2006). This is a particularly important consideration because the largest trends in precipitation and streamflow over the United States have been observed in the fall (Karl and Knight 1998; Small et al. 2006).
There is ample evidence that fall precipitation and low to median stream flows (Lins and Slack 1999; Douglas et al. 2000; Small et al. 2006) have increased across the central and eastern United States. There is also evidence that precipitation (Gan 1998; Dery and Wood 2005) and streamflow (Zhang et al. 2001; Dery and Wood 2005) are decreasing in southern and eastern Canada and increasing in northern high latitudes (Zhang et al. 2001). In this study, we address the question of whether it is possible to identify a single mode of variability that links fall precipitation anomalies in the central United States with those in northern and eastern Canada. To address this objective, the dominant pattern of fall precipitation variability over North America is identified from multiple data sources and linked to a hemispheric-scale circulation pattern that resembles the circumglobal teleconnection (Branstator 2002) and modulates the moisture transport over the different regions of the continent.
2. Data
The monthly gridded (2.5° × 2.5°) Precipitation Reconstruction over Land (PREC/L) precipitation data (Chen et al. 2002) covering the period 1948–2004 are used in this study. The dataset is derived from gauge observations collected in the Global Historical Climatology Network (GHCN), version 2, and the Climate Anomaly Monitoring System (CAMS) datasets. Gauge observations are sparse in eastern Canada and primarily limited to areas near the U.S. border and along the east coast. Monthly satellite precipitation data from the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP) dataset (1979–2004) are also used to check the robustness of the precipitation signal over remote regions of northern Canada and to extend the analysis to include the North Pacific. Monthly wind and specific humidity data from the National Centers for Environmental prediction (NCEP) global reanalysis dataset (Kalnay et al. 1996) were used to estimate the monthly-mean, column-integrated mean moisture flux, 300-hPa streamfunction, and 300-hPa rotational wind from 1948 to 2004. The seasonal cycle was removed from all of the data by subtracting the monthly mean before calculating fall [September–November (SON)] mean anomalies.
3. Methods and results
a. The leading mode of fall precipitation variability
To identify the leading mode of fall precipitation variability over North America between 1948 and 2004, principal component analysis (PCA) was performed on fall (SON) precipitation over the United States and Canada from 25° to 60°N using a correlation matrix instead of a covariance matrix to highlight large-scale patterns in precipitation and avoid focusing on isolated areas such as the Gulf Coast where the variance of the precipitation tends to be large. The spatial pattern of the leading mode [the first empirical orthogonal function (EOF1)] of precipitation is estimated by correlating the leading principal component (PC1) and the precipitation time series at each grid point (Fig. 1a). The sparseness of precipitation stations in northern Canada presents a potential problem for using gridded precipitation datasets such as the PREC/L in the study of large-scale precipitation variations. In regions where there are relatively few stations, the spatial variability of precipitation is not well represented in the gridded data. To check the consistency of the leading mode of precipitation, the PCA was repeated using the 2.5° gridded CMAP satellite precipitation from 1979 to 2004. The CMAP satellite precipitation also allows the domain of the analysis to be expanded to include the extratropical Pacific (20°–60°N, 120°E–30°W). The leading principal component of the precipitation from each datasets was also correlated with the fall average 300-hPa streamfunction to identify the signature of a large-scale dynamic control on precipitation.
The leading principal component of PREC/L precipitation (1948–2004) explains 15.6% of the variance over the North American domain. The pattern associated with PC1 exhibits large correlations of one sign (positive) across most of the interior of the United States west of the Appalachian Mountains, with large opposite-signed correlations (negative) in the coastal regions along the Gulf of Alaska in southern Alaska and northern British Columbia and in eastern Quebec and the Canadian Maritime provinces. The pattern indicates a tendency for wet conditions in the central United States to occur during dry periods along the Gulf of Alaska and eastern Canada. The spatial pattern of the leading mode of fall precipitation variability in North America is tripolar, with the precipitation anomalies in the central United States having a sign opposite those along the eastern and western coasts of Canada.
The correlation between PC1 and the 300-hPa streamfunction indicates the presence of a wave pattern that appears to originate in the western tropical Pacific and stretches across North America and into the North Atlantic (Fig. 1a). The wave pattern includes a negative anomaly over the western United States and positive anomalies near the Aleutian Islands and across eastern North America. The leading principal component, however, explains very little of the variance in precipitation across eastern Canada (i.e., low correlation with precipitation across the region). We repeated the principal component analysis using the PREC/L data over the same domain from 1979 to 2004, the period of the CMAP satellite data. The leading principal component explains 20.7% of the variance over the North American domain over the period 1979–2004. The tripole pattern is also more clearly visible associated with the leading PC of the later period when the precipitation across most of eastern Canada from the Great Lakes to the Maritime Provinces and Newfoundland is more strongly anticorrelated with precipitation in the central United States (Fig. 1b). A wave pattern similar to that shown in Fig. 1a is observed when the leading PC from the latter period is correlated with the 300-hPa streamfunction, but with the positive center over eastern North America shifted well to the north over the Great Lakes and eastern Canada.
The leading principal component of CMAP precipitation explains 10.9% of the variance over the larger domain that includes North America and the extratropical Pacific. The pattern associated with PC1 captures the tripole, with the precipitation anomalies in the central United States strongly anticorrelated with precipitation anomalies in the Gulf of Alaska and eastern Canada. The pattern associated with the lead PC of the CMAP satellite precipitation also shows that the precipitation signal over North America is part of a larger wavelike pattern of alternating wet and dry anomalies that stretches from east of Japan well into the North Atlantic. The correlation with the 300-hPa streamfunction indicates a wave train with an apparent origin in the western Pacific with a positive center over the Great Lakes. In Figs. 1b and 1c, the positive center of the 300-hPa streamfunction anomalies is over eastern Canada and not the southeastern United States, consistent with the stronger negative precipitation anomaly over eastern Canada.
The PC1 time series from the PREC/L (1948–2004) and CMAP (1979–2004) observations are shown in Fig. 1d. A 3-yr running average smoother was applied to the time series to highlight the interannual fluctuations in the time series. The principal component time series show that a shift toward positive precipitation anomalies occurred in the early 1960s with two periods of anomalously wet conditions observed across the central United States (PC1 positive) in the early 1970s and mid-1980s. Dry periods in the central United States (PC1 negative) are observed in the early 1950s, late 1970s, and late 1980s. A similar pattern is observed in PC1 of the CMAP precipitation after 1979, with a correlation of 0.84 between the two leading principal component time series (unsmoothed) over the common period 1979–2004. As the precipitation datasets differ both in terms of the use of satellite data and the inclusion of ocean areas, the consistency of the leading pattern and associated time series between the two estimates, as well as the similar patterns in the 300-hPa streamfunction anomalies, together provide strong evidence for a physically consistent tripole mode of variability in fall precipitation over North America. The rest of this analysis focuses on the PREC/L precipitation and National Center for Atmospheric research (NCAR)–NCEP reanalysis data to take advantage of the longer 1948–2004 period.
The leading precipitation PC and associated streamfunction anomalies suggest that atmospheric dynamics are playing a key role in the dominant fall precipitation regime over the Pacific–North America region. The pattern of the correlation with streamfunction demonstrates that positive (negative) precipitation anomalies in the central United States (eastern and western Canada) are associated with anomalous southerly (northerly) winds over the region, suggesting a role for the large-scale circulation in modulating the moisture transport over different regions of North America. In the following two sections, we isolate the leading mode of the Northern Hemispheric meridional wind in fall and link them to anomalous moisture transport and precipitation over North America.
b. Circulation anomalies conducive for the precipitation tripole
The observed streamfunction anomalies (Fig. 1) exhibit a zonally extended wavenumber-5 structure that extends from the western tropical Pacific to the North Atlantic. With positive centers south of Japan and over the Gulf of Alaska, the pattern resembles the winter circumglobal teleconnection (CGT) pattern (Branstator 2002; Branstator and Selten 2009). Branstator (2002) demonstrated the importance of a wavelike pattern that is able to extend all the way around the Northern Hemisphere by closely following the waveguide defined by the core of westerlies. The analysis demonstrated that disturbances in the presence of the Asian jet stream tend to be of shorter wavelength and enhanced zonal extent than disturbances over regions with weaker meridional wind gradients and produce teleconnections that are correlated over much larger distances. Here, we adopt an approach similar to Branstator (2002) for isolating the fall CGT pattern and relating it to the observed precipitation tripole.
To isolate the CGT pattern, principal component analysis is performed on the meridional component of the fall mean 300-hPa rotational wind (Vrot) from the equator to 45°N. The meridional rotational wind is chosen because it is typically associated with the smaller-scale disturbances that theory predicts should be preferentially trapped in the subtropical jet (Branstator 2002). The equator to 45°N domain is chosen to highlight variability in the meridional wind field associated primarily with the subtropical jet (Branstator 2002). The two leading principal components of Vrot explain 18.2% and 13.6% of the total variance, respectively. The patterns associated with the two leading PCs are wave patterns that stretch across most of the Northern Hemisphere from Asia to Europe (Fig. 2). The pattern of the first PC (Fig. 2a.) places rotational wind anomalies of opposite signs over the central United States and the Gulf of Alaska (North Atlantic), consistent with the observed circulation anomalies in Fig. 1. Over North America, the pattern of the second PC (Fig. 2b) is spatially shifted approximately a quarter of a wavelength, with the strongest Vrot anomalies observed across New England, eastern Canada, and the Gulf of Alaska, with opposite signed anomalies over the Canadian prairies and the north central United States. The leading principal components of Vrot were also correlated with the 300-hPa streamfunction anomalies (Figs. 2a,b) to highlight the consistency with the wave pattern in Fig. 1. The pattern of the streamfunction anomalies associated with PC1 (Fig. 2a) is similar to the pattern in Fig. 1, only shifted to the east by approximately 10°. The strong correlation between PC1 of Vrot and the leading PC of the PREC/L precipitation over North America further corroborates this relationship.
The correlations between the unsmoothed time series of PC1 of PREC/L precipitation (1948–2004) and PC1 and PC2 of Vrot are 0.66 and −0.42, respectively. The smoothed time series of PC1 is plotted with the smoothed time series of PC1 of precipitation to highlight the low-frequency fluctuations in the time series (Fig. 2c). The time series both display three prominent decadal peaks near 1955, 1972, and 1985, suggesting that this mode of atmospheric variability is an important factor in the fall precipitation regime across the Pacific–North American region. During the wet periods of the early 1970s and mid-1980s, when PC1 of precipitation is strongly positive, PC2 of Vrot is strongly negative (Fig. 2d). The wettest periods in the central United States occur when PC1 of Vrot is strongly positive and PC2 is strongly negative. When this is the case, both modes of variability are associated with enhanced southerly 300-hPa meridional wind anomalies over the central United States and negative wind anomalies over eastern Canada and the west coast. Although we have not performed a rigorous trend analysis in this study, the leading PCs of 300-hPa meridional wind and precipitation have both increased over the period 1948–2004. The strong correlation between the leading mode of the 300-hPa meridional wind and the leading principal component of precipitation then suggests that the CGT might be an important factor in producing the observed trend in fall precipitation over North America. Directly linking the CGT pattern to trends in precipitation and moisture transport is beyond the scope of the current study and will be rigorously addressed in a future publication.
c. Circulation anomalies and anomalous moisture transports
The relationship between the Vrot anomalies and precipitation is addressed by estimating precipitation and meridional moisture flux anomalies during the extreme phases of the two leading modes of Vrot. The extreme phases of Vrot are defined as the six years when PC1 > 0.5 standard deviation and PC2 < −0.5 (positive phase) or the five years when PC1 < −0.5 and PC1 > 0.5 (negative phase). We define the extreme phases of the teleconnection this way because the PCs suggest that the extreme wet or dry periods after the mid-1960s appear to occur when the two time series are of opposite signs and near their relative maxima. Composites of the precipitation and the column-integrated zonal and meridional moisture flux anomalies at each grid point were then estimated for each phase. The precipitation time series at each grid point were divided by the standard deviation so that the precipitation composites have units of standard deviations. The composite precipitation and moisture flux anomalies are shown in Fig. 3.
The positive phase precipitation and moisture flux anomalies are both positive over the central United States (Fig. 3a). The vectors indicate that the anomalous southerly moisture flux over the central United States is associated with an anticyclonic circulation anomaly centered over the eastern United States and a weak cyclonic anomaly over the Rockies. Negative moisture flux anomalies are also observed in the Gulf of Alaska and along the east coast. A prominent cyclonic circulation anomaly is also observed east of the Aleutian Islands, although the precipitation anomalies are extremely weak along the Alaska coast during the positive phase composite. Anomalous cyclonic circulation is also observed over the Atlantic, which contributes to northerly moisture flux and negative precipitation anomalies over eastern Canada. The results suggest that a wave pattern is also observed in the moisture flux anomalies associated with the alternating areas of anticyclonic and cyclonic circulation anomalies stretching from the Gulf of Alaska, across North America, and into the central Atlantic that modulate the transport of moisture over the continent.
During the negative phase (Fig. 3b), large negative precipitation and northerly meridional moisture flux anomalies are observed over the central United States and the Canadian prairies, with positive anomalies of both observed along the Gulf of Alaska coast and along the east coast of North America. A wave pattern is also visible in the moisture flux anomalies in the composite of the negative phase years, with cyclonic moisture flux anomalies over the Gulf of Alaska and eastern United States and weak anticyclonic anomalies over the Atlantic.
The difference between the positive phase and negative phase was also plotted (Fig. 3c). The difference field clearly shows the tripole precipitation pattern, with positive differences in precipitation exceeding 2 standard deviations over the central United States and negative differences of 1.3 standard deviations in southern Alaska and eastern Canada. The large positive difference between the precipitation anomalies over the central United States during the different phases is also visible in the moisture flux. The southerly moisture flux anomalies over the central United States can be explained by a much stronger cyclonic flow over the Rockies and anticyclonic flow to the east of the Great Lakes. Similarly, the large negative precipitation differences over eastern Canada can also be explained by the anticyclonic circulation anomalies east of the Great Lakes along with enhanced cyclonic circulation over the Atlantic. The negative precipitation anomalies along the southern coast of Alaska appear to be associated with anomalous northerly moisture flux and anticyclonic circulation anomalies over the Gulf of Alaska and cyclonic anomalies over the Rockies. The difference field (Fig. 3c) clearly shows the similarity between the direction of the anomalous moisture flux over the different regions of North America and the 300-hPa streamfunction anomalies observed in Figs. 1 and 2. The precipitation and moisture flux anomalies defined by the different phases of the Vrot anomalies are clearly consistent with a tripole precipitation pattern and show that the large-scale circulation anomalies are playing a primary role in modulating the regional moisture flux.
4. Discussion
In this study, we have identified a tripole pattern in fall precipitation over North America from two different precipitation datasets and linked the pattern to regional moisture flux changes and hemispheric-scale circulation anomalies. Our analysis shows that the leading mode of fall precipitation variability over North America is linked to a hemispheric-scale wave pattern that produces anomalies of one sign in the central and north central United States and the opposite sign in New England, eastern Canada, and costal regions bordering the Gulf of Alaska. The seasonal precipitation anomalies in the three regions are linked through a zonally extended wavelike circulation pattern that reaches from the western Pacific across North America into the North Atlantic. The results demonstrate a consistent connection between circulation anomalies, the meridional moisture flux and precipitation over the different regions of North America.
The observed streamfunction anomalies exhibit a zonally extended wavenumber-5 structure with positive centers south of Japan and over the Gulf of Alaska and resemble the winter circumglobal teleconnection (CGT) described by previous studies. A CGT index was estimated from the rotational component of the 300-hPa meridional wind from the equator to 45°N. The precipitation and moisture flux anomalies during different phases of the CGT-like pattern are consistent with those of the tripole precipitation pattern. The CGT has been previously observed in winter (Branstator 2002) and summer (Ding and Wang 2005), but this study is the first to identify its signature in fall precipitation over the North Pacific and North America and to suggest that it might be a major factor in controlling the seasonal mean transport of moisture into different regions of the continent.
The recent study of Branstator and Selten (2009) found that the climate trend in a large ensemble of general circulation model runs has the same structure as the CGT and highlights the importance of this intrinsic mode of variability on trends in regional climate. If future changes in regional climate over North America are likely to strongly reflect this mode of intrinsic variability, the results presented in this study suggest that the pattern of precipitation trends over North America in the future climate might strongly resemble the precipitation tripole pattern. Quantifying how intrinsic modes of variability connected to the strength of the jet streams (such as the CGT) impact regional precipitation is particularly important because climate model experiments predict that a warming of the poles relative to the equator under doubled CO2 is likely to weaken the baroclinic zones and shift storm tracks and the eddy-driven polar jet to the north (Yin 2005). Such a shift is consistent with increases in precipitation and storminess over some regions of North America and decreases over others (Salathé 2006). Other likely scenarios involve changes in tropical SSTs and circulations that might possibly displace the subtropical jet through changes in the position or strength of the Hadley cell (Risbey et al. 2002). We are currently investigating potential tropical forcing of the hemispheric-scale wave patterns identified in this study, the seasonal cycle of CGT variability, and the relationships to trends and decadal-scale variability. We also plan a detailed study of CGT-induced precipitation trends in the doubled CO2 scenarios used in the Fourth Assessment Report (AR4).
The results presented in this manuscript have not considered the role of other teleconnections such as the El Niño–Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO), or the Pacific–North America (PNA) pattern on fall precipitation variability. This is because fall precipitation over the central United States is largely uncorrelated with the SST-based indices of ENSO or the PDO (not shown). The precipitation in the central United States is weakly anticorrelated to the PNA index (not shown). The pattern of correlation between fall precipitation and the 300-hPa streamfunction suggests a much more zonally extended (i.e., nearly hemispheric) teleconnection pattern than that typically associated with the regional PNA pattern, motivating the current focus on the CGT. The relationship, if any, between the CGT and the PNA, PDO, or ENSO is beyond the scope of this work.
Acknowledgments
This research was supported, in part, by grants from the National Science Foundation (ATM-0741600, EAR-0809783, and EAR 0811099) and Tufts University School of Engineering.
REFERENCES
Branstator, G., 2002: Circumglobal teleconnections, the jet stream waveguide, and the North Atlantic Oscillation. J. Climate, 15 , 1893–1910.
Branstator, G., and Selten F. , 2009: “Modes of variability” and climate change. J. Climate, 22 , 2639–2658.
Chen, M., Xie P. , Janowiak J. E. , and Arkin P. A. , 2002: Global land precipitation: A 50-yr monthly analysis based on gauge observations. J. Hydrometeor., 3 , 249–266.
Dery, S. J., and Wood E. F. , 2005: Decreasing river discharge in northern Canada. Geophys. Res. Lett., 32 , L10401. doi:10.1029/2005GL022845.
Ding, Q., and Wang B. , 2005: Circumglobal teleconnection in the Northern Hemisphere summer. J. Climate, 18 , 3483–3505.
Douglas, E. M., Vogel R. M. , and Kroll C. N. , 2000: Trends in flood and low flows in the United States: Impact of spatial correlation. J. Hydrol., 240 , 90–105.
Gan, T. Y., 1998: Hydroclimatic trends and possible climatic warming in the Canadian prairies. Water Resour. Res., 34 , 3009–3015.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Karl, T. R., and Knight R. W. , 1998: Secular trends of precipitation amount, frequency, and intensity in the United States. Bull. Amer. Meteor. Soc., 79 , 231–241.
Kingston, D. G., McGregor G. R. , Hannah D. M. , and Lawler D. M. , 2006: River flow teleconnections across the northern North Atlantic region. Geophys. Res. Lett., 33 , L14705. doi:10.1029/2006GL026574.
Lins, H. F., and Slack J. R. , 1999: Streamflow trends in the United States. Geophys. Res. Lett., 26 , 227–230.
Risbey, J. S., Lamb P. J. , Miller R. L. , Morgan M. C. , and Roe G. H. , 2002: Exploring the structure of regional climate scenarios by combining synoptic and dynamic guidance and GCM output. J. Climate, 15 , 1036–1050.
Salathé E. P. Jr., , 2006: Influences of a shift in North Pacific storm tracks on western North American precipitation under global warming. Geophys. Res. Lett., 33 , L19820. doi:10.1029/2006GL026882.
Small, D., Islam S. , and Vogel R. M. , 2006: Trends in precipitation and streamflow in the eastern U.S.: Paradox or perception? Geophys. Res. Lett., 33 , L03403. doi:10.1029/2005GL024995.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32 , L18701. doi:10.1029/2005GL023684.
Zhang, X., Harvey K. D. , Hogg W. D. , and Yuzyk T. R. , 2001: Trends in Canadian streamflow. Water Resour. Res., 37 , 987–998.
The patterns associated with the leading PC of PREC/L precipitation for (a) 1948–2004, (b) 1979–2004 and (c) CMAP satellite precipitation 1979–2004 (shaded contours), and the correlation between the leading principal component and the 300-hPa streamfunction. The leading PC explains (a) 15.8%, (b) 20.7%, and (c) 10.8% of the total variance in fall precipitation. The pattern is presented as the correlation between the leading principal component (PC1) and precipitation at each gridpoint. (d) The leading principal component time series of the PREC/L and CMAP precipitation. The unfiltered PREC/L time series (thin dashed line) is shown with the smoothed time series (3-yr running average) of the PREC/L (thick black line) and satellite (thick red line) precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
The patterns associated with the leading PC of PREC/L precipitation for (a) 1948–2004, (b) 1979–2004 and (c) CMAP satellite precipitation 1979–2004 (shaded contours), and the correlation between the leading principal component and the 300-hPa streamfunction. The leading PC explains (a) 15.8%, (b) 20.7%, and (c) 10.8% of the total variance in fall precipitation. The pattern is presented as the correlation between the leading principal component (PC1) and precipitation at each gridpoint. (d) The leading principal component time series of the PREC/L and CMAP precipitation. The unfiltered PREC/L time series (thin dashed line) is shown with the smoothed time series (3-yr running average) of the PREC/L (thick black line) and satellite (thick red line) precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
The patterns associated with the leading PC of PREC/L precipitation for (a) 1948–2004, (b) 1979–2004 and (c) CMAP satellite precipitation 1979–2004 (shaded contours), and the correlation between the leading principal component and the 300-hPa streamfunction. The leading PC explains (a) 15.8%, (b) 20.7%, and (c) 10.8% of the total variance in fall precipitation. The pattern is presented as the correlation between the leading principal component (PC1) and precipitation at each gridpoint. (d) The leading principal component time series of the PREC/L and CMAP precipitation. The unfiltered PREC/L time series (thin dashed line) is shown with the smoothed time series (3-yr running average) of the PREC/L (thick black line) and satellite (thick red line) precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
(a),(b) The patterns of the two leading PCs (filled contours) and (c),(d) associated time series of the meridional component of the fall 300-hPa rotational wind. The leading EOFs (shaded contours) of the meridional rotation wind are presented as the correlation between the principal component and the original time series at each gridpoint. The correlation between the leading PC and the fall 300-hPa streamfunction is also shown with a contour interval of 0.10 and the zero contour line in bold. The original (thin dotted) and smoothed (solid) PC time series are shown along with the smoothed PC1 of precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
(a),(b) The patterns of the two leading PCs (filled contours) and (c),(d) associated time series of the meridional component of the fall 300-hPa rotational wind. The leading EOFs (shaded contours) of the meridional rotation wind are presented as the correlation between the principal component and the original time series at each gridpoint. The correlation between the leading PC and the fall 300-hPa streamfunction is also shown with a contour interval of 0.10 and the zero contour line in bold. The original (thin dotted) and smoothed (solid) PC time series are shown along with the smoothed PC1 of precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
(a),(b) The patterns of the two leading PCs (filled contours) and (c),(d) associated time series of the meridional component of the fall 300-hPa rotational wind. The leading EOFs (shaded contours) of the meridional rotation wind are presented as the correlation between the principal component and the original time series at each gridpoint. The correlation between the leading PC and the fall 300-hPa streamfunction is also shown with a contour interval of 0.10 and the zero contour line in bold. The original (thin dotted) and smoothed (solid) PC time series are shown along with the smoothed PC1 of precipitation.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
Precipitation (shaded contours) and column-integrated moisture flux anomalies (vectors) for seasons when (a) PC1 of 300 hPa is strongly positive and PC2 of 300 hPa Vrot is strongly negative and (b) PC1 is strongly negative and PC2 is strongly positive. (c) The difference between (a) and (b). The precipitation anomalies have units of standard deviations; the moisture flux is in kg m−2.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
Precipitation (shaded contours) and column-integrated moisture flux anomalies (vectors) for seasons when (a) PC1 of 300 hPa is strongly positive and PC2 of 300 hPa Vrot is strongly negative and (b) PC1 is strongly negative and PC2 is strongly positive. (c) The difference between (a) and (b). The precipitation anomalies have units of standard deviations; the moisture flux is in kg m−2.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
Precipitation (shaded contours) and column-integrated moisture flux anomalies (vectors) for seasons when (a) PC1 of 300 hPa is strongly positive and PC2 of 300 hPa Vrot is strongly negative and (b) PC1 is strongly negative and PC2 is strongly positive. (c) The difference between (a) and (b). The precipitation anomalies have units of standard deviations; the moisture flux is in kg m−2.
Citation: Journal of Hydrometeorology 11, 5; 10.1175/2010JHM1273.1
