Relationship between values and trends of two circulation indices and temperature and rainfall in Argentina


A. P. Alessandro, Dto. de Cs. de la Atmósfera y los Océanos, FCEN, UBA, Buenos Aires, Argentina. E-mail:


Extreme values of two circulation indices are examined for the time series covering the period 1960–2006. The influence of circulation on surface temperature and rainfall in Argentina is analysed using a zonal index Z and a meridional index R at the 1000 and 500 hPa levels. Seasonal means for all the years representing extreme negative values of R(Z) at 1000 and/or 500 hPa, determine mainly positive (negative) temperature anomalies with increased rainfall in both indices. These characteristics revert for extreme positive values of R(Z). At 1000 hPa, seasonal anomaly trends of R are negative and significant. Trends of Z are all positive and significant, except for winter when they are not significant. Signs of trends at 500 hPa are the same as at 1000 hPa, except for Z in winter and R in summer, whose values are not significant. Annual and seasonal trends of the anomalies of Z differences between 500 and 1000 hPa, are mostly negative, which implies a decrease in baroclinicity in southernmost South America. Analysis of the statistical values obtained for both variables generated by extreme values of R and Z, and of the signs and significance of index trends at 1000 hPa indicates that both circulation indices might contribute to surface temperature warming during the last years of the period 1960/2006. The increase in R, whose influence is greater than that of Z, might contribute to greater rainfall. Copyright © 2011 Royal Meteorological Society

1. Introduction

Various authors have used circulation indices for quantitative descriptions of atmospheric circulation. Among the first studies, Namias and Clapp (1951) introduced the zonal circulation index, defined as the difference between hemispherically averaged pressures, over sub-tropical and a sub-polar latitude circles. They studied the behaviour of the mean, the quasi-periodic variations and the relationship with the amplitude and motion of long waves.

In the south of South America, Schwerdtfeger (1951) and Prohaska (1952) introduced local and regional circulation indices, and Minetti et al. (1987) applied several circulation indices, calculating pressure differences between pairs of meteorological stations, to study long-period fluctuations (2–20 years) and relate them with precipitation.

The use of some indices, such as the Southern Oscillation Index (SOI (Tahití-Darwin)), the Trans-Polar Index (Jones et al., 1999), the Antarctic Oscillation Index (Jones et al., 2003), the Monsoon Index (Kinter et al., 2002; Lee et al., 2005, Kajikawa et al., 2010), the North Atlantic Oscillation Index (NAOI) (Nordli et al., 2005), the Mediterranean Oscillation Index (MOI) and a new Mediterranean Circulation Index (MCI) (Feidas et al., 2007) has become widespread. Niedywiedy et al. (2009), compared precipitation patterns using three simple regional indices of atmospheric circulation: westerly circulation, southerly circulation and cyclonicity (C) index as well as the relationship between precipitation and NAOI. Indices are used for objective descriptions of different atmospheric situations. These are some examples of the use of indices in meteorological research.

Alessandro (2003a, 2003b, 2005a, 2005b) studied blocking phenomena in southernmost South America in the period 1988–1998, using a zonal index I, defined as I = U (30°S) + U (60°S) − 2U (45°S), where U is the zonal wind component at 500 hPa at given latitudes and longitudes. The author also examined the effects of blocking on temperature and precipitation in Argentina.

Index I was applied later over the same region (Alessandro, 2008) to analyse the influence of strong westerly flow at around 45°S on those two variables. When this index was calculated at 70°W, temperature anomalies were positive over almost all Argentina and higher than normal precipitation was confined to the south of 42°S. Recently, Kreienkamp et al. (2010) used a circulation index capable of tracing blocked states of the hemispheric circulation which is relevant for a mid-latitude window between 40 and 60°N and between 30°W and 30°E.

Using data for the period 1980–1988 provided by the European Centre for Medium-Range Weather Forecasts, Alessandro (1998a, 1998b, 2001) applied regional circulation indices to examine long waves over South America. The author studied the development of migratory pressure systems in relation with persistent values of zonal (Z) and meridional (R) circulation indices at the 500 and 1000 hPa levels. As a result, a strong relationship was observed between anomalies of circulation indices and synoptic development in Argentina. High values of the zonal index are related to strong westerly winds over Patagonia, with the rapid passage of small-amplitude troughs and ridges. On the other hand, low values represent opposite situations, with blocking produced by anticyclones located far to the south and cut-off lows at low latitudes, as well as cyclogenesis in the extreme northeast of the country. For that period, seasonal values of Z were positive during the four seasons of the year at both levels studied, according to the definition of that variable. A peak was observed in summer at 500 hPa. During this season, geopotential height at 35°S grows markedly due to the southward displacement of the polar vortex. At 1000 hPa, maximum values were observed in winter, when anticyclonic circulation prevailed at 35°S, in contrast to the slight variations in pressure observed throughout the year at 55°S, as can be seen in the Climate Atlas for Argentina (1960).

Alessandro (1998a, 1998b) also related the extreme deviations of Z anomalies with precipitation and surface temperature through zonal index Z at nine sites in Argentina, which represent different climatic regions. As a result, long westerly waves were observed to have significant influence on precipitation. In addition, there is significant correlation with temperature, which is positive at 500 hPa and negative at 1000 hPa.

For a given longitude, difference Z (DZ) calculated between 500 and 1000 hPa and two latitudes gives an idea of the baroclinicity. Alessandro (1998a, 1998b) calculated DZ in two latitude ranges, 35–55°S and 20–40°S. The two DZ values were compared and the percentage of total baroclinicity at lower latitudes was calculated with respect to higher latitudes. Alessandro found that in warm months, the contribution of baroclinicity at low latitudes was smaller. This is because in summer the polar front does not reach low latitudes in general. On the other hand, the percentages of baroclinicity obtained in winter for both latitude ranges were similar (about 50%). This means that the influence of the polar front is similar in both latitude ranges.

Meridional index R at 1000 hPa showed the best correlation with temperature, since meridional circulation at this level has the strongest effect on temperature (Alessandro, 1998a). With southerly (northerly) flow, temperature anomalies at the selected sites are negative (positive). Under mean conditions, R > 0 corresponds to increased pressure over the Pacific and decreased pressure over the Atlantic, which can be given by the strengthening of the Pacific anticyclone or greater cyclonic activity at 40°S over the Atlantic, with consequent advection of cold air over Argentina. At both levels, the meridional circulation index behaves similarly with respect to temperature. On the contrary, if strengthening occurs of the Atlantic anticyclone (R < 0) or cyclonic activity enhances over the Pacific, the sign of the index reverts as well as the sign of temperature anomalies, a consequence of the advection of warm air caused by the western edge of the Atlantic anticyclone. When absolute values of anomalies of RR) are greater than or smaller than the standard deviation, there is good correlation between the indices and rainfall. If ΔR < 0 rainfall is greater than normal: for ΔR > 0 rainfall is smaller.

Based on the relationship found between Z, R and both meteorological variables in Argentina during 1988–1998, this paper analyses the time series of the indices over a much longer period. This study will make it possible to examine fluctuations and trends of both indices related to changes in atmospheric circulation in the south of the southern hemisphere, as well as the incidence of circulation and the indices on temperature and rainfall during that period.

2. Data and methodology

Daily surface temperatures and daily rainfall data from 43 weather stations were provided by the Argentinean National Meteorological Service (SMN). Daily 500 and 1000 hPa geopotential heights at 1200 Z as well as graphical outputs of geopotential heights and 1000/500 hPa thickness were obtained from the reanalyses of the National Center of Environmental Protection (NCEP) ( over a 2.5° × 2.5° grid. All data cover the period 1960–2006.

The indices used, R and Z, are defined as follows. The difference in geopotential height between two longitudes, one to the west and the other to the east of the continent, at a given latitude: R = H(40°S, 90°W) − H(40°S, 50°W), where H is the height in geopotential meters of the selected isobaric surfaces (1000 and 500 hPa). Latitude 40°S was selected because at 1000 hPa, the southern extreme of subtropical anticyclones is located at 40°S and at 500 hPa this latitude is immersed in the westerlies and close to its peak at this level. The difference in height between two latitudes, one subtropical and one subpolar, at 70°W: Z = H(35°S, 70°W) − H(55°S, 70°W). Those latitudes were selected because at 1000 hPa, 35°S corresponds to the subtropical high-pressure strip (Minetti and Vargas, 1983) and 55°S, to the area of greatest interdiurnal pressure variability (Lichtenstein, 1976; Trenberth, 1991).

At 500 hPa, 35°S represents the northern boundary of the westerly flow in summer, though it is well immersed in the westerlies in winter (Fernandez and Necco, 1982). The greatest kinetic energy of perturbations concentrates at 55°S.

Extreme values of both indices are obtained by comparing the mean value of the index for the selected period (month, season, year) with the difference between the mean (month, season, year) of the entire index series and its standard deviation σ.

These values are classified in two groups: ‘negative’ and ‘positive’ of monthly and seasonal data. For the seasonal data, the group ‘negative’ is composed of those years whose seasonal R(Z) ((Rse (Zse)) are smaller than Rmse − σse(Zmse − σse) and ‘positive’ composed of the R(Z) that are greater than Rmse + σse(Zmse + σse), where Rmse (Zmse) and σse correspond to the mean seasonal and the seasonal standard deviation of R (Z) in the period 1960–2006.

For example, mean values of R of the 46 winters (June, July and August) are used to calculate the mean (Rmwi) and the standard deviation (σwi). The difference and the sum of those two values are calculated, and are then compared with each winter in the series. The comparison results in two series, one composed of the winters whose values are smaller than Rmwi − σwi and the other one of the winters whose values are greater than Rmwi + σwi. Once the selection is made, calculations are made for each group of the seasonal temperatures, accumulated rainfall and the frequency of rainy days for each meteorological station. The same procedure is applied for each month and year.

With this analysis, the number of seasonal cases of Z and R is low (between 7 and 10), despite a 46 year series used. To have a greater number of cases, a much longer (not available) time series would be necessary.

Temperature is analysed by means of its anomalies (ΔT), which are calculated for each group as the difference between the mean (monthly, seasonal or annual) temperature under extreme conditions of R and/or Z and the mean (monthly, seasonal or annual) temperature of the entire 1960–2006 series.

Precipitation is estimated through the quotient (rac) between the mean accumulated (monthly, seasonal or annual) precipitation under extreme conditions and the mean rainfall for the entire series. Values rac > 1 at a given location indicate that the rainfall for the selected case (+ or −) is greater than normal.

Maps are plotted of the mean geopotential height and its anomalies at 1000 and 500 hPa for both groups and indices in order to relate the calculated indices with the meteorological situations and understand the precipitations and temperatures they generate.

In order to relate the 1000 and 500 hPa levels and analyse variations in baroclinicity, the difference of Z (DZ) between the 500 and 1000 hPa levels is calculated as well as its anomalies.

DZ is defined at λ = 70°W as:

equation image

To examine the variation over time of the indices and their differences, linear trends are calculated for index values in the 46 year period under study. Magnitudes are analysed using the Mann–Kendall test (Brooks et al., 1953) and the significance of the slope of the regression line at the 95% level. Significant trend equations are bolded. Temperature and rainfall that take place under extreme conditions of R and Z are associated with the signs of the trends of both circulation indices.

It is worth mentioning that due to space constraints, only the seasonal fields of geopotential heights, temperature and rainfall will be shown and analysed in detail, although monthly and annual fields were also calculated.

Finally, to assess the contribution of the variation of these indices on rainfall, temperature and baroclinicity in the period 1960–2006, mean values of the latter variables are compared in two different periods. The periods correspond to the first (1960–1974) and the last 15 years (1992–2006) of the series. Temperature and accumulated rainfall fields are estimated of the differences between both periods. The periods were selected arbitrarily, without considering any changes or inflection points in the series.

3. Results and discussion

3.1. Meridional index R

3.1.1. Mean fields

Figures 1 and 2 show the seasonal fields of geopotential height anomalies at 1000 hPa for the negative (−) and positive (+) groups (R1000− and R1000+) and the 500 hPa fields corresponding to the groups selected for 1000 hPa. For R1000− a negative anomaly can be observed dominating the Pacific, and a positive anomaly over the Atlantic, during the four seasons of the year. This indicates the weakening of the Pacific anticyclone, which can be due to the passage of a greater number of cyclones or a smaller number of migratory anticyclones, which on average move across the country from the Pacific at around 41°S (Alessandro and Lichtenstein, 1995). Another possible explanation would be enhanced cyclonic activity over the Atlantic at around 40°S.

Figure 1.

Seasonal anomalies of geopotential height fields at 1000 hPa for R1000− (a,c,e,g) and corresponding fields at 500 hPa (b,d,f,h), in summer (a,b), autumn (c,d), winter (e,f) and spring (g,h)

Figure 2.

Seasonal anomalies of geopotential height fields at 1000 hPa for R1000+ (a,c,e,g) and corresponding fields at 500 hPa (b,d,f,h), in summer (a,b), autumn (c,d), winter (e,f) and spring (g,h)

The negative anomaly from the Pacific penetrates farther into the continent in winter and spring when it peaks at − 22.5 gpm. Consequently, the positive anomaly over the Atlantic is farther away from the continent during these two seasons. The positive anomaly implies the strengthening of the high-pressure systems reaching the Atlantic and, in this case, the peak is located farther to the south of the quasi-stationary anticyclone. Such layout gives rise to a flow from the north-northeast, which brings moist and warm air to the east of the country.

The geopotential height fields at 1000 hPa for winter and spring are similar, as are the fields for summer and autumn.

On the other hand, anomaly signs are opposite for R1000+ where the Pacific anticyclone becomes stronger or the Atlantic anticyclone weakens. In winter, the greatest area of the continent is affected by the negative anomaly centred over the Atlantic. In autumn, the greatest area is affected by the positive anomaly. 1000 hPa fields show greater differences in the location of both anomalies than in R1000−. The corresponding mean fields at 500 hPa are mostly in phase with those at 1000 hPa.

Seasonal fields of geopotential height at 500 hPa are shown in Figure 3. Those fields were obtained for the years resulting from the application of index R at 500 hPa. Although the 500 hPa fields calculated for both groups are similar to those at 1000 hPa, there are slight differences in the location of anomalous centres, which in general are displaced slightly to the east.

Figure 3.

Seasonal anomalies of geopotential height fields at 500 hPa for R500− (a,c,e,g) and R500+ (b,d,f,h) in summer (a,b), autumn (c,d), winter (e,f) and spring (g,h)

3.1.2. Temperature and rainfall under extreme conditions Temperature

although few temperature anomalies are greater than 1, the seasonal fields formed by the years of negative group calculated at 1000 hPa (R1000−) (Figure 4) present positive temperature anomalies over almost the entire country, according to the dominating northerly flow (Figure 1). The exception occurs in summer, with a much weaker northeasterly flow (Lichtenstein, 1980) which prevents further warming in the centre and north of the country.

Figure 4.

Seasonal anomalies of temperature fields at 1000 hPa for R1000− (a–d) and R1000+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h)

Similarly to the annual field, the spring field shows the penetration of the negative anomaly into the south of Patagonia, which may be caused by the passage of migratory cyclones.

Anomaly fields of geopotential heights (Figure 2) of R1000+ show a dominating flow from the south over a great part of the country, which responds to the presence of the Atlantic and Pacific anomalies, and causes ΔT < 0 during practically all seasons of the year (Figure 4). The strongest negative anomaly over the Atlantic appears in winter covering the entire country, with the greatest ΔT < 0. The signs of ΔT calculated through R at 500 hPa (Figure 5) are similar to those for 1000 hPa in both groups. However, summer values for R500− are negative over the entire country.

Figure 5.

Seasonal anomalies of temperature fields at 500 hPa for R500− (a–d) and R500+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h) Rainfall

the relationships were calculated for accumulated rainfall (Figures 6 and 7) as described in Section '2. Data and methodology'. Values of r estimated at 1000 and 500 hPa for negative group, show a dominance of values rac > 1 in the four seasons, although in winter they stretch over a larger area, except for some regions in the west of the country at around 40°S. In winter, the mean fields of geopotential height anomalies at 1000 hPa (Figure 1) show a negative anomaly entering from the Pacific to north-central and eastern Argentina. This implies the passage of low pressure systems or intense weakening of the Pacific anticyclone and stronger cyclonic circulation in the east, with penetration of northeasterly flow. At 500 hPa, these anomalies are displaced to the west and have similar characteristics.

Figure 6.

Cumulative seasonal rainfall relationship fields at 1000 hPa for R1000− (a–d) and R1000+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h)

Figure 7.

Cumulative seasonal rainfall relationship fields at 500 hPa for R500− (a–d) and R500+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h)

The greatest area covered by rac > 1 values is observed in winter followed by summer. The most remarkable difference between both seasons is the intensification of the northwesterly low pressure system (DNOA). In winter, it presents a negative anomaly (−5 gpm) and in summer the anomaly is positive (5 gpm), which indicates that northerly or northeasterly flow is weaker and therefore rainfall is less in the north or northeast of the country. The other difference is that accumulated values reveal rainfall shortages in the south of the country during summer, where a positive anomaly of geopotential height at 1000 hPa is observed.

Whilst the geopotential height fields at 1000 hPa for summer and autumn are similar, as are the fields for winter and spring, at 500 hPa, to the south of 40°S, there are differences between the two latter fields in the extent of penetration of the Pacific anomaly into the country, as well as of the positive anomaly to the north of that latitude. Because of this, rainfall in spring is weak with values of rac smaller than one to the north of 35°S. Rainfall is greater in autumn than in summer because the northward flow at 1000 hPa is normal (the anomaly for that season of the year is zero): however, rainfall is low in the south of Patagonia. Values greater than 1 in both relations occur over practically the entire country in summer, autumn and winter.

For R1000+ and R500+, coefficients rac are smaller than one over almost the whole country, mainly because of the weakening of the northeasterly flow and the dominance of southerly flow at 500 and 1000 hPa. Agreement between values of rac relationships smaller than one is better than for the other group, except for the centre of the country at 1000 hPa in winter. Although weak, the annual field presents values greater than 1 to the south of 34°S. The rest of the country remains with practically normal values.

3.2. Index Z

3.2.1. Mean fields

Figure 8 presents the seasonal fields of geopotential heights of Z1000− and Z500− A strong positive anomaly is observed in the south of the country or at around 70°W over the Atlantic and the Pacific Oceans. Positive anomalies at 500 hPa are in phase with those at 1000 hPa, except over the continent to the north of 40°S in summer, where decreased geopotential heights are observed at 500 hPa. The anomaly has opposite sign in the fields of Z+ (Figure 9). Negative anomalies at 1000 hPa stretch from the south to great part of the country. A positive anomaly located north of 40°S can be seen in the 500 hPa fields. This anomaly is not observed at lower levels. The mentioned characteristics also have a place in the monthly fields of each season.

Figure 8.

Seasonal anomalies of geopotential height fields at 1000 hPa for Z1000− (a,c,e,g) and corresponding fields at 500 hPa (b,d,f,h), in summer (a,b), autumn (c,d), winter (e,f) and spring (g,h)

Figure 9.

Seasonal anomalies of geopotential height fields at 1000 hPa for Z1000+ (a,c,e,g) and corresponding fields at 500 hPa (b,d,f,h), in summer (a,b), autumn (c,d), winter (e,f) and spring (g,h)

3.2.2. Temperature and rainfall under extreme conditions Temperature

calculated monthly ΔT (not shown) indicate the dominance of negative values in all the cases of Z and positive values in the opposite cases. Sign comparison of temperature anomalies obtained from calculating Z1000 and Z500 shows that the greatest differences are in the geographic boundaries of the country. For example, in January temperature anomaly fields for Z− are practically equal at both levels. For Z500+, positive values stretch farther to the north over Patagonia. Of the 12 months of the year, the largest differences between both levels in Z− are observed in April. Seasonal temperature anomalies are shown in Figure 10.

Figure 10.

Seasonal anomalies of temperature fields at 1000 hPa for Z1000− (a–d) and Z1000+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h)

As already mentioned, the monthly fields of geopotential height anomalies are similar within each season of the year. However, differences between the temperature anomalies they generate require an in depth analysis that goes beyond the scope of this paper. In general, there is a weakening of the quasi-stationary anticyclones of the Pacific and Atlantic oceans in Z− (Figure 8). The weakening of the latter contributes to keep temperatures lower in the north and east of the country. Negative geopotential height anomalies at both levels, generally at latitudes located farther to the north of Patagonia, suggest the passage of synoptic systems that causes temperature to drop.

Westerly winds over Patagonia are weaker. According to the positive geopotential height anomalies, easterly winds cause temperature to drop. Those winds tend to neutralize the effects of forced subsidence caused by the westerlies, particularly in winter and spring, when mean fields of geopotential height at 1000 hPa show that winds have a southerly component.

On the other hand, in Z+ (Figure 9), flow from the west is stronger than that from the east over Patagonia, with consequent temperature warming due to subsidence (Alessandro, 2007). However, some mean fields of geopotential heights indicate the presence of winds with a southerly component, which cause temperatures to remain low in the southernmost extreme of the country.

The situation is almost normal at 1000 hPa over northern Argentina (Figure 9), while at 500 hPa, to the north of Patagonia, positive values of geopotential height anomalies dominate, which is associated with temperature rise over the region. Rainfall

rainfall (Figure 11) was subjected to the same procedure as temperature. For Z−, a general increase is observed in rainfall in the four seasons. It is worth clarifying that values for southern Argentina are not completely reliable due to data gaps in the study period. Anomalies at 1000 hPa show the transport of moisture from the Atlantic to Patagonia. However, in the north, the weakening of the Atlantic anticyclone, which is also observed at 500 hPa, causes low rainfall in the north and northeast of the country during some months.

Figure 11.

Cumulative seasonal rainfall relationship fields at 1000 hPa for Z1000− (a–d) and Z1000+ (e–h) in summer (a,e), autumn (b,f), winter (c,g) and spring (d,h)

Opposite signs are observed for Z+, the most frequent being rac < 1. Again, it is in the north where values of r with opposite signs appear, e.g. in summer and winter. For this group, westerly flow dominates at 1000 hPa, and causes air to dry due to subsidence. The north of the country, under almost normal conditions at 1000 hPa receives moisture from the humid tropics and, therefore, more precipitation. Rainfall, however, is greatly inhibited by the positive anomaly at 500 hPa (Figure 9).

Monthly observations reveal that these features tend to remain, with varying areas of influence. Given the randomness of rainfall, the areas are less homogeneous than for temperature.

3.3. Trends of R and Z

Table 1 shows the linear trend equations for the four seasons of the year at 1000 and 500 hPa, as well as the 1000–500 hPa difference of Z, in the period 1960–2006. Significant values, according to Mann–Kendall's test and the slope test, are in bold.

Table 1. Seasonal linear trends of R, Z and DZ
 1000 hPa500 hPa1000 hPa500 hPaDZ
  1. x = 1 to 46 years.

Summer− 0.26x + 6.280.31x − 8.050.79x− 17.750.28x − 6.40− 0.50x + 11.36
Autumn− 0.28x + 6.72− 0.30x + 7.120.57x− 12.900.07x − 1.66− 0.50 + 11.22
Winter− 0.27x + 6.54− 0.26x + 6.290.42x − 9.37− 0.05x + 1.00− 0.47x + 10.47
Spring− 0.33x + 8.00− 0.13x + 3.100.50x− 8.400.85x− 19.15− 0.02x + 0.50
Annual− 0.29x + 6.70− 0.11x + 2.600.56x− 13.480.20x − 6.38− 0.46x + 10.43

Linear trends of monthly R at 1000 and 500 hPa for the 46 years are mostly negative, which mirrors in the sign of annual and seasonal trends, the exception being July.

At both levels the values of R decrease over time, except for summer at 500 hPa. Trends are negative for monthly values in most months.

The trends of Z at 1000 hPa are positive for all months, while at 500 hPa they are negative in February, April, May, August and September. Seasonal trends are shown in Table 1, with positive values at 500 and 1000 hPa, except for winter at 500 hPa when the trend is small and negative.

In general, when trends of Z are positive at 500 and 1000 hPa, the trend of DZ is positive when the trend at 1000 hPa is smaller than at 500 hPa, as happens in October, November, March, summer, autumn, spring and in the annual series. When the trends of Z are positive at 500 and 1000 hPa, the trend of DZ is negative whenever the trend at 1000 hPa is greater than at 500 hPa, as happens in January, June and July. If trends at both levels are similar, there is practically no variation in DZ (December).

If the trend of Z at 500 hPa decreases and the trend at 1000 hPa increases, DZ decreases (February, April, May, August, September, winter). To illustrate these results, Figure 12(a) and (b) show the example of the latter case and the first combination mentioned in the previous paragraph. ‘Es’ represents the thickness between 500 and 1000 hPa at 35°S and ‘Es’, indicates the thickness resulting from decreased or increased trends of Z at 500 or 1000 hPa, assuming that Es at 55°S remains unchanged. In (a) Es′< Es, which means that DZ decreases. In (b) Es′> Es, and the difference grows.

Figure 12.

Scheme of DZ change, (a) if the trend of Z at 500 hPa decreases and increases at 1000 hPa, (b) if the trend increases at both levels, but more at 500 than at 1000 hPa

In general, the trend of DZ is significant when the trend is significant at any of both levels (e.g. October), but it is not significant when both levels are significant (e.g. March). Negative trends of DZ imply a decrease in baroclinicity.

Alessandro (2007) showed that decreasing DZ over time in the southernmost area of the region under study is mainly due to warming of the air layer expanding from about 850 to 500 hPa.

3.4. Relationship between trends of indices and temperature and rainfall

In the period 1960–2006, R− leads to warming temperatures and increased seasonal rainfall. Positive extremes lead to a drop in both variables over most of the country. A slight increase in both variables is assumed considering these results and that for this period the trend of R at 1000 hPa is negative and significant during the four seasons of the year as well as at 500 hPa except for summer.

Z− cause ΔT < 0 over almost the entire country, as well as dominating rac > 1 over a smaller area. Opposite characteristics are observed in the months when Z+ is positive. According to these results and considering the positive trends of Z at 1000 and 500 hPa, which are significant in general at the latter level, Z is assigned an increase in temperature and a drop in rainfall over most parts of Argentina.

3.5. Mean fields of two sub-periods of the analysed period

Figure 13 shows the temperature and rainfall anomalies obtained as the difference between mean values of a recent and an old period. For all seasons, the south of the country has positive thickness anomalies but only summer anomalies of the 500/1000 thickness for both periods are shown in Figure 14. Mean ΔT was positive for almost all the country (Figure 13(a)), except for part of the centre and a small area in Patagonia. These differences are due to those that occur in winter and spring, which have negative values over that region.

Figure 13.

(a) Mean temperature differences between the period 1992/2006 and 1960/1974, (b) same as (a) but for rainfall

Figure 14.

Summer thickness 500/1000 in the period (a) 1960/1974 and (b) 1992/2006

The analysis of the signs of the trends of both indices and of the ΔT obtained for the extreme cases of those indices is in agreement with the temperature rise in the most recent years of the period under study.

Rainfall also presents mostly positive values (Figure 13(b)), the largest in the centre and northeast of the country, followed by Patagonia, but the effects of R and Z on precipitation are different. The mean fields in Figure 1 show that the positive anomaly of geopotential height at 1000 hPa covers a large part of the Atlantic, which brings moisture even to Patagonia. However, these values decrease because of the positive trend of Z, which implies a stronger westerly flow (Alessandro, 2007) mainly over Patagonia. This is consistent with the negative anomaly at both levels in the south of the country during all the seasons of the year (Figure 9(b)) and a positive anomaly at 500 hPa over northern Patagonia. Therefore, R would cause rainfall to rise and Z would make it drop, although according to the significance of the trends, the influence of R is greater than the influence of Z. Therefore, Z coincides with the mean rainfall difference field.

The decrease over time of DZ anomalies implies a decrease in baroclinicity between 35 and 55°S in Argentina. The growth of thickness during summer in recent years can be observed in Figure 14, when anomalies changed sign from negative to positive.

4. Conclusions

The seasons with extreme negative of R (R−) (positive (R+)) values at 1000 hPa and/or 500 hPa present for the period 1960–2006 the following observations.

  1. A negative (positive) geopotential height anomaly dominating the Pacific Ocean and a positive (negative) one over the Atlantic. The geopotential height anomaly fields at 1000 hPa in winter and spring are similar, as are those for summer and autumn. Flow is from the north-northeast and brings warm and moist air into the east and northeast of the country.

  2. Temperature anomalies are positive (negative) over almost the entire country except for the summer in the cases when Rmse < Rse − σ. Most anomalies are smaller than one (greater than − 1).

  3. Predominant increase (decrease) is observed in rainfall, i.e. ra > 1(ra < 1).

The seasons with extreme negative of Z (Z−) [positive (Z+)] values at 1000 hPa and/or 500 hPa present the following observations

  1. A positive (negative) anomaly at around 55°S and 70°W and a slight decrease (increase) in the strength of subtropical anticyclones.

  2. Temperature anomalies are negative (positive) over almost the entire country.

  3. Increased (decreased) rainfall dominates over a smaller area than temperature.

According to the Mann–Kendall and the slope tests at the 95% confidence level, the seasonal trends of R at 1000 hPa are negative and significant. At 500 hPa, they are negative, except for summer when it is not significant.

Seasonal trends of Z at 1000 hPa are positive and significant, except for winter. For this season, the trend at 500 hPa is negative and not significant. The remaining trends are positive at this level. The annual and seasonal trends of difference of Z (DZ) between 500 and 1000 hPa are negative, and significant except for spring.

In the period under study, the analysis of DZ reveals a drop in baroclinicity, which implies annual and seasonal variations in atmospheric circulation over southern South America.

The signs of the trends of both circulation indices and the values of temperature anomalies obtained for their the extreme cases, might contribute to the warming observed in the last years of the analysed period 1961–2006. Under the same assumptions, R would cause rainfall to rise and Z would make it drop, although according to the significance of the trends, the influence of R is greater than the influence of Z.


This study was supported by grant PICT N° 38277 of the AGENCIA and grant 228 of the UBA.