1 OCTOBER 2012

BERMAN ET AL.

6781

Eastern Patagonia Seasonal Precipitation: Influence of Southern Hemisphere Circulation and Links with Subtropical South American Precipitation ANA LAURA BERMAN AND GABRIEL SILVESTRI Centro de Investigaciones del Mar y la Atmo´sfera, Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas and Departamento de Ciencias de la Atmo´sfera y los Oce´anos, Facultad de Ciencias Exactas y Naturales, FCEN, Universidad de Buenos Aires, and Unidad Mixta Internacional: Instituto Franco-Argentino sobre Estudios de Clima y sus Impactos,* Buenos Aires, Argentina

ROSA COMPAGNUCCI Departamento de Ciencias de la Atmo´sfera y los Oce´anos, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, and Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas, Buenos Aires, Argentina (Manuscript received 8 September 2011, in final form 8 April 2012) ABSTRACT Some aspects of the seasonal precipitation over eastern Patagonia, the southernmost area of South America east of the Andes Cordillera, are examined in this paper. Results indicate that the central-north areas, the southern continental region, and the southernmost islands are three independent regions of seasonal precipitation, and that each of them is associated with specific patterns of atmospheric circulation. Precipitation over the central-north region is significantly related to the precipitation over a wide area of southern South America east of the Andes during the four seasons. Enhanced (reduced) precipitation over this area is associated with weakened (intensified) westerly flow in the region. Precipitation over the southern continental area has a close connection with the dipolar pattern of precipitation over subtropical South America during spring, summer, and autumn. The anomalies of atmospheric circulation at low and upper levels associated with the subtropical dipole are also able to modulate the intensity of the westerlies over the south of eastern Patagonia, affecting the regional precipitation. Precipitation over the islands of the southernmost part of eastern Patagonia is connected with subtropical precipitation in summer and winter. The activity of frontal systems associated with migratory perturbations moving to the east along the Southern Hemisphere storm tracks modulates the variability of seasonal precipitation over this region.

1. Introduction Patagonia is a wide region extended over southern South America on both sides of the Andes Cordillera (Fig. 1). The regional atmospheric circulation at lower and upper levels is strongly influenced by the typical westerly flow of subtropical and subpolar latitudes of

* The Unidad Mixta Internacional (3351): Instituto FrancoArgentino sobre Estudios de Clima y sus Impactos is sponsored by the Centre National de la Recherche Scientifique, the Consejo Nacional de Investigaciones Cientı´ficas y Te´cnicas, and the Universidad de Buenos Aires.

Corresponding author address: Ana Laura Berman, DCAO, FCEN, Universidad de Buenos Aires, Intendente Guiraldes 2160, Ciudad Universitaria, Buenos Aires C1428EGA, Argentina. E-mail: [email protected] DOI: 10.1175/JCLI-D-11-00514.1 Ó 2012 American Meteorological Society

the Southern Hemisphere (e.g., Paruelo et al. 1998; Garreaud 2009; Garreaud et al. 2009). The zonal atmospheric flow and transport of humid air from the Pacific Ocean are blocked by the Andes Cordillera, provoking an important contrast between the dry conditions in the Argentinean region east of the Andes, referred to as eastern Patagonia (EPAT) and the humid Chilean side (Prohaska 1976). Regional precipitation is mostly associated with the activity of frontal systems linked with migratory surface cyclones moving to the east along the Southern Hemisphere storm tracks (e.g., Trenberth 1991; Berbery and Vera 1996). Although precipitation takes place over the western slope of the Andes, it can cross over the mountains and fall over the eastern slope and adjacent areas (e.g., Hoffmann 1975). Furthermore, part of the total amount of precipitation is of the orographic type induced by the upward movement of the low-level flow from the Pacific across the

6782

JOURNAL OF CLIMATE

VOLUME 25

FIG. 1. Patagonia region (topography is shaded; units: meters). The five selected meteorological stations and the CMAP grid points are indicated with white circles and black squares, respectively (see the text for more details).

western slope of the mountains. The combination of processes associated with the orography and the atmospheric dynamics produces both the amount and variability of precipitation near the Andes that are higher than in areas far from the mountains (e.g., Jobba´gy et al. 1995). Barrett et al. (2009) analyzed the influence of the Andes Cordillera on the occurrence of both frontal and orographic precipitation over southern South America, finding an increment of total precipitation over the region due to the high topographic barrier. Moreover, Garreaud (2007) showed negative (positive) correlations between the 850-hPa zonal wind and the monthly precipitation over the east (west) side of southern Andes, meaning that precipitation decreases (increases) with intensified westerly flow at the Argentinean (Chilean) side and vice versa [see also Fig. 13 in Compagnucci (2011), provided by R. D. Garreaud 2011, personal communication]. Easterly winds and advection of moisture from the adjacent areas of the Atlantic Ocean can also produce daily precipitation over EPAT (e.g., Mayr et al. 2007a,b). Although the standard deviation of annual precipitation has low magnitudes over EPAT, the interannual standard deviation normalized by the annual mean has a maximum over this region (see Fig. 7 in Garreaud et al. 2009). Changes in the occurrence of precipitation induced by atmospheric circulation variability enhance the vulnerability of areas, with dry climate like EPAT increasing desertification, water erosion, and soil compaction. Such processes can affect different systems (e.g., biosphere) or human activities, producing an important socioeconomic damage. However, the analysis of the

interannual variability of precipitation over EPAT is an issue that has still not received adequate attention, perhaps because EPAT has been historically a region little inhabited, where complete and reliable information of precipitation is available for only a few meteorological stations. To our knowledge, there are few studies published in international journals focused on the variability of precipitation over this region. Aravena and Luckman (2009) made important progress describing some characteristics of the spatiotemporal variability of annual precipitation on both sides of the Andes Cordillera. Gonzalez and Vera (2010) and Gonzalez et al. (2010) described the influence of the Indian and Pacific Oceans on the interannual variability of winter precipitation over northwestern EPAT. These works describe Rossby wave trains extending from both tropical basins toward southern South America, affecting the precipitation over specific areas adjacent to the Andes Cordillera. Moreover, precipitation trends on EPAT were analyzed by Castan˜eda and Gonzalez (2008), finding that precipitation enhanced in areas of the north and south but reduced in regions of the center and west during the second half of the twentieth century. Different authors analyzed forcings of precipitation variability in areas that, among other regions, include EPAT. The pioneer study of Pittock (1980) describes the influence of the Southern Hemisphere circulation patterns on the annual precipitation over southern South America. Alessandro (2005, 2008) analyzed the increment of precipitation over areas of the south of South America by effect of blocking episodes in the surrounding oceans. Moreover, Schneider and Gies

1 OCTOBER 2012

BERMAN ET AL.

(2004) showed the influence of El Nin˜o–Southern Oscillation (ENSO) events on southernmost South American precipitation on both sides of the Andes Cordillera through changes in the westerlies. Furthermore, different authors have demonstrated that the interannual variability of precipitation over areas of EPAT is modulated by the southern annular mode (e.g., Gillett et al. 2006; Silvestri and Vera 2009) and the Southern Oscillation index (e.g., Pittock 1980; Kiladis and Diaz 1989; Aravena and Luckman 2009). In the light of the previous comments, it is clear that there are some aspects of the variability of precipitation over EPAT that must be investigated to have a more complete comprehension of the climate variability in this remote region of the world. In particular, the principal modes of spatiotemporal variability of seasonal precipitation over EPAT, the associated anomalies of atmospheric circulation, and the relations with the precipitation over other areas of southern South America have not been analyzed in previous papers. Therefore, the aim of this study is to investigate such characteristics of the climate variability in southern South America. Results presented in this paper describe not only aspects of the present climate in this region but also can be an important contribution to understand the causes of the notorious climate changes that occurred in the past (e.g., Gilli et al. 2005; Haberzettl et al. 2005). The article is organized as follows: data and methodology are described in section 2, the main results are presented in section 3, and the conclusions are summarized in section 4.

2. Data and methodology a. Data The National Meteorological Service of Argentina (NMSA) provided the monthly-mean values of precipitation corresponding to the meteorological stations Trelew (TRE), Comodoro Rivadavia (CR), Rı´o Gallegos (RG), and Ushuaia (USH) (Fig. 1). Although there are more meteorological stations in the region under study, most of them have large gaps or cover short periods. Only the four selected stations have information without missing values in the daily data with which the NMSA constructed the monthly means. The Chilean station Punta Arenas (ARE) (Fig. 1) is not located in the area of EPAT, but it is included in the analysis for a better description of the seasonal precipitation over southernmost South America. Monthly-mean precipitation corresponding to the station ARE was taken from the Global Historical Climatology Network-Monthly (GHCN-M, version 3). The Climate Prediction Center

6783

Merged Analysis of Precipitation (CMAP) described by Xie and Arkin (1997) is also used. The CMAP data merge observations from rain gauges with precipitation estimates from several satellite-based algorithms covering the area of EPAT in a uniform grid of 2.58 latitude 3 2.58 longitude resolution (Fig. 1). An additional point that gives singular importance to the CMAP dataset is the fact that it allows the study of connections between the precipitation over areas of EPAT and precipitation over other regions of South America. Atmospheric conditions in the Southern Hemisphere are described using monthly-mean fields of geopotential height at 850 hPa (Z850) and 200 hPa (Z200) from the National Centers for Environmental Prediction– National Center for Atmospheric Research (NCEP– NCAR) reanalysis (Kalnay et al. 1996). The study is made for the period 1979–2009 because it is the period covered by the available dataset of precipitation (NMSA and CMAP). Moreover, during this period the NCEP–NCAR reanalysis includes information from satellites, assuring a better quality of the data at high latitudes of the Southern Hemisphere. Monthly anomalies are defined as departures from the corresponding monthly means computed over the entire 1979–2009 period. Seasonal means of such monthly anomalies are calculated for all dataset considering the seasons as in the Southern Hemisphere: December– February (summer), March–May (autumn), June–August (winter), and September–November (spring).

b. Methodology Linear correlations among the time series of seasonal precipitation registered in the five selected meteorological stations are analyzed as a first approach of the spatial homogeneity. The CMAP dataset was used in several investigations of precipitation variability over tropical and subtropical areas of South America (e.g., Zhou and Lau 2001; Berbery and Barros 2002), but there are no antecedents about the use of this dataset in studies focused on southern regions. Therefore, seasonal correlations between the precipitation from the meteorological stations and the CMAP data are analyzed to demonstrate that the precipitation CMAP and the five stations represent the same time variability and spatial relations. Consequently, the analysis of the spatiotemporal variations of regional precipitation is made considering the CMAP data. The principal modes of seasonal precipitation variability are obtained by principal component analysis (PCA) applied to the CMAP data in the area of EPAT (see Fig. 1) using a correlation input matrix among the grid points’ time series that is, namely, the S mode (e.g.,

6784

JOURNAL OF CLIMATE

VOLUME 25

TABLE 1. Correlations between the time series of precipitation corresponding to the five meteorological stations depicted in Fig. 1. One (two) asterisk(s) indicate correlations that are statistically significant at the 95% (99%) level for a Student’s t test. There are no values significant at the 90% level. Summer TRE CR RG USH ARE

0.65** 20.10 20.07 20.22

CR 20.15 0.03 20.28

Autumn RG

USH

0.03 0.53**

0.25

RG

USH

CR RG USH ARE

TRE

CR

RG

0.44* 0.05 0.13 20.20

0.27 20.07 0.10

20.15 0.56**

Winter

CR RG USH ARE

TRE

CR

0.58** 0.58** 20.21 20.01

0.41* 20.17 0.01

USH

0.00

Spring

20.26 0.48**

0.08

Richman 1986). The analytic formulation as well as specific characteristics of this procedure were presented and discussed by Compagnucci and Richman (2008). This methodology defines the pattern time series for clusters of points with similar covariability [principal component (PC) scores] and the concomitant areas where these pattern time series are representative (PC loadings) (e.g., Castan˜eda and Compagnucci 2005). Therefore, the PC scores of each mode, which are nondimensional time series, can be considered as a time index representing the precipitation variability in a specific area of EPAT. The varimax orthogonal rotation (Kaiser 1958, 1959) is applied to eliminate possible degeneration in the solutions due to eigenvectors’ sampling errors (North et al. 1982) and to improve the accuracy of the solutions (Richman 1986). The number of retained components is preview by the Scree test (Cattell 1966) and determined through an iterative rotation from two-component retention and adding one more in each step until the Tucker’s congruence coefficients, which test the resulting matches between components and the dataset, give values in the range of 0.92–0.82, considered the borderline match (Richman 1986). To capture both the explained variance and the rate of variability, we calculate the correlation and regression coefficients between the PC scores obtained from the PCA of precipitation CMAP over EPAT (referred as PCACMAP) and the precipitation over the South American continent south of 108S. The same relations are calculated with the fields of Z850 and Z200 to analyze the anomalies of atmospheric circulation at lower and upper levels, respectively. The analysis for the upper troposphere was also performed considering the streamfunction at 200 hPa from the NCEP–NCAR reanalysis, but the results are similar to those of Z200 even in tropical latitudes.

TRE CR RG USH ARE

0.49** 0.05 20.10 0.03

CR 0.08 0.26 0.07

RG

USH

20.12 0.51**

20.15

Regressions with Z850 and Z200 are performed with the standardized series of both variables because the PC scores are standardized series. Therefore, regressions presented in this paper do not have dimensions. A linear trend was removed from all analyzed time series. Special care was taken in the calculation of all correlations and regressions, ensuring that neither was dominated by few specific extreme cases.

3. Results Correlations among the five selected series of precipitation describe a common pattern of spatial relations during summer, autumn, and spring (Table 1). In fact, three common characteristics are detected in these seasons: 1) significant linear relations among the northeastern stations TRE and CR, but both are not connected with the other stations; 2) significant relations among the southern stations RG and ARE, but both are independent of the other ones; and 3) the southernmost station USH has no connection with any other station. A similar pattern is observed in winter, with the only exception that the southern station RG has a significant connection with the northern stations TRE and CR. These relations for seasonal precipitation are, in general, in agreement with the conclusions of Aravena and Luckman (2009) for annual means. The poor relation of the precipitation in Ushuaia with the nearby stations Rio Gallegos and Punta Arenas could be associated with the very particular location of this station. In fact, Ushuaia is located in the area of the Cordillera Darwin (see Fig. 1 in Schneider and Gies 2004), which means that very special conditions can affect the local precipitation. Relations mentioned in the previous paragraph are suitably represented by the CMAP data. The spatial

1 OCTOBER 2012

BERMAN ET AL.

6785

FIG. 2. Correlations between the time series of precipitation corresponding to the five meteorological stations depicted in Fig. 1 and the CMAP data during summer. Areas where positive (negative) values are statistically significant at the 90%, 95%, and 99% level of a Student’s t test are shaded in light, medium and dark gray, respectively. Contours: 60.2, 60.3, 60.37, 60.47, 60.60, and 60.80. Negative contours are dashed.

correlation fields of each station’s precipitation time series with CMAP data show that CMAP is consistent with precipitation measured at the five meteorological stations. Figure 2 describes these relations during summer, and similar features are detected in autumn, winter, and spring (figures not shown). Precipitation registered at the stations TRE and CR have significant positive correlation with CMAP over a wide area of northern EPAT (Figs. 2a and 2b), precipitation at the stations RG and ARE have significant positive correlation with CMAP over the southern portion of EPAT (Figs. 2c and 2d), and precipitation at USH has significant positive correlation with CMAP over the island of Tierra del Fuego (Fig. 2e). Moreover, correlation among time series of precipitation at the meteorological stations and CMAP corresponding to the closest point to each station (Fig. 1) results, in all cases, higher than 0.85. Such specific series of CMAP precipitation have among themselves the same patterns of relation asthose described by Table 1 for the meteorological stations (not shown). A more detailed analysis is performed in the following subsection, but this preliminary description indicates that the gridded CMAP precipitation reproduces the main characteristics of time variability and spatial relations detected by the meteorological observations in EPAT. Therefore, the leading modes of seasonal precipitation over EPAT and the correlations extended over the South American continent are obtained from the CMAP dataset. These modes of precipitation variability and the associated anomalies of atmospheric circulation are described for each season in the following subsections.

a. Subregions of EPAT precipitation and links with subtropical areas The Scree test for the eigenvalues of the PCACMAP suggests that only the first three components

are significant because they are clearly outside of the noise tail (not shown). Furthermore, the congruence coefficients of the iteratively varimax rotated solution are lower than the 0.82 limit if four or more component are retained. Therefore, we considered the first three varimax rotated components of the PCA-CMAP, which explain almost 70% of the total variance and define similar homogenous subregions of precipitation variability in the four seasons (Fig. 3). These components subdivide the precipitation CMAP over EPAT in regions centered near the location of the meteorological stations Trelew, Rio Gallegos, and Ushuaia, coinciding with the subregions inferred from Table 1 and Fig. 2. The corresponding PC scores are depicted in Fig. 4. These time series represent the temporal variability of standardized anomalies of seasonal precipitation in each subregion. In particular, the occurrence of extreme seasonal precipitation can be detected considering such series. Center and north of EPAT is a homogeneous area defined by the component that explains more variance in summer, winter, and spring, whereas it is the second in order of explained variance in autumn (pattern 1 in Fig. 3). This component has strong positive correlation with the precipitation over a broad area of the surrounding Atlantic, the Pampa region (center of Argentina), and the eastern Pacific in 258S–408S. The variability of precipitation over most of the continental area of EPAT south of 458S corresponds to the second component in order of explained variance in all seasons except in autumn, when it is the component of higher variance (pattern 2 in Fig. 3). Except in winter, this subregion has strong connections with the dipolar pattern of precipitation typical of subtropical South America east of the Andes. In fact, significant positive values of correlation are observed

6786

JOURNAL OF CLIMATE

VOLUME 25

FIG. 3. Patterns of seasonal correlations between the first three PC score time series of the precipitation in EPAT and the CMAP data. Areas where positive (negative) values are statistically significant at the 90%, 95%, and 99% level of a Student’s t test are shaded in light, medium, and dark red (green), respectively. Contours: 60.2, 60.3, 60.37, 60.47, 60.60, and 60.80. The variance explained by each PCACMAP mode is indicated in the corresponding panel.

1 OCTOBER 2012

BERMAN ET AL.

6787

FIG. 4. PC scores corresponding to the PCA-CMAP patterns depicted in Fig. 3 for pattern 1: blue line, pattern 2: red line, and pattern 3: green line.

in the northeast of Argentina, southern Brazil, western Paraguay, and north of Uruguay [the region referred to as southeastern South America (SESA)] during the period spring–summer–autumn. In contrast, significant negative correlations are detected in an elongated band extending from the coast of Brazil at 208S to, at least, 358S in the surrounding Atlantic. Negative correlations with precipitation over the eastern part of the central Andes onto the plateau known as the Bolivian Altiplano are also detected in summer and autumn. The variability of precipitation over the surroundings of the island of Tierra del Fuego at the southernmost part of EPAT defines the third subregion explaining less variance in the four seasons (pattern 3 in Fig. 3). During summer, inverse connections are observed with the variability of precipitation over the surroundings of the Andes in 208–408S and over the Bolivian Altiplano. Moreover, inverse relations are detected with precipitation over the Pampa region and the adjacent Atlantic in winter.

b. Anomalies of atmospheric circulation related to EPAT precipitation 1) PATTERN 1 Correlations between precipitation over the centernorth of EPAT (pattern 1 of Fig. 3) and the atmospheric circulation at low levels (Z850) show negative

nonsignificant relations over most of subtropical South America and the adjacent Pacific Ocean during summer (Fig. 5a1). Positive relations are observed at midand subpolar latitudes being significant in areas of the southern Pacific and Atlantic Oceans. A similar structure of atmospheric circulation over Patagonia is detected in the upper troposphere, where the magnitudes in the center of negative correlations over the southern continent are significant (Fig. 5b1). Significant relations are also detected at all levels between 700 and 200 hPa (not shown), indicating that weakened westerlies take place in the entire troposphere. These relations imply weakened (intensified) westerly flow at low levels over southern South America associated with more (less) precipitation over the center-north of EPAT. In autumn, the relations of this subregion with the atmospheric circulation at low and upper levels are different from those observed in summer. In fact, the negative relations at low levels over the continental subtropical region are located over the western sector of the continent and over the Pacific, shifted to the west with respect to the position in summer (Fig. 5a2). Furthermore, the center of positive relations located at high latitudes over the Pacific is shifted to the east, forming a coupled meridional dipole with the subtropical negative center extended across the entire troposphere (Fig. 5b2). These significant anomalies imply weakened low-level westerlies over Patagonia, producing more

6788

JOURNAL OF CLIMATE

VOLUME 25

FIG. 5. Seasonal correlations between the first PC score time series of the precipitation CMAP in EPAT (corresponding to pattern 1 in Fig. 3) and (left) Z850 and (right) Z200. Areas with statistically significant positive (negative) values at 90%, 95%, and 99% level for a Student’s t test are shaded in light, medium, and dark red (green), respectively. Solid (dashed) black lines indicate positive (negative) regressions (contours: 0.1; see text for more details).

precipitation over northern EPAT. Moreover, the negative correlations with low-level circulation in northwestern Argentina suggest an intensification of the northerly transport of humid air over areas to the east

of the Andes that could reach the north of EPAT, producing more precipitation in this region. Inverse anomalies of atmospheric circulation are associated with less precipitation over EPAT.

1 OCTOBER 2012

BERMAN ET AL.

During winter, the anomalies of atmospheric circulation and associated processes are similar to those described in autumn (Figs. 5a3 and 5b3). The structure of circulation in the lower troposphere corresponds to the pattern described by Rutllant and Fuenzalida (1991) and Compagnucci and Vargas (1998) associated with abundant and frequent snowfalls over the central Andes the winter after an El Nin˜o year. The connection between this mode of precipitation variability and a dipolar structure of atmospheric circulation extended through the whole troposphere with centers at subtropical and subpolar latitudes of the eastern Pacific persists during spring but with lower values of correlation (Figs. 5a4 and 5b4). Therefore, enhanced (reduced) precipitation over this subregion of EPAT is associated with weakened (intensified) westerlies in the area.

2) PATTERN 2 During summer, the low-level circulation associated with precipitation over the southern continental area of EPAT (pattern 2 in Fig. 3) is characterized by a cyclonic center covering almost all of the South American continent south of 208S (Fig. 6a1). Anticyclonic centers at subpolar latitudes over the Amundsen and Bellingshausen Seas and over the subtropical Atlantic near the South American coast are also observed. This pattern of correlation implies weakened (intensified) westerly flow over southern EPAT associated with enhanced (reduced) precipitation over the region. The structure of upper-level circulation resembles the Pacific–South America 2 pattern (PSA2; e.g., Mo and Higgins 1998) extending from the east of Australia to southern South America and western Atlantic (Fig. 6b1). Previous studies demonstrated the influence of PSA2 on the summer precipitation over tropical and subtropical South America (e.g., Mo and Paegle 2001), but our results show that this pattern of atmospheric circulation also affects the precipitation over the southernmost areas of the continent. Moreover, this structure of lowand upper-level circulation agrees with the pattern described by Diaz and Aceituno (2003) for episodes of enhanced convective cloudiness over SESA. In fact, they demonstrated that enhanced convection over SESA during austral spring and summer is associated with weakened convection over the South Atlantic convergence zone (SACZ). The atmospheric conditions during such episodes are mainly characterized by anomalous anticyclonic circulation at the middle and upper troposphere centered over the subtropical western Atlantic. This anomaly is part of a wavelike quasi-barotropic structure extending across the Pacific from Australia to New Zealand, and it is connected with a stronger-than-

6789

average subtropical jet stream and a reinforced northwesterly flow of warm and humid air toward SESA (e.g., Vera et al. 2006 and references herein). Similar characteristics were detected by Doyle and Barros (2002) in the analysis of low-level water vapor transport during anomalous precipitation events connected with the SACZ-SESA dipole. The studies of Doyle and Barros (2002) and Diaz and Aceituno (2003) were focused on the variability of convective cloudiness and precipitation over subtropical South America east of the Andes. In contrast, our analysis has been focused on the variability of precipitation over the southernmost areas of the South American continent. However, we found that the circulation patterns associated with the precipitation over EPAT are similar to those suggested by the other authors for the area SESA-SACZ. These connections between EPAT and SESA-SACZ have not been previously addressed in the scientific literature. The anomalies of low-level circulation in autumn are quite similar to those observed in summer (Fig. 6a2). The main differences are the lower magnitudes of negative correlations over the continent and the increment of positive correlations at polar latitudes. The most prominent feature of the anomalous circulation at upper levels is the structure of opposite phases such as those typically associated with the southern annular mode (SAM) (Fig. 6b2). A wave train similar to the Pacific– South America 1 pattern (PSA1; Mo and Higgins 1998) is also detected. These influences of the SAM and PSA1 on the variability of precipitation over the southernmost areas of South America east of the Andes during autumn have not been described in previous studies. During winter, the anomalies of atmospheric circulation associated with precipitation over the south of EPAT are different from those observed in summer and autumn. In fact, the anomalous low-level circulation is characterized by an anticyclonic center over the Drake Passage and a cyclonic center extended in the southern Pacific reaching the Chilean coast (Fig. 6a3). Such a pattern of atmospheric circulation can be associated with weakened westerlies over southernmost South America, favoring the increment of precipitation over southern areas of EPAT. Inverse anomalies of circulation imply intensified westerlies associated with less precipitation. The weak connection between precipitation over the south of EPAT and the regional atmospheric circulation can be the cause of the lack of relation between precipitation over this region and over SESA (see Fig. 3b3). Moreover, the upper-level circulation does not show wave trains (PSA2, etc.) acting as remote forcing of the variability of precipitation over southern EPAT (Fig. 6b3). This characteristic could be explained by the results of Berbery et al. (1992), who

6790

JOURNAL OF CLIMATE

VOLUME 25

FIG. 6. As in Fig. 5, but for the second PC score time series (corresponding to pattern 2 in Fig. 3).

demonstrated that the latitudinal gradient of absolute vorticity hinders the meridional propagation of waves emanating from the western tropical Pacific during austral winter.

In spring, the anomalies of atmospheric circulation at both low and upper levels recover the characteristics observed in summer and autumn. In fact, a center of negative correlation (cyclonic anomalies) over southern

1 OCTOBER 2012

BERMAN ET AL.

South America and the adjacent Atlantic together with a center of positive correlation (anticyclonic anomalies) over the Amundsen and Bellingshausen Seas are detected at low levels (Fig. 6a4). A PSA1 is clearly detected at upper levels as well as the typical structure of the SAM (Fig. 6b4). Although the influence of the SAM on the variability of precipitation over areas of southern South America during austral spring was suggested by Silvestri and Vera (2009), the simultaneous impact of this atmospheric mode and the PSA1 on EPAT has not been described in previous studies. The consequences of these atmospheric anomalies on the precipitation over EPAT and over subtropical areas of South America have been discussed in the previous paragraphs. The dynamic mechanisms connecting the variability of precipitation over both regions of the continent (see Fig. 3b4) are clear.

3) PATTERN 3 During summer, the anomalies of atmospheric circulation associated with precipitation over the islands of southernmost EPAT (pattern 3 in Fig. 3) have characteristics opposite of those detected in the other two modes of precipitation. In fact, precipitation over southernmost EPAT has a positive relation with the low-level circulation over most of the South American continent south of 308S, whereas negative relations are detected at subpolar and polar latitudes (Fig. 7a1). The anomalous circulation extends to the upper levels (Fig. 7b1), generating reinforced westerly flow in the entire troposphere over the south of South America. These conditions favor the passage of migratory perturbations that move to the east, increasing the precipitation in the region. The lack of significant correlations could be a consequence of strong interdiurnal changes affecting the seasonal means or strong interannual changes, which are not analyzed in this study. Moreover, the anticyclonic center over the central Andes can inhibit the upward movements over the western slopes of the mountains, reducing the precipitation in these regions. It could be the mechanism that explains the inverse relation between precipitation over southernmost EPAT and precipitation over the central Andes and the Bolivian Altiplano (see Fig. 3c1). In autumn, the anomalies of atmospheric circulation associated with precipitation over this subregion are different from those observed in summer, and the anomalous pattern suggests the important influence of cold fronts on the precipitation variability. In fact, the inverse relations between the precipitation and the lowlevel circulation in areas of southern South America and the southern Atlantic configure a structure of atmospheric anomalies favorable to outbreaks of cold air from polar latitudes to the south of EPAT (Fig. 7a2).

6791

Although the correlations are not significant, the inverse relation between the precipitation and the upper-level circulation over the Drake Passage and the southern Atlantic is also detected (Fig. 7b2). The cyclonic center extending into the entire troposphere over the Drake Passage and the surrounding South Atlantic is more intense during winter (Figs. 7a3 and 7b3). The frontal activity (cold fronts) associated with this anomalous circulation can explain the occurrence of more precipitation over southernmost EPAT. At the same time, the cyclonic circulation can reduce the flow of moisture from the subtropical Atlantic to eastern South America, explaining the inverse relation between precipitation over southernmost EPAT and precipitation over SESA (see Fig. 3c3). The inverse relations between precipitation over this subregion and the atmospheric circulation over the Drake Passage and the surrounding Atlantic persist during spring (Figs. 7a4 and 7b4). As was previously described, this structure of anomalous circulation can produce an increment of precipitation over southernmost EPAT due to the activity associated with the passage of cold fronts.

c. Relations between precipitation over the east and west sides of the southern Andes Section 1 mentioned that the low-level zonal wind in Patagonia is negatively (positively) correlated with monthly precipitation variations to the east (west) of the Andes Cordillera. These relationships are also clearly detected in the seasonal time scale considering indexes of zonal wind and precipitation on both sides of the mountains. In fact, the low-level zonal wind averaged in the northern and southern areas of Patagonia (Fig. 8) has significant negative (positive) correlation with precipitation over regions to the east (west) of the Andes during the four seasons (Table 2). However, poor relations are found among precipitation registered at the same latitude over both sides of the mountains, describing a lack of west–east dipole between the hyperhumid conditions to the west of the Andes and the semiarid conditions to the east. This characteristic is also observed in the seasonal correlations of CMAP data with the PC scores of the PCA-CMAP (see Fig. 3). The zonal wind has influence on the seasonal variability of EPAT precipitation, but the effects of other forcings can produce precipitation over EPAT that does not vary in phase with that registered to the west of the Andes. Results presented in the previous section suggest that precipitation over northern EPAT can be influenced by the northerly transport of humid air over areas to the east of the Andes in autumn and winter. Blocking anticyclones at higher latitudes over the western South

6792

JOURNAL OF CLIMATE

VOLUME 25

FIG. 7. As in Fig. 5, but for the third PC score time series (corresponding to pattern 3 in Fig. 3).

Atlantic produce easterly winds and advection of moisture from the adjacent areas of the Atlantic Ocean that penetrates into the eastern continent up to hundreds of kilometers, producing stratiform daily

precipitation over EPAT (e.g., Prohaska 1976; Mayr et al. 2007b). Intense daily precipitation over different regions of EPAT is also associated with persistence of strong easterly winds and advection of moisture from the

1 OCTOBER 2012

6793

BERMAN ET AL.

TABLE 2. Correlations between the time indexes of zonal wind and precipitation corresponding to the areas depicted in Fig. 8. One (two) asterisk(s) indicate correlations that are statistically significant at the 95% (99%) level for a Student’s t test. There are no values significant at the 90% level. Summer

NWIND NWPP

NWPP

NEPP

0.43*

20.42* 20.03

NWPP

NEPP

0.55**

20.58** 0.02

SWPP

SEPP

0.60**

20.50** 20.09

SWPP

SEPP

SWIND SWPP

0.44*

20.59** 20.14

SWPP

SEPP

SWIND SWPP

0.51**

20.46* 20.08

SWPP

SEPP

SWIND SWPP

0.63**

20.48** 20.11

SWIND SWPP

Autumn

NWIND NWPP Winter

NWIND NWPP FIG. 8. Areas considered for indexes of 850-hPa zonal wind and precipitation in Patagonia. The two boxes limited by solid lines indicate the areas where the zonal wind is averaged to define the indexes for northern (NWIND) and southern (SWIND) Patagonia. The four shaded boxes indicate the areas where the CMAP data are averaged to define the indexes of precipitation for northwestern (NWPP), northeastern (NEPP), southwestern (SWPP), and southeastern (SEPP) Patagonia.

Atlantic (Frumento 2000; Mayr et al. 2007a). Although these daily events are not detected with the seasonal correlations analyzed in this paper, they are additional local processes that explain part of the precipitation variability in areas of EPAT contributing to the lack of relation with precipitation variability registered to the west of the southern Andes. Such local forcings can also contribute to the different variability of precipitation over the northern and southern continental areas of EPAT (patterns 1 and 2 of Fig. 3; see Fig. 4), even though both regions are influenced by the westerlies.

4. Concluding remarks Different aspects of the variability of seasonal precipitation over EPAT, the South American region south of 408S east of the Andes, have been described in this paper. Connections between precipitation over this remote region of the world and precipitation over other areas of South America were studied as well as the associated anomalies of atmospheric circulation. Three homogeneous subregions of precipitation extending over almost similar areas of EPAT are detected in the four seasons. In fact, the center-north region, the southern continental area, and the southernmost islands are independent regions of variability of

NWPP

NEPP

0.54**

20.44* 20.03

Spring

NWIND NWPP

NWPP

NEPP

0.57**

20.48** 0.09

seasonal precipitation. Moreover, each subregion is associated with specific anomalies of atmospheric circulation and has particular links with precipitation over other areas of South America. Precipitation over the center-north region of EPAT has significant positive correlation with precipitation over the surrounding Atlantic and the Pampa region (center of Argentina) during the four seasons. The enhancement (reduction) of seasonal precipitation over this part of EPAT is associated with weakened (enhanced) westerly flow over the region. Enhanced (weakened) northerly transport of humid air over areas to the east of the Andes can also produce more (less) precipitation in the region during autumn and winter. Although the westerlies have influence on precipitation over both sides of the Andes Cordillera, the characteristics of the low-level flow in autumn–winter affecting the precipitation over the north of EPAT contribute to the lack of negative correlation between precipitation over this region and that registered over areas west of the mountains. Daily precipitation associated with easterly winds and strong advection of moisture from the adjacent Atlantic toward the north of EPAT described in previous studies can contribute to the lack of inverse relation among precipitation over both sides of the Andes during the four seasons. Precipitation over the southern continental subregion has a strong connection with precipitation over

6794

JOURNAL OF CLIMATE

subtropical areas east of the Andes during spring, summer, and autumn. In fact, positive correlations are detected with precipitation over the SESA region, while negative connections are observed in an elongated band extended along the coast of Brazil and the surrounding Atlantic in 208S–358S. The connection between precipitation over subpolar and subtropical regions is established via an atmospheric pattern constituted by anomalous cyclonic circulation over most of southern South America at low levels associated with the Pacific– South America wave train extending from the western Pacific at upper levels. A similar structure of atmospheric circulation was presented by Diaz and Aceituno (2003) for episodes of enhanced convective cloudiness over subtropical areas, but our results indicate that this atmospheric pattern is also able to connect the variability of precipitation over subpolar and subtropical regions. In Patagonia, the anomalous cyclonic (anticyclonic) circulation implies weakened (enhanced) westerly flow that increases (reduces) the precipitation over southern EPAT. In subtropical latitudes, strong (weak) subtropical jet stream and reinforced (weakened) northwesterly flow of warm and humid air produce more (less) precipitation over SESA but a weakened (reinforced) SACZ, reducing (increasing) the precipitation farther to the north. This subpolar–subtropical precipitation link and the influence of the Pacific–South America patterns on the variability of precipitation over southernmost South America have not been previously addressed in the scientific bibliography. As in the northern areas, precipitation over both sides of the Andes is modulated by the westerlies, but there is a lack of negative correlation among precipitation over southern EPAT and that registered at the same latitude to the west of the Andes. Additional local forcings of EPAT precipitation variability could produce such disconnection. In particular, previous studies described events of intense daily precipitation over southern areas of EPAT associated with strong easterly winds and advection of moisture from adjacent areas of the Atlantic. Such local processes could also contribute to the lack of significant relation between the variability of precipitation over the northern and southern continental areas of EPAT even though both regions are influenced by the westerlies. Precipitation over the islands located in the southernmost area of EPAT is associated with the activity of frontal systems in the region. Moreover, precipitation over these islands is not connected with precipitation over other regions of southern South America during autumn and spring. Nevertheless, significant negative correlations with precipitation over areas of the central Andes and the Pampa region are detected in summer and winter, respectively.

VOLUME 25

Results presented in this paper contribute to a better knowledge of the variability of seasonal precipitation over southern South America. This issue is very important not only for descriptions of the present climate but also to infer past conditions. In fact, the information from paleoclimatic deposits located in southernmost South America reveals important climate changes in the region (e.g., Boninsegna et al. 2009). Our results can help to understand the possible conditions in this remote region of the world during specific past periods as well as the causes of such climate fluctuations. Acknowledgments. Comments and suggestions provided by three anonymous reviewers were very helpful in improving this paper. The study was financed by Grants UBACYT EX016, UBACYT 20020100101049, AGENCIA-MINCYT PICT-2007-00438, PICT-20102110, CONICET-PIP-114-201001-00250, and MINCYTMEYS-ARC/11/09. REFERENCES Alessandro, A., 2005: Bloqueos simultaneos en el Atlantico y Pacifico Sur y sus influencias sobre la Republica Argentina. Rev. Bras. Meteor., 20, 277–300. ——, 2008: Temperature and precipitation conditions in Argentina associated with strong westerly mid-latitute. Rev. Bras. Meteor., 23, 126–142. Aravena, J.-C., and B. H. Luckman, 2009: Spatio-temporal rainfall patterns in southern South America. Int. J. Climatol., 29, 2106–2120. Barrett, B. S., R. D. Garreaud, and M. Falvey, 2009: Effect of the Andes Cordillera on precipitation from a midlatitude cold front. Mon. Wea. Rev., 137, 3092–3109. Berbery, E., and C. Vera, 1996: Characteristics of the Southern Hemisphere winter storm track with filtered and unfiltered data. J. Atmos. Sci., 53, 468–481. ——, and V. Barros, 2002: The hydrologic cycle of the La Plata basin in South America. J. Hydrometeor., 3, 630–645. ——, J. Nogues-Paegle, and J. Horel, 1992: Wavelike Southern Hemisphere extratropical teleconnections. J. Atmos. Sci., 49, 155–177. Boninsegna, J. A., and Coauthors, 2009: Dendroclimatological reconstructions in South America: A review. Palaeogeogr. Palaeoclimatol. Palaeoecol., 281, 210–228. Castan˜eda, M., and R. Compagnucci, 2005: Temporal variability of lower stratosphere temperature. Stud. Geophys. Geod., 49, 573–596. ——, and M. Gonzalez, 2008: Statistical analysis of the precipitation trends in the Patagonia region in southern South America. Atmo´sfera, 21, 303–317. Cattell, R., 1966: The Scree test for the number of factors. Multivariate Behav. Res., 1, 245–276. Compagnucci, R. H., 2011: Atmospheric circulation over Patagonia since the Jurassic to present: A review through proxy data and climatic modeling scenarios. Biol. J. Linn. Soc., 103, 229–249. ——, and W. M. Vargas, 1998: Inter-annual variability of the Cuyo rivers’ streamflow in the Argentinean Andean mountains and ENSO events. Int. J. Climatol., 18, 1593–1609.

1 OCTOBER 2012

BERMAN ET AL.

——, and M. B. Richman, 2008: Can principal component analysis provide atmospheric circulation or teleconnection patterns? Int. J. Climatol., 28, 703–726. Diaz, A., and P. Aceituno, 2003: Atmospheric circulation anomalies during episodes of enhanced and reduced convective cloudiness over Uruguay. J. Climate, 16, 3171–3185. Doyle, M., and V. Barros, 2002: Midsummer low-level circulation and precipitation in subtropical South America and related sea surface temperature anomalies in the South Atlantic. J. Climate, 15, 3394–3410. Frumento, O., 2000: The severe precipitation event of April 1998 in north-east Patagonia. Preprints, Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Boulder, CO, Amer. Meteor. Soc., 406–407. Garreaud, R. D., 2007: Precipitation and circulation covariability in the extratropics. J. Climate, 20, 4789–4797. ——, 2009: The Andes climate and weather. Adv. Geosci., 7, 1–9. ——, M. Vuille, R. Compagnucci, and J. Marengo, 2009: Presentday South American climate. Palaeogeogr. Palaeoclimatol. Palaeoecol., 281, 180–195. Gillett, N., T. D. Kell, and P. D. Jones, 2006: Regional climate impacts of the southern annular mode. Geophys. Res. Lett., 33, L23704, doi:10.1029/2006GL027721. Gilli, A., D. Ariztegui, F. S. Anselmetti, J. A. McKenzie, V. Markgraf, I. Hajdas, and R. D. McCulloch, 2005: MidHolocene strengthening of the southern westerlies in South America—Sedimentological evidences from Lago Cardiel, Argentina (498S). Global Planet. Change, 49, 75–93. Gonzalez, M., and C. Vera, 2010: On the interannual wintertime rainfall variability in the southern Andes. Int. J. Climatol., 30, 643–657. ——, M. Skansi, and F. Losano, 2010: A statistical study of seasonal winter rainfall prediction in the Comahue region (Argentina). Atmo´sfera, 23, 277–294. Haberzettl, T., and Coauthors, 2005: Climatically induced lake level changes during the last two millennia as reflected in sediments of Laguna Potrok Aike, southern Patagonia (Santa Cruz, Argentina). J. Paleolimnol., 33, 283–302. Hoffmann, J. A. J., 1975: Maps of mean temperature and precipitation. Climatic Atlas of South America, Vol. 1, WMO/ UNESCO, 1–28. Jobba´gy, E. G., J. M. Paruelo, and R. J. C. Leo´n, 1995: Estimacio´n del re´gimen de precipitacio´n a partir de la distancia a la cordillera en el noroeste de la Patagonia. Ecol. Austral, 5, 47–53. Kaiser, H., 1958: The varimax criterion for analytic rotation in factor analysis. Psychometrika, 23, 187–200. ——, 1959: Computer program for varimax rotation in factor analysis. Educ. Psychol. Meas., 19, 413–420.

6795

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471. Kiladis, G., and H. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2, 1069–1090. Mayr, C., and Coauthors, 2007a: Holocene variability of the Southern Hemisphere westerlies in Argentinean Patagonia (528S). Quat. Sci. Rev., 26, 579–584. ——, and Coauthors, 2007b: Precipitation origin and evaporation of lakes in semi-arid Patagonia (Argentina) inferred from stable isotopes (d18O, d2H). J. Hydrol., 334, 53–63. Mo, K. C., and R. Higgins, 1998: The Pacific–South American modes and tropical convection during the Southern Hemisphere winter. Mon. Wea. Rev., 126, 1581–1596. ——, and J. N. Paegle, 2001: The Pacific–South American modes and their downstream effects. Int. J. Climatol., 21, 1211–1229. North, G., T. Bell, R. Cahalan, and F. Moeng, 1982: Sampling errors in the estimation of empirical orthogonal functions. Mon. Wea. Rev., 110, 699–706. Paruelo, J., A. Beltran, E. Jobbagy, O. Sala, and R. Golluscio, 1998: The climate of Patagonia: General patterns and controls on biotic processes. Ecol. Austral, 8, 85–101. Pittock, A. B., 1980: Patterns of climatic variation in Argentina and Chile—I. Precipitation, 1931–60. Mon. Wea. Rev., 108, 1347–1361. Prohaska, F., 1976: The climate of Argentina, Paraguay and Uruguay. Climates of Central and South America, W. Schwerdtfeger, Ed., World Survey of Climatology, Vol. 12, Elsevier, 13–72. Richman, M., 1986: Rotation of principal components. J. Climatol., 6, 293–335. Rutllant, J., and H. Fuenzalida, 1991: Synoptic aspects of the central Chile rainfall variability associated with the Southern Oscillation. Int. J. Climatol., 11, 63–76. Schneider, C., and D. Gies, 2004: Effects of El Nin˜o–Southern Oscillation on southernmost South America precipitation at 538S revealed from NCEP–NCAR reanalysis and weather station data. Int. J. Climatol., 24, 1057–1076. Silvestri, G., and C. Vera, 2009: Nonstationary impacts of the southern annular mode on Southern Hemisphere climate. J. Climate, 22, 6142–6148. Trenberth, K., 1991: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48, 2159–2178. Vera, C., and Coauthors, 2006: Toward a unified view of the American monsoon systems. J. Climate, 19, 4977–5000. Xie, P., and P. 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. Zhou, J., and K.-M. Lau, 2001: Principal modes of interannual and decadal variability of summer rainfall over South America. Int. J. Climatol., 21, 1623–1644.