• Cavalcanti, I. F., N. J. Ferreira, M. G. A. J. da Silva, and M. A. F. da Silva Dias, 2009: Tempo e Clima no Brasil. Oficina de Textos, 463 pp.

  • Crossett, C. C., and N. D. Metz, 2017: A climatological study of extreme cold surges along the African highlands. J. Appl. Meteor. Climatol., 56, 17311738, https://doi.org/10.1175/JAMC-D-15-0191.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fortune, M., and V. Kousky, 1983: Two severe freezes in Brazil: Precursors and synoptic evolution. Mon. Wea. Rev., 111, 181196, https://doi.org/10.1175/1520-0493(1983)111<0181:TSFIBP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gan, M., and V. Rao, 1994: The influence of the Andes Cordillera on transient disturbances. Mon. Wea. Rev., 122, 11411157, https://doi.org/10.1175/1520-0493(1994)122<1141:TIOTAC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., 2000: Cold air incursions over subtropical South America: Mean structure and dynamics. Mon. Wea. Rev., 128, 25442559, https://doi.org/10.1175/1520-0493(2000)128<2544:CAIOSS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., and J. M. Wallace, 1998: Summertime incursions of midlatitude air into subtropical and tropical South America. Mon. Wea. Rev., 126, 27132733, https://doi.org/10.1175/1520-0493(1998)126<2713:SIOMAI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Girardi, C., 2006: O poço dos Andes. Bull. Braz. Soc. Meteor., 30, 39–48.

  • Hackbart, E., 2010: Argentina se cobre de branco e sul do Brasil registra neve e raro episódio de chuva congelante. Anima Meteorologia, http://anitameteorologia.blogspot.com/2010/07/argentina-se-cobre-de-branco-e-gelo-e.html.

  • Hamilton, M. G., and J. R. Tarifa, 1978: Synoptic aspects of a polar outbreak leading to frost in tropical Brazil, July 1972. Mon. Wea. Rev., 106, 15451556, https://doi.org/10.1175/1520-0493(1978)106<1545:SAOAPO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ingemi, D., 2012: La prima ondata di freddo polare in Sud-America porta la neve su alcune città argentine, fiocchi pure sui quartieri di Buenos Aires e venti di tempesta sulla Terra del Fuoco. MeteoWeb, http://www.meteoweb.eu/2012/06/la-prima-ondata-di-freddo-polare-in-sud-america-porta-la-neve-su-alcune-citta-argentine-fiocchi-pure-sui-quartieri-di-buenos-aires-e-venti-di-tempesta-sulla-terra-del-fuoco/138462/.

  • Krishnamurti, T., M. Tewari, D. Chakraborty, J. Marengo, P. Silva Dias, and P. Satyamurty, 1999: Downstream amplification: A possible precursor to major freeze events over southeastern Brazil. Wea. Forecasting, 14, 242270, https://doi.org/10.1175/1520-0434(1999)014<0242:DAAPPT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liebmann, B., G. N. Kiladis, L. V. Carvalho, C. Jones, C. S. Vera, I. Bladé, and D. Allured, 2009: Origin of convectively coupled Kelvin waves over South America. J. Climate, 22, 300315, https://doi.org/10.1175/2008JCLI2340.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lupo, A., J. Nocera, L. Bosart, E. Hoffman, and D. Knight, 2001: South American cold surges: Types, composites, and case studies. Mon. Wea. Rev., 129, 10211041, https://doi.org/10.1175/1520-0493(2001)129<1021:SACSTC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marengo, J., 2001: Infoclima – July’13, 2001. INPE/CPTEC, http://infoclima1.cptec.inpe.br/~rinfo/infoclima/jul_2001.shtml.

  • Marengo, J., A. Cornejo, P. Satyamurty, C. Nobre, and W. Sea, 1997: Cold surges in tropical and extratropical South America: The strong event in June 1994. Mon. Wea. Rev., 125, 27592786, https://doi.org/10.1175/1520-0493(1997)125<2759:CSITAE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Metz, N. D., H. M. Archambault, A. F. Srock, T. J. Galarneau, and L. F. Bosart, 2013: A comparison of South American and African preferential pathways for extreme cold events. Mon. Wea. Rev., 141, 20662086, https://doi.org/10.1175/MWR-D-12-00202.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parmenter, F., 1976: A Southern Hemisphere cold front passage at the equator. Bull. Amer. Meteor. Soc., 57, 14351440, https://doi.org/10.1175/1520-0477(1976)057<1435:ASHCFP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pezza, A. B., and T. Ambrizzi, 1999: Um Estudo das Flutuações de Temperatura para o Período de Inverno da América do Sul, Correlacionando a Patagônia com o Sul do Brasil. Rev. Bras. Meteor., 14, 2334.

    • Search Google Scholar
    • Export Citation
  • Pezza, A. B., and T. Ambrizzi, 2005: Cold waves in South America and freezing temperatures in São Paulo: Historical background (1888-2003) and case studies of cyclones and anticyclones tracks. Rev. Bras. Meteor., 20, 141158.

    • Search Google Scholar
    • Export Citation
  • Revista Cafeicultura, 2010: Especial: 35 Anos de Geada de 1975—Entenda o que foi a geada negra que dizimou todas as plantações de café do Paraná. Cafeicultura, accessed 21 December 2016, http://www.revistacafeicultura.com.br/index.php?mat=34022.

  • Russo, S., and et al. , 2014: Magnitude of extreme heat waves in present climate and their projection in a warning world. J. Geophys. Res. Atmos., 119, 12 50012 512, https://doi.org/10.1002/2014JD022098.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seluchi, M., Y. Serafini, and H. Le Treut, 1998: The impact of the Andes on transient atmospheric systems: A comparison between observations and GCM results. Mon. Wea. Rev., 126, 895912, https://doi.org/10.1175/1520-0493(1998)126<0895:TIOTAO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sturcke, J., 2007: Buenos Aires has first snow 1918. The Guardian, accessed 29 April 2017, https://www.theguardian.com/world/2007/jul/10/argentina.weather.

  • Uccellini, L. W., and D. R. Johnson, 1979: The coupling of upper and lower tropospheric jet streaks and implications in development of severe convective storms. Mon. Wea. Rev., 107, 682703, https://doi.org/10.1175/1520-0493(1979)107<0682:TCOUAL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weatherboy, 2007: Inverno 2000 – O inverno mais rigoroso dos últimos 50 anos. Accessed 21 December 2016, https://www.youtube.com/watch?v5GRgS21VILlY.

  • Xu, Q., 1990: A theoretical study of cold air damming. J. Atmos. Sci., 47, 29692985, https://doi.org/10.1175/1520-0469(1990)047<2969:ATSOCA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Geographic locations of the selected regions used in this study, denoted by MP1, MP2, MR, and EP. The area delimited by all of the four regions is denoted as ALL.

  • View in gallery

    Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 17 Jul 2010. The primary cold front is denoted by a black line and the secondary front by a gray line. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

  • View in gallery

    Meteorological fields for 1200 UTC 17 Jul 2010: (a) air temperature at 2 m (shaded; °C), sea level pressure (contours; hPa), and 850-hPa winds (vectors; m s−1); (b) 1000–500-hPa thickness (contours; dam), thermal wind (vectors; m s−1), and air temperature anomaly at 2 m (shaded; °C); (c) 850-hPa winds (vectors; m s−1), thermal advection (shaded; × 10−4 °C s−1), rates of change of geopotential height [black contours; dam (12 h)−1], and omega (red contours; Pa s−1); (d) 500-hPa relative vorticity advection (shaded; × 10 −9 m s−2) and geopotential height (solid contours; dam) and 700-hPa omega (dashed contours; Pa s−1); (e) 925-hPa thermal advection (shaded; × 10−4 °C s−1, color bar), 500-hPa relative vorticity (contours; 10−9 s−1), and 250-hPa streamlines with wind speed (shaded; scale at top-right corner; m s−1); and (f) anomaly of sea level pressure (shaded; hPa), anomaly of 500-hPa geopotential height (contours; dam), and anomaly of 250-hPa wind (vectors; m s−1). The highlighted box indicates its associated pattern.

  • View in gallery

    As in Figs. 3a–f, but for the MP pattern composite.

  • View in gallery

    Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 12 Jul 2007. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

  • View in gallery

    As in Fig. 3, but for 1200 UTC 12 Jul 2007.

  • View in gallery

    As in Figs. 3a–f, but for the MR pattern composite.

  • View in gallery

    Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 17 Jul 2000. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

  • View in gallery

    As in Fig. 3, but for 1200 UTC 17 Jul 2000.

  • View in gallery

    As in Figs. 3a–f, but for the EP pattern composite.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 173 173 33
PDF Downloads 68 68 6

Classification of Extreme Cold Incursions over South America

View More View Less
  • 1 Institute of Astronomy, Geophysics and Atmospheric Sciences, University of São Paulo, São Paulo, Brazil
© Get Permissions
Full access

Abstract

Cold-air incursions across South America present a variety of problems, sometimes by advancing to the Amazon basin and causing agricultural losses related to extreme low-temperature events. The synoptic conditions that produce cold-air incursions are relatively well understood; however, the most affected region depends on the route by which cold air spreads across the continent. Therefore, the classification of extreme cold-air incursions allows a better understanding of the particularities directly related to the aforementioned losses. In this work, similarities and differences among extreme cold surges were found through time series correlation of anomaly temperatures in four selected areas and compositing techniques from ERA-Interim reanalysis datasets, resulting in three distinct patterns: meridional penetration (MP), meridional restriction (MR), and east penetration (EP). The patterns identified here enable a more detailed understanding of the synoptic patterns and forcing mechanisms associated with extreme cold-air incursions and therefore can be used for operational weather forecasting.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Isaque Saes Lanfredi, isaque.lanfredi@usp.br

Abstract

Cold-air incursions across South America present a variety of problems, sometimes by advancing to the Amazon basin and causing agricultural losses related to extreme low-temperature events. The synoptic conditions that produce cold-air incursions are relatively well understood; however, the most affected region depends on the route by which cold air spreads across the continent. Therefore, the classification of extreme cold-air incursions allows a better understanding of the particularities directly related to the aforementioned losses. In this work, similarities and differences among extreme cold surges were found through time series correlation of anomaly temperatures in four selected areas and compositing techniques from ERA-Interim reanalysis datasets, resulting in three distinct patterns: meridional penetration (MP), meridional restriction (MR), and east penetration (EP). The patterns identified here enable a more detailed understanding of the synoptic patterns and forcing mechanisms associated with extreme cold-air incursions and therefore can be used for operational weather forecasting.

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Isaque Saes Lanfredi, isaque.lanfredi@usp.br

1. Introduction

Cold-air incursions across South America can present particular characteristics due to the presence of the Andes Mountains, which contribute to the advancement of cold fronts to the southwest of the Amazon basin. Besides organizing the convection in northern Brazil (Garreaud and Wallace 1998), the continental fronts are also known to cause impacts that are especially felt by agriculture because of freeze events, with significant losses in coffee production and consequent economic losses (Revista Cafeicultura 2010).

The 1975 freeze event, for instance, was responsible for huge losses in coffee production, with a drop of almost 51% in Brazilian production (Marengo et al. 1997), amounting to about $75 million (U.S. dollars) in damages (Revista Cafeicultura 2010). This situation motivated the studies of Parmenter (1976) and Girardi (2006), who analyzed the event based on the visualization and interpretation of satellite images. Years later, multiple subsequent case studies (Hamilton and Tarifa 1978; Fortune and Kousky 1983; Marengo et al. 1997; Krishnamurti et al. 1999) sought to understand the incursions based on the synoptic and dynamic aspects of the related weather systems, as will be described in the following paragraphs. More recently, the works of Liebmann et al. (2009) and Metz et al. (2013) have stood out, the former describing the cold waves convectively coupled with Kelvin waves and the latter by better demonstrating the trajectory of their incursions, compared with other regions in the Southern Hemisphere.

The initial phase of the cold-air incursions is directly associated with the amplification of a trough–ridge system over southern Chile and Argentina, with a frontogenetic process along the eastern side of the Andes Mountains. Krishnamurti et al. (1999) investigated the reasons for this amplification, showing the contribution of constructive interference between baroclinic waves and planetary waves of greater magnitude at the initial stage. The maximum amplification occurs under the influence of the Andes mountain range through the ageostrophic cold-air advection on its east side, as stated by Marengo et al. (1997), in agreement with Gan and Rao (1994) and Seluchi et al. (1998).

According to Xu (1990), the momentum balance on the eastern side of the Andes is responsible for forcing the cold air mass to move northward through the effect known as mountain channeling. North of 18°S, the mountains show a change in orientation, and the channeling diminishes (Garreaud 2000). However, the pressure gradient force associated with the surface cold air continues to push it northward, so that it spreads from the southwest Amazon basin, eventually crossing the equator.

The cold advection and the differential advection of anticyclonic vorticity, in association with the advance of cold air through the interior of the continent toward the low latitudes, promote the development of a postfrontal continental anticyclone in the lower levels of the troposphere (Marengo et al. 1997; Lupo et al. 2001). This anticyclone moves northward through the effect of topography but also intensifies due to a feedback mechanism explored by Marengo et al. (1997): the intensification of the trough in the middle and upper troposphere by cold advection along the eastern side of the Andes promoted by the trough–ridge system in the southern part of South America. When the cold-air incursion is mature, the trough promotes the development of an extratropical cyclone in the Atlantic, whose association with the topographic effect can build an efficient mechanism for cold-air displacement northward (Lupo et al. 2001).

According to the theoretical explanations above, the synoptic and dynamic aspects behind the intense cold-air incursions throughout South America are well known. However, they are seen as a whole, with few works seeking to explore the different routes by which the cold-air incursions spread across the continent.

Among these studies, Fortune and Kousky (1983) examined the behavior of two severe cold events, from May 1979 and July 1981, observing that the frontal trough in May 1979 formed 10° to the east in relation to the July 1981 one, and this difference in positioning made the 1979 incursion happen faster than that of 1981. In the 1981 event, the formation of the trough farther west enhanced the interaction with the Andes topography, leading to the formation of a cutoff low in southern Brazil, and this was responsible for retaining the cold air mass for a long time over the continent, before moving on to areas of southeastern Brazil and the adjacent Atlantic Ocean.

In addition, Lupo et al. (2001) created a classification system for cold-air incursions in South America according to the longitudinal positioning of the cold fronts. To achieve their goals, the authors subjectively identified all of the cold fronts that affected South America between 1992 and 1996 from reanalysis data and carefully identified differences in behavior between the fronts that affected the longitudinal bands between 55°–65°W and 40°–50°W.

In brief, in the 55°–65°W band they identified type 1 and type 2 incursions. Type 1 incursions are weak and hardly persistent, with a zonal incursion in subtropical areas of Argentina and usually with a short-lived surface anticyclone. Type 2 incursions are more intense, sometimes penetrating to low latitudes and the Amazon basin. Type 3 events occur between 40° and 50°W and are the most common mode of cold-air incursions, which mostly take place in southern and southeastern Brazil. About 60% of type 2 incursions precede type 3 ones, when the cold air moves to the south and southeast regions of Brazil. However, type 2 incursions represent only 25% of all the cold-air incursions in South America.

These two latitudinal bands are also studied in the works of Cavalcanti et al. (2009) and Metz et al. (2013), who used the reanalysis data to explore, respectively, the annual climatological frequency of cold fronts and the frequency of days with temperatures below 16°C at 925 hPa. Considering the entire South American domain, the results in both studies show higher frequencies in southern locations and lower frequencies in northern locations, with preferred paths in the abovementioned cold frontal longitudinal bands.

Thus, the findings of Fortune and Kousky (1983), Lupo et al. (2001), Cavalcanti et al. (2009), and Metz et al. (2013) converge on the fact that there are preferred routes for cold-air incursions, which is evident even in extreme cases. Within this context, this study takes advantage of this fact to complement the existing knowledge, proposing a classification system for the most intense cold-air incursions. The article thus presents and compares these incursions through their associated synoptic and dynamic aspects, using reanalysis data to make composites and analyze case studies over preselected regions.

This paper is organized as follows. Section 2 presents the methodology used and the justification for the selected regions, while section 3 presents the characterization of the different patterns through the composite fields and the representative events. Section 4 summarizes the study and presents the conclusions.

2. Data and methodology

a. Choice of regions

Following a similar methodology outlined by Crossett and Metz (2017) for cold waves over the African continent, this study delimited the regions of interest in four boxes (called MR, EP, MP1, and MP2; Fig. 1), which represent the preferred locations of cold-air incursions from the literature.

Fig. 1.
Fig. 1.

Geographic locations of the selected regions used in this study, denoted by MP1, MP2, MR, and EP. The area delimited by all of the four regions is denoted as ALL.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

The MP1 and MP2 boxes (for meridional penetrations 1 and 2) represent cold-air incursions that propagate meridionally east of the Andes, according to Cavalcanti et al. (2009) and Metz et al. (2013). These incursions are also related to the type 2 cold waves of Lupo et al. (2001), but with greater flexibility for the longitudinal range of the chosen boxes’ domains, to best represent the topographically confined flow. This approach considers the 900-km average zonal extension of the Rossby radius of deformation (away from the Andes), from which the low-level blocked flow departs from geostrophic winds (Garreaud 2000).

The east penetration (EP) box represents the cold-air incursions that propagate through eastern South America, preferentially affecting the south and east regions of Brazil, according to Cavalcanti et al. (2009) and Metz et al. (2013). This region is also related to type 3 incursions, according to Lupo et al. (2001).

The meridional restriction (MR) box is related to the likely existence of intense cold-air incursions that affect mainly southern locations, with some form of meridional restriction. This agrees with the meridional reduction of the cold frontal climatological frequency (Cavalcanti et al. 2009; Metz et al. 2013) and also with the existence of transient incursions with zonal displacement (Lupo et al. 2001).

Localities farther south like Argentine Patagonia were not considered in this study because the frontal influence in this part of the continent is different from that farther north, including inverse relationships of temperature between Patagonia and southeastern Brazil during the winter cold-air incursions (Pezza and Ambrizzi 1999). This study therefore prioritized the intense cold-air incursions that affect large regions at relatively low latitudes, including the Amazon basin.

b. Data processing

The data considered for the purposes of this study are from the ERA-Interim reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF), with 1.5° of horizontal resolution at 0000 and 1200 UTC (Dee et al. 2011). Time series of daily average surface temperatures from May to September were used in an objective procedure to identify the extreme cold events, whose anomalies considered the period 1979–2016 as the climatological database. In each box (MP1, MP2, EP, and MR; Fig. 1), the daily time series were obtained by considering area averages of the corresponding grid points. We also considered the area average of the four boxes joined together as a unique area, which we called ALL.

The identification of the extreme cases of low temperatures covers the 2000–16 period, which provides a collection of recent significant events and an update of Pezza and Ambrizzi (2005), who examined the most significant cases throughout the whole of South America from a historical perspective for the 1888–2003 period. However, the July 2000 cold wave, though included in the 1888–2003 period, was also considered in this study as it was a very intense case and also exemplified one of the identified patterns found in this study.

For each time series, for the whole ALL area and the four boxes separately, the temperature values were ordered from low to high. However, in order to identify only the most significant events, the temperature values were previously normalized according to the normal distribution hypothesis around the climatological mean for each month. This methodology assumes a symmetrical distribution of temperatures for each locality around the mean value. However, it is necessary to consider that the extremes of low temperature tend to form a substantial tail toward the lowest values, as can be noticed in the results of Garreaud (2000). Thus, in the following results, the real probability of occurrence of the low normalized temperatures tends to be higher than that theoretically predicted by the normal distribution.

It is also important to emphasize that ordered values of temperature are by themselves insufficient to properly distinguish the most important cold-air incursions. The identification of each event was made by using reanalysis meteorological fields of temperature, sea surface pressure, and geopotential height to identify the meteorological systems on the dates corresponding to these values. This procedure avoided the false association of two anomalous values for the same event (and vice versa), also permitting the determination of the coldest temperature for each cold-air incursion.

After identifying the main cold events in the time series corresponding to each of the four boxes, the next step consisted of creating a composite of the meteorological fields of the reanalysis variables, to represent how the cold-air incursion affects each locality. To do this, seven cold events for the MP2 box, seven cold events for the MR box, and seven cold events for the EP box were selected for the realization of the respective composites. During this stage, day 0 was defined as the date of the minimum daily temperature in each region, and consequently days −2, −1, and +1 could be defined. Further details of the composites, including the use of the MP2 box temperatures series (instead of MP1), are described in section 3a.

c. Main cold-air incursions and pattern identification

Table 1 shows the 10 most intense cold-air incursions over South America from 2000 to 2016, ordered according to the lowest normalized temperatures of the daily series corresponding to the spatial average of all four boxes (ALL area). The information in the middle column in Table 1 came partly from an analysis of the meteorological fields’ reanalysis variables (section 2b). Complementary information (extreme temperatures, snowfall, etc.) came from the meteorological stations of the Brazilian National Institute of Meteorology [Instituto Nacional de Meteorologia (INMET)], the meteorological station of the Institute of Astronomy, Geophysics and Atmospheric Sciences of the University of São Paulo (IAG/USP), and several online sources of communication.

Table 1.

The 10 most intense cold incursions in South America according to the temperature anomalies in the ALL area. The first column gives the date of the event, with the corresponding anomaly value shown in the last column, in units of standard deviation σ for the whole ALL area, for which the value is 2.95°C. The middle column shows some highlights for each event, according to the media information collected by the authors.

Table 1.

Table 2 presents a list with the strongest cold-air incursions during the peak intensity, with the entries ordered by their respective normalized temperature values in the ALL area. Table 2 also includes information for each studied box (MP1, MP2, EP, and MR), where the ordinal values corresponding to their respective boxes are not presented, as in the ALL area. For box MP1, for instance, the ordinal values indicate the positions at which certain events are ranked among the most intense cold-air incursions within the specified box.

Table 2.

List of the 32 main cold-air intrusions ranked by temperature anomaly for the whole ALL area. The first column contains the rank; the second, the event date; the subsequent columns contain, for each region, the temperature anomaly values in units of standard deviation σ and the ranking with only the temperature anomalies of that region considered. The boldface numbers in columns 4, 6, and 7 indicate the events selected for the composites, whose associated patterns are shown in the last column. The boldface and italic numbers represent the most intense events for the corresponding regions analyzed separately (case studies).

Table 2.

The values presented in Table 2 allow us to identify the box most affected by the corresponding cold-air incursion. Therefore, the cold-air incursions can be classified according to the respective box most affected. Putting together all the events that most affect the MP2 box, for instance, it is to be expected that the composite map of the associated meteorological fields presents typical characteristics of the cold-air incursions that propagate meridionally to the Amazon basin. Thus, for each box, if the most intense events are grouped together, it is reasonable to suppose that almost all of these events are related to similar meteorological systems.

At this point, it is important to highlight that the intensities of cold-air incursions should be observed with caution when considering the normalized temperature values in Table 2. For instance, the second-ranked cold-air incursion in the ALL area is ranked third for the MP2 box, with a standard deviation of −4.60, and is also ranked second for the EP box, with a standard deviation of −3.63. This case shows that the third-ranked event for MP2 has a standard deviation with a greater magnitude than the second event for the EP box. This fact is due to the deviations of the climatological distributions of the temperature in relation to the theoretical normal distribution, which implies that the use of the ranked value is more appropriate than the use of the normalized temperature to determine the box most affected by a certain cold event.

On the other hand, the correlation values of normalized temperatures between the boxes can provide more information on the different routes of the intense cold-air incursions over South America. Thus, Table 3 shows the correlation matrix between the 32 values of Table 2 for the MP1, MP2, EP, and MR boxes when they were correlated with each other. Initially, one may note a high correlation between the extreme values for the MP1 and MP2 boxes and the opposite for the EP and MR boxes. Furthermore, there are moderate to high correlations between MP2, MP1, and MR. On the other hand, moderate-to-low correlation is observed between EP and the other boxes.

Table 3.

Linear correlation matrix for the different regions studied showing the temperature anomalies given in Table 2.

Table 3.

These results make sense when considering the preferred longitudinal bands for the cold-air incursion mentioned by Cavalcanti et al. (2009) and Metz et al. (2013). The values for the western regions are correlated with each other with moderate-to-strong intensity, but the correlations are lower between the eastern region, the EP box, and the western ones, the MP1 and MP2 boxes. This means that the preferred longitudes for cold-air incursions are observed even for the most intense events. In other words, there is a split between those mostly affecting the continental region located to the west and those affecting the eastern regions and the maritime area. Other valuable information involves the higher correlation between MP1 and MP2 when compared to MP1 and MR, meaning that not all strong cold events propagate meridionally toward the Amazon basin with efficiency.

Thus, it is interesting to analyze the intense cold-air incursions in terms of their zonal and meridional displacements. As a result, this work considered three patterns of propagation, as presented in the following classification:

  • pattern 1, cold air mass with MP;
  • pattern 2, cold air mass with MR; and
  • pattern 3, cold air mass with EP.

By definition, the MP pattern represents the strongest cold-air incursions that mainly affect the MP1 and MP2 boxes; the MR pattern represents the strongest cold-air incursions that mainly affect the MR box and the EP pattern represents the strongest cold-air incursions that mainly affect the EP box. Thus, if compared to Lupo et al. (2001), the MP and MR patterns are classified as type 2, representing the incursions that affected the continental regions of South America, associated with intense anticyclogenesis over Argentina. The EP pattern, in turn, is related to type 3.

3. Characterization of patterns

In this section, the description of the patterns is based on the analyses and interpretation of composites of fields related to the selected dates when the cold-air incursions were identified. To reinforce understanding, each of the three patterns is related to its corresponding case study, which consists of the most intense cold event for the corresponding box. To begin, we clarify the details of the procedure used to make the composites, which followed an objective approach but included certain subjective criteria.

a. Approach to the composites

The cold events used to obtain the composites fields for each pattern are specified by the boldface values in Table 2. The selection criterion consisted of choosing the most representative events of cold-air incursions for each region of interest, using ranked values.

In this sense, taking as an example the event of 17 July 2010 (Table 2, row 8), we see that it is presented as the most intense cold-air incursion in the MP1 and MP2 boxes. However, it is in 10th position in the MR box and in 26th position in the EP box. Thus, in addition to being an MP pattern incursion (because of the greater intensity in the MP1 and MP2 boxes), this event also presents considerably greater intensity in the boxes corresponding to the MP pattern in relation to the boxes corresponding to the other patterns. Therefore, this event was chosen for the MP composite pattern incursion.

The technique described above was also applied to the selection of the other MP pattern cold-air incursion events, as well the selection of seven events for the MR and EP composites. At this point, it is important to stress that only the events most representative of MP2, characterized by the efficient propagation toward the low latitudes (given the box location), were used to select the events representative of the MP pattern. In this way, although there are MP pattern events that mainly affect the MP1 box, this box is located between the MP2 and MR boxes. Therefore, the choice of events through the values in the MP2 box best distinguishes MP pattern incursions from MR pattern incursions through composite maps.

The procedures above were made to avoid the selection of cold events that similarly affect all the considered boxes and make clear the classification from the composite images. The dates related to the specific days with the lowest temperatures were denoted as day 0 and were adopted as the reference date for that specific event.

b. The event of 17 July 2010: MP

The 17 July 2010 event was classified as the eighth most intense cold-air incursion occurring in the ALL area, but at the same time it is the most intense cold-air incursion in the southwestern Amazon basin (MP2 box). From Table 2, the observed anomaly in the ALL area is −2.85σ, but for the MP2 box it is −5.64σ. This event was characterized by an intense extratropical anticyclone of 1044 hPa across Argentina (Table 1) and featured the continental penetration of a cold-air incursion to very low latitudes. According to INMET observations, the temperature in Rio Branco, located at 6°S, varied from 10.5° to 14.7°C during 17 July and dropped to 9.9°C on 19 July. For locations in the southern part of Amazonas state, like Humaitá and Lábrea, the minimum temperature reached 13°C at the apex of the event (INMET stations).

Figure 2 presents the time evolution of the cold fronts for the event of 17 July 2010, created by consulting meteorological fields through ERA-Interim reanalysis data. It shows that this event was a combination of two cold fronts: the first with the typical track associated with intense frontal systems over the continent (Marengo et al. 1997; Garreaud 2000; Lupo et al. 2001) and the second confined to the area over the interior of the continent. Together, they formed a sequence of cold days for almost a week over part of the Amazon basin. During the frontolysis of the first system and the subsequent frontogenesis, the postfrontal anticyclone formed two high pressure systems on 18 July, with one of them confined to the area over the continent, reaching Paraguay and Bolivia.

Fig. 2.
Fig. 2.

Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 17 Jul 2010. The primary cold front is denoted by a black line and the secondary front by a gray line. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

Figure 3 presents superimposed meteorological fields from reanalysis variables, for synoptic and dynamic characterization of the 17 July 2010 event, at 1200 UTC on day 0. From the analysis of Figs. 3d–f, one can identify the upper-level vortex crossing the Andes Mountains over northern Chile. The vortex was generated by the interaction of the frontal trough with the topography, in which the southern part followed the westerlies and the northern part was trapped, forming a cutoff low.

Fig. 3.
Fig. 3.

Meteorological fields for 1200 UTC 17 Jul 2010: (a) air temperature at 2 m (shaded; °C), sea level pressure (contours; hPa), and 850-hPa winds (vectors; m s−1); (b) 1000–500-hPa thickness (contours; dam), thermal wind (vectors; m s−1), and air temperature anomaly at 2 m (shaded; °C); (c) 850-hPa winds (vectors; m s−1), thermal advection (shaded; × 10−4 °C s−1), rates of change of geopotential height [black contours; dam (12 h)−1], and omega (red contours; Pa s−1); (d) 500-hPa relative vorticity advection (shaded; × 10 −9 m s−2) and geopotential height (solid contours; dam) and 700-hPa omega (dashed contours; Pa s−1); (e) 925-hPa thermal advection (shaded; × 10−4 °C s−1, color bar), 500-hPa relative vorticity (contours; 10−9 s−1), and 250-hPa streamlines with wind speed (shaded; scale at top-right corner; m s−1); and (f) anomaly of sea level pressure (shaded; hPa), anomaly of 500-hPa geopotential height (contours; dam), and anomaly of 250-hPa wind (vectors; m s−1). The highlighted box indicates its associated pattern.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

The cutoff low affected the surface pressure field through the formation of an inverted trough to the north of the frontal anticyclone (Fig. 3a), which later became associated with the low pressure center responsible for the secondary continental front (Fig. 2). The persistence of southerly winds, related to the split of the postfrontal high and the frontogenesis, maintained the cold advection at lower levels over central-western Brazil and neighboring countries (Figs. 3c,e), causing the most significant anomalies to be located over Bolivia and adjacent areas of Brazil (Fig. 3b).

Figure 4 shows the superimposed meteorological fields of the composite obtained by the selection of the seven selected events of cold-air incursions of the MP pattern. Each of the image types in Fig. 3 (case study) is distributed within the columns of Fig. 4 (composite), which is also associated with the time evolution from day −2 to day +1 (always at 1200 UTC). When comparing the sequence of Figs. 3b, 3c, and 3e and the sequence of Figs. 4b, 4c, and 4e, in both figures the MP pattern is characterized by an intense and persistent cold advection at lower levels over Bolivia and central-western Brazil (Figs. 4c,e), with the most significant anomalies located in the MP1 box (Fig. 4b).

Fig. 4.
Fig. 4.

As in Figs. 3a–f, but for the MP pattern composite.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

In addition, the Fig. 3 cutoff low is reflected in the detached trough over northern Chile and in its associated geopotential anomaly field in the composites (Figs. 4d,f). Thus, the interaction of the frontal trough with the Andes is considerable, reinforcing the channeling effect imposed by the topography. This happens because the trough–ridge system, over the mountains, forces the low-level southern winds through the acceleration provided by the predominance of the pressure gradient force at the leading edge of the postfrontal anticyclone that is intensifying (Garreaud 2000). Consequently, the cold-air advection is enhanced, contributing to the effective displacement of the postfrontal high to the north (Figs. 3a, 3f, 4a, and 4f) (Marengo et al. 1997; Garreaud 2000; Lupo et al. 2001).

Therefore, the MP pattern is marked by the interaction of the frontal trough with the Andes Mountains and the establishment of temperature anomalies mainly between central-western Brazil and neighboring countries, related to the persistent cold advection of temperatures east of the Andes. If a second frontal system appears (as in the case study), all of the described MP features are intensified as the postfrontal anticyclone gets stuck, together with associated cold air, over low latitudes.

c. The event of 12 July 2007: MR

The 12 July 2007 cold wave was the ninth most intense in the last 15 years, with an anomaly of −2.74σ in the ALL area and −2.85σ in the MR box. It was associated with a postfrontal high of 1036 hPa and brought snow to the city of Buenos Aires for the first time in 89 years. It also came with very low temperatures in Argentina, with a record of −22°C in Bariloche (Table 1).

Figure 5 shows the trajectory of some of the major synoptic systems, which followed the typical characteristics for the great continental cold-air incursions (Garreaud 2000). However, contrary to what may be observed in Fig. 2, there was no bipartition of the postfrontal high, and the cold wave persisted for less time over the continent.

Fig. 5.
Fig. 5.

Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 12 Jul 2007. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

Figure 6 shows that the most significant low temperature anomalies (Fig. 6b) are concentrated more to the south than the MP pattern (Fig. 3b) although the cold front was able to advance to 10°S (Fig. 5). The frontal trough advances over northern Chile, but does not result in the formation of a cutoff low (Figs. 6d–f) nor in the formation of a secondary front (Fig. 5). Thus, even though the cold front reaches the Amazon basin, these dynamic processes make it difficult for effective penetration of the cold-air incursion to the north, leaving the most significant anomalies to the south.

Fig. 6.
Fig. 6.

As in Fig. 3, but for 1200 UTC 12 Jul 2007.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

Moreover, Fig. 7 shows the jet stream to be very pronounced over the entire width of the continent (Fig. 7e), which is also observed in the case study (Fig. 6e). Garreaud (2000) analyzed the importance of the entrance region of the upper-level jets for the generation of an additional dynamic support for the cold-air incursions, as Uccellini and Johnson (1979) showed that a thermally direct circulation exists in this region leading to southerly winds at low levels of the troposphere. Thus, for the MR pattern, the absence of the jet stream entrance over the continent does not allow for a contribution of southerly ageostrophic winds to the cold-air incursion intensification.

Fig. 7.
Fig. 7.

As in Figs. 3a–f, but for the MR pattern composite.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

Nevertheless, the composites show that the temperature drops considerably over the MR box (Fig. 7b), where cold advection and subsidence are observed (Fig. 7c) under the frontal trough (Fig. 7e). These processes are related to the intensification of the frontal trough via the feedback mechanism described in section 1 (Marengo et al. 1997; Lupo et al. 2001). However, from day 0 the cold advection and subsidence begins to be located to the north of the frontal trough (Fig. 7e). Consequently, the feedback ends, and, together with the lack of dynamic support for the more effective advancement of cold air to the north, causes the rapid dissipation of the cold-air incursion, making the southernmost region (the MR box) the most affected by the corresponding pattern.

MR events are also characterized by the presence of a transient cyclonic circulation during the days preceding the maximum intensification of the cold air mass. In Fig. 7f, the anomalous fields allow us to observe a negative zone of a surface pressure anomaly, which appears around 45°S and 45°W on day −2 and then moves rapidly southeast. In the evolution of the mean pressure field in Fig. 7a, this anomaly does not even set up a surface cyclone. However, a separate analysis of the seven cases included in the composite reveals the presence of the extratropical cyclone (not shown), which is masked by the average procedure of the composites due to variability in its position. The case study of the 2007 event follows as an example, where, in Fig. 6f, the cyclone still appears to be defined but is centered outside the domain of the figure. Its trajectory was not shown in Fig. 5 to avoid an excess of information and to preserve aesthetics.

In summary, the MR events are characterized by considerable drops in temperature over the corresponding box (Fig. 7b), due to cold advection and subsidence (Fig. 7c). This fact is related to the considerable development of the ridge–trough system (Fig. 7d) by the feedback process, days before the maximum intensification of the cold air. The presence of the jet across the continent (Figs. 6e and 7e) hinders substantial cooling in lower latitudes because of the lack of dynamic support by the thermally direct circulation.

d. The event of 17 July 2000: EP

The 17 July 2000 cold-air incursion was the most intense in the last 17 years, as well as one of the most significant events in the last 100 years (Table 1), due to the extremely low temperature values, freezes, and snowfall. The extreme freezes were registered in almost all parts of southern Brazil and in parts of the interiors of the São Paulo, Rio de Janeiro, and Minas Gerais states, together with the fourth occurrence of snow in the last 100 years in Porto Alegre city (Weatherboy 2007). In addition, the minimum temperature reached 0.2°C in São Paulo (state of São Paulo, IAG station) on 17 July, and the maximum temperature reached −2.0°C in São Joaquim (State of Santa Catarina, INMET station) (Table 1).

According to Table 2, the day of 13 July 2000 presents the most negative temperature anomaly in the ALL area for the whole series considered in this study: −3.59σ. However, over box EP, the greatest negative temperature anomaly occurred on 17 July 2000 (not shown), with a deviation of −3.86σ. This difference occurs because the 2000 cold wave was made up of several successive cold-air incursions.

Each cold-air incursion was accompanied by an intense cold front, with its associated postfrontal high. Considering the formation days of the postfrontal high over Argentina and its dissipation by the ocean (which determine the beginning and the end of a cooling period in South America), the analysis of the reanalysis fields allows us to divide the cold wave into four pulses: the first pulse from 6 to 15 July, second pulse from 15 to 18 July, third pulse from 18 to 21 July, and fourth pulse from 21 to 26 July 2000. Thus, the second pulse brought the most intense cold wave during 15 years of reanalysis for the EP box, and as a result, this second pulse was selected to be the representative case of cold-air incursion for the EP pattern.

Figure 8 shows the cold front track and its associated extratropical cyclone and anticyclone. One can observe the frontal system propagating quickly over South America, forming on 14 July over Argentina and dissipating on 18 July 2000 over the northeast region of Brazil. The high pressure system initially spread to northern Argentina, following the usual path for large cold-air incursions. But in the following days, the anticyclone continued moving across the continent, reaching the southeast of Brazil before arriving at Espirito Santo state.

Fig. 8.
Fig. 8.

Frontal evolution with the tracks of cyclones and anticyclones for the cold incursion of 17 Jul 2000. The letters H and L denote, respectively, the locations of high and low pressure centers. The corresponding dates are indicated in the side boxes at 24-h intervals, always at 1200 UTC.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

The cold front was accompanied by an intense extratropical cyclone (Figs. 9a,f) and was associated with a large frontal trough over the ocean (Figs. 9d,e), yielding to strong cold advection, subsidence, and a consequent increase of the geopotential height (Fig. 9c). These features agree with the low-level anticyclogenesis described by Marengo et al. (1997) from a quasigeostrophic perspective, according to which the cold advection centered around the layer of 700 hPa causes subsidence and an increase in the geopotential height at 850 hPa. In this sense, the lowest absolute temperatures where frosts and negative values occurred (Table 1) appear concentrated in the EP box (Fig. 9b), on the path of the postfrontal anticyclone (Fig. 8).

Fig. 9.
Fig. 9.

As in Fig. 3, but for 1200 UTC 17 Jul 2000.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

The features cited in Fig. 9 are maintained when analyzing the composites of Fig. 10. Furthermore, Fig. 10f allows us to identify the cyclone in Fig. 9f, where the composites show that it persists from day −2 through day +1. As in Figs. 9d and 9f, the cyclone is associated with a large frontal trough, with its axis extending from the ocean to the continent, passing through the highlighted EP box (Figs. 9d,f). Finally, both Figs. 9c and 10c show intense cold advection, subsidence, and the positive increase trend of geopotential height at 850 hPa in the EP box and adjacent regions.

Fig. 10.
Fig. 10.

As in Figs. 3a–f, but for the EP pattern composite.

Citation: Weather and Forecasting 33, 5; 10.1175/WAF-D-17-0159.1

The southerly winds induced by the pressure gradient between the cyclone and the anticyclone (Figs. 10a,f) particularly contribute to the cold-air incursion from northern Argentina through advection (Fig. 10c). This fact agrees with the results of Marengo et al. (1997) and shows the importance of the cyclone (and not only the anticyclone) for the cold-air incursion, as observed by Lupo et al. (2001) and emphasized by Pezza and Ambrizzi (2005). Moreover, Metz et al. (2013) states that the Brazilian plateau, while smaller than the Andes, also contributes to the incursion of cold air to the north, through the same effects produced by the Andes Mountains. In this way, the trajectory of the postfrontal anticyclone in Fig. 8 follows the Brazilian Highlands pathways described by Metz et al. (2013).

When considering the work of Lupo et al. (2001), it can be summarized that EP events fall into the category of cold-air incursions that affect the eastern latitudinal band. Because they also penetrate into the continent forced by the orography of the Andes, they fit into type 2 incursions of Lupo et al. (2001), which later become type 3 as they move eastward. In this transformation, the cyclone and the Brazilian Highlands, respectively, play important roles through the cold advection promoted by a pressure gradient with the anticyclone in the continent (Marengo et al. 1997; Pezza and Ambrizzi 2005) and by the topographic effects.

e. Comparison

Up to this point, the intense cold-air incursions have been identified, classified, and characterized in their synoptic and dynamic aspects, according to the three incursion patterns corresponding to their respective boxes: the MP, MR, and EP patterns. Each pattern contains its unique characteristics, which will now be highlighted through comparative analysis.

First, the MP pattern is characterized by the greater interaction of the frontal trough with the Andes over northern Chile (see Fig. 4d). Consequently, southerly winds were stronger, promoting colder advection and subsidence at lower levels over Bolivia and central-western Brazil (cf. Figs. 3c, 3e, 4c, and 4e with Figs. 6c, 6e, 7c, and 7e, and Figs. 9c, 9e, 10c, and 10e). Therefore, the most significant temperature anomalies are located farther north in comparison to the other patterns, as discussed in section 3b (cf. Figs. 3b and 4b with Figs. 6b and 7b, and with Figs. 9b and 10b).

In some cases (four of the seven cases selected for the composite), the frontal trough over northern Chile (Fig. 4d) causes the upper-level vortex formation (cutoff low) (Fig. 3d), which can form the continental secondary frontal system (Fig. 2). When it appears, the postfrontal anticyclone can split into two parts (Fig. 2), where one of them becomes stuck east of the Andes Mountains, which causes even more intense and persistent cold events over lower latitudes.

In the MR pattern, the jet stream appears pronounced over the entire width of the continent (Figs. 6e and 7e). It differs from the composites of the MP (Fig. 4e) and EP (Fig. 10e) patterns, where its entrance is located over the southern region of Brazil. This particularity blocks the contribution of the direct thermal circulation (Uccellini and Johnson 1979; Garreaud 2000), impeding the more effective meridional advance of cold air.

Nevertheless, the cooling caused by the MR events is associated with the most significant temperature anomalies during their intensity peak (cf. Figs. 4a,b with Figs. 7a,b, and with Figs. 10a,b). This results from the efficient feedback that intensifies the frontal trough and the associated cold-air incursion, as described in section 3c and evaluated by Marengo et al. (1997) and Lupo et al. (2001). However, although noteworthy for the MR pattern, this feedback also occurs in the EP and MP pattern before the peak of a cold event, as can be seen in Figs. 4e, 7e, and 10e through the cold advection under the frontal trough. Therefore, the feedback process is a feature of the intense cold-air incursions of any pattern.

In the EP pattern, the lower-level cold advection is considerably more intense in the corresponding box (cf. Figs. 10c,e with Figs. 4c,e and with Figs. 7c,e), resulting in the lowest temperatures over the states of Santa Catarina, Parana, and São Paulo (cf. Figs. 10a,b with Figs. 4a,b and with Figs. 7a,b). The EP events also feature a faster propagation of the cold front in relation to the other patterns of incursions, in agreement with Fortune and Kousky (1983) when taking into account the eastern longitudinal band of this pattern (see section 1 and cf. Fig. 8 with Fig. 2 and with Fig. 5). Additional features include the substantial contribution of the extratropical cyclone and the Brazilian Highlands pathway to the cold-air incursion, as already described in section 3d.

In summary, the MP and MR patterns are characterized by cold-air incursions that mainly affect western longitudinal bands (MR1, MR2, and MR boxes; Fig. 1). Both are considerably influenced by the topography of the Andes, being differentiated by the meridional position of the most-affected region. The EP pattern mainly affects southern Brazil and adjacent countries, receiving support from the cyclone over the ocean and the influence of the Brazilian Highlands. Furthermore, the cutoff low over northern Chile, the continuous jet stream over the entire South America width, and the substantial cyclone constitute key identifying features for the MP, MR, and EP patterns, respectively.

4. Conclusions

This paper introduces a classification system for the intense cold-air incursions over South America, based on the categorization of events in three different patterns: meridional penetration (MP), meridional restriction (MR), and east penetration (EP). It thus complements the previous literature as it presents unique results regarding the synoptic and dynamic characterization of the extreme events according to their regional areas.

For the different patterns of incursions, this study allowed us to observe the importance of the zonal positioning of the upper-level trough and its interaction with the orography of the Andes and with the highlands of Brazil. Some features already reported in previous studies, such as the contribution of the extratropical cyclone, the jet entry position, and the feedback mechanism (Uccellini and Johnson 1979; Marengo et al. 1997), were reevaluated within the context of the identified patterns and shown to be key features in the process of pattern identification. In addition, the pattern differentiation of the meridionality of the incursions is an innovation in this study since it was not identified by previous studies.

In fact, cold-air incursions were considered without a classification system by the majority of previous studies, as in Marengo et al. (1997), Garreaud and Wallace (1998), Garreaud (2000), and Pezza and Ambrizzi (2005). The studies of Fortune and Kousky (1983), Cavalcanti et al. (2009), and Metz et al. (2013) did mention the two longitudinal regions of propagation, but only the work of Lupo et al. (2001) classifies the cold-air incursions, which are detailed for the most intense cases and by their meridional propagation in this paper. The meridional analysis allowed us to observe that the MP and MR patterns belong to the type 2 incursions of the abovementioned work.

Another highlight is the methodology used to identify the patterns of cold-air incursions. The locations of the boxes were determined based on the literature cited in the paragraph above, but not arbitrarily. In this sense, previous knowledge allowed us to identify the preferred longitudinal bands for the penetration of the cold fronts, which were statistically detailed for the extreme events through the analysis of the correlations. Thus, it became possible to determine the existence of three patterns of cold-air incursions, which were later characterized using the composites. Therefore, the determination of the patterns did not come immediately with the determination of the boxes, which resulted in a greater degree of credibility for the patterns found.

From a practical point of view, this study presents a contribution to weather forecasting through the interpretation of the numerical model outputs. Through the correct understanding of the meteorological fields in their key features for each pattern, a forecast of the localities to be most affected becomes more reliable. Consequently, it becomes possible to make more satisfactory decisions related to mitigating the economic and social damages caused by extreme cold. This means potential decreases in losses of coffee production and other crops vulnerable to frosts at low temperatures.

Moreover, this article furthers the study by Pezza and Ambrizzi (2005), who examined the records of the intense cold-air incursions over South America from 1888 to 2003, by presenting a complete record up to 2016. The results indicate that extreme cold events have been taking place recently, and so the importance of the continuity of related studies is considerable. In addition, it is generally agreed upon by most scientists that extreme events will continue to take place within the context of climate change (Russo et al. 2014), and cold-air incursions are no exception (Pezza and Ambrizzi 2005).

Finally, new studies can be carried out that explore other parts of the subcontinent not reported upon in this study, such as the inverse temperature relationship between Patagonia and southeast Brazil during the winter (Pezza and Ambrizzi 1999). The cold waves over Patagonia might be related to a fourth pattern over South America, and this is a reason why its determination can contribute to expanding our current knowledge of the cold waves to the entire subcontinent.

Acknowledgments

The authors thank Coordenação de Aperfeiçoamento Pessoal de Nível Superior (CAPES) and the São Paulo Research Foundation (Grant 2018/16658-9) for financial support. We also thank the reviewers for their thoughtful comments that have contributed substantially to the improvement of this paper.

REFERENCES

  • Cavalcanti, I. F., N. J. Ferreira, M. G. A. J. da Silva, and M. A. F. da Silva Dias, 2009: Tempo e Clima no Brasil. Oficina de Textos, 463 pp.

  • Crossett, C. C., and N. D. Metz, 2017: A climatological study of extreme cold surges along the African highlands. J. Appl. Meteor. Climatol., 56, 17311738, https://doi.org/10.1175/JAMC-D-15-0191.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dee, D., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fortune, M., and V. Kousky, 1983: Two severe freezes in Brazil: Precursors and synoptic evolution. Mon. Wea. Rev., 111, 181196, https://doi.org/10.1175/1520-0493(1983)111<0181:TSFIBP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gan, M., and V. Rao, 1994: The influence of the Andes Cordillera on transient disturbances. Mon. Wea. Rev., 122, 11411157, https://doi.org/10.1175/1520-0493(1994)122<1141:TIOTAC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., 2000: Cold air incursions over subtropical South America: Mean structure and dynamics. Mon. Wea. Rev., 128, 25442559, https://doi.org/10.1175/1520-0493(2000)128<2544:CAIOSS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garreaud, R. D., and J. M. Wallace, 1998: Summertime incursions of midlatitude air into subtropical and tropical South America. Mon. Wea. Rev., 126, 27132733, https://doi.org/10.1175/1520-0493(1998)126<2713:SIOMAI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Girardi, C., 2006: O poço dos Andes. Bull. Braz. Soc. Meteor., 30, 39–48.

  • Hackbart, E., 2010: Argentina se cobre de branco e sul do Brasil registra neve e raro episódio de chuva congelante. Anima Meteorologia, http://anitameteorologia.blogspot.com/2010/07/argentina-se-cobre-de-branco-e-gelo-e.html.

  • Hamilton, M. G., and J. R. Tarifa, 1978: Synoptic aspects of a polar outbreak leading to frost in tropical Brazil, July 1972. Mon. Wea. Rev., 106, 15451556, https://doi.org/10.1175/1520-0493(1978)106<1545:SAOAPO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ingemi, D., 2012: La prima ondata di freddo polare in Sud-America porta la neve su alcune città argentine, fiocchi pure sui quartieri di Buenos Aires e venti di tempesta sulla Terra del Fuoco. MeteoWeb, http://www.meteoweb.eu/2012/06/la-prima-ondata-di-freddo-polare-in-sud-america-porta-la-neve-su-alcune-citta-argentine-fiocchi-pure-sui-quartieri-di-buenos-aires-e-venti-di-tempesta-sulla-terra-del-fuoco/138462/.

  • Krishnamurti, T., M. Tewari, D. Chakraborty, J. Marengo, P. Silva Dias, and P. Satyamurty, 1999: Downstream amplification: A possible precursor to major freeze events over southeastern Brazil. Wea. Forecasting, 14, 242270, https://doi.org/10.1175/1520-0434(1999)014<0242:DAAPPT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liebmann, B., G. N. Kiladis, L. V. Carvalho, C. Jones, C. S. Vera, I. Bladé, and D. Allured, 2009: Origin of convectively coupled Kelvin waves over South America. J. Climate, 22, 300315, https://doi.org/10.1175/2008JCLI2340.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lupo, A., J. Nocera, L. Bosart, E. Hoffman, and D. Knight, 2001: South American cold surges: Types, composites, and case studies. Mon. Wea. Rev., 129, 10211041, https://doi.org/10.1175/1520-0493(2001)129<1021:SACSTC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marengo, J., 2001: Infoclima – July’13, 2001. INPE/CPTEC, http://infoclima1.cptec.inpe.br/~rinfo/infoclima/jul_2001.shtml.

  • Marengo, J., A. Cornejo, P. Satyamurty, C. Nobre, and W. Sea, 1997: Cold surges in tropical and extratropical South America: The strong event in June 1994. Mon. Wea. Rev., 125, 27592786, https://doi.org/10.1175/1520-0493(1997)125<2759:CSITAE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Metz, N. D., H. M. Archambault, A. F. Srock, T. J. Galarneau, and L. F. Bosart, 2013: A comparison of South American and African preferential pathways for extreme cold events. Mon. Wea. Rev., 141, 20662086, https://doi.org/10.1175/MWR-D-12-00202.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parmenter, F., 1976: A Southern Hemisphere cold front passage at the equator. Bull. Amer. Meteor. Soc., 57, 14351440, https://doi.org/10.1175/1520-0477(1976)057<1435:ASHCFP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pezza, A. B., and T. Ambrizzi, 1999: Um Estudo das Flutuações de Temperatura para o Período de Inverno da América do Sul, Correlacionando a Patagônia com o Sul do Brasil. Rev. Bras. Meteor., 14, 2334.

    • Search Google Scholar
    • Export Citation
  • Pezza, A. B., and T. Ambrizzi, 2005: Cold waves in South America and freezing temperatures in São Paulo: Historical background (1888-2003) and case studies of cyclones and anticyclones tracks. Rev. Bras. Meteor., 20, 141158.

    • Search Google Scholar
    • Export Citation
  • Revista Cafeicultura, 2010: Especial: 35 Anos de Geada de 1975—Entenda o que foi a geada negra que dizimou todas as plantações de café do Paraná. Cafeicultura, accessed 21 December 2016, http://www.revistacafeicultura.com.br/index.php?mat=34022.

  • Russo, S., and et al. , 2014: Magnitude of extreme heat waves in present climate and their projection in a warning world. J. Geophys. Res. Atmos., 119, 12 50012 512, https://doi.org/10.1002/2014JD022098.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seluchi, M., Y. Serafini, and H. Le Treut, 1998: The impact of the Andes on transient atmospheric systems: A comparison between observations and GCM results. Mon. Wea. Rev., 126, 895912, https://doi.org/10.1175/1520-0493(1998)126<0895:TIOTAO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sturcke, J., 2007: Buenos Aires has first snow 1918. The Guardian, accessed 29 April 2017, https://www.theguardian.com/world/2007/jul/10/argentina.weather.

  • Uccellini, L. W., and D. R. Johnson, 1979: The coupling of upper and lower tropospheric jet streaks and implications in development of severe convective storms. Mon. Wea. Rev., 107, 682703, https://doi.org/10.1175/1520-0493(1979)107<0682:TCOUAL>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weatherboy, 2007: Inverno 2000 – O inverno mais rigoroso dos últimos 50 anos. Accessed 21 December 2016, https://www.youtube.com/watch?v5GRgS21VILlY.

  • Xu, Q., 1990: A theoretical study of cold air damming. J. Atmos. Sci., 47, 29692985, https://doi.org/10.1175/1520-0469(1990)047<2969:ATSOCA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
Save