1. Introduction
Deep convection frequently occurs east of the Andes Cordillera, in the northern part of Argentina. From October to March, heavy hail stones that reach the ground can produce severe damage to cultivated land. Such events have particularly large consequences, for example in the Mendoza area, a region famous for its vineyards. In this paper, we study the main dynamical conditions that are required to generate the warm and moist air masses necessary for deep moist convection, and how they can be met in the complex orographic situation of northern Argentina and the Mendoza region. The height of the Andes is such that humidity cannot originate from the Pacific Ocean area: crossing the Andes results in a very strong drying of the air masses. Humid air masses are usually carried to the Mendoza area by the so-called low-level jet (LLJ). The LLJ consists of a low-level northerly circulation that is concentrated in a relatively narrow region located toward the east of the mountain barriers (Stensrud 1996; Marengo et al. 2002; Douglas et al. 2000; Paegle 2000). It is known to be the main cause for the transport of warm and moist air from the Amazon River basin toward central South America. The LLJ has been the focus of the recent South American Low-Level Jet Experiment (SALLJEX) in situ experiment and has been extensively studied in previous papers (Berri and Inzunza 1993; Nogues-Peagle and Mo 1997; Mo and Higgins 1996; Li and Le Treut 1999; Saulo and Nicolini 2000; Silva Dias 2000; Berbery and Collini 2000; Salio et al. 2002; Berbery and Barros 2002; Wang and Fu 2004; Liebmann et al. 2004; Marengo et al. 2004). The LLJ has also been studied through simulations using both global and mesoscale models (Saulo et al. 2000; Misra et al. 2000; Nicolini and Saulo 2000; Vernekar et al. 2003). A well-documented situation is when the northerly winds within the LLJ may weaken or even change direction and become southerlies in the presence of the South Pacific anticyclone (Wang and Fu 2004). This LLJ wind reversal is generally associated with a temperature decrease and a moisture depletion that results in deep convection suppression over the Mendoza area.
The objective of the paper is to show that the association between the LLJ and convection in the Mendoza region is not systematic. There is an alternative source for moist air, which does not require the presence of the LLJ, and can also provide the humidity needed for deep convection. A key meteorological factor affecting the origin of air masses and humidity entering northern Argentina is the position of the anticyclones traveling from west to east, from the Pacific to the Atlantic Ocean. From the detailed documentation of several test cases we show in particular that, after the South Pacific anticyclones have triggered the suppression of convection described above, when they cross the continent farther to the east and their western edge reaches the east coast, temperature may increase again, and deep convection may reappear over the Mendoza area. This process can occur independently from the LLJ; it is then a direct consequence of the transport of warm and moist air from Uruguay, southeast Brazil, and also directly from the Atlantic Ocean. This implies that convection in northern Argentina is not necessarily the consequence of a moisture supply from the Amazonian basin.
We document this possibility through the analysis of individual cases, rather than statistical studies using various types of averaging: we explain further why a purely statistical approach is not easy and would require the development of sophisticated tools that are beyond the scope of the present paper. A detailed analysis of real test cases is therefore necessary at this exploratory stage to point out the links between complex processes occurring at different scales.
The succession of test cases is chosen to illustrate how a convection situation associated with the presence of an active LLJ can be modified by the motion of an anticyclone through the South American continent, with phases when the convection is suppressed and phases when it resumes, but following a dynamical scheme that is not necessarily linked with the LLJ. Section 2 describes the methods and the data being used. Section 3 details a reference case where deep convection is observed together with the LLJ occurrence. Section 4 presents a series of test cases where this initial situation is affected by different positions of the anticyclone. Comments, summary, and conclusions are presented in section 5.
2. Method and data being used
To study the local and regional atmospheric conditions, we use a combination of data from various sources. The analyses of the European Centre for Medium-Range Weather Forecasts (ECMWF) provide an important source of information. They are available every 6 h, on a grid whose horizontal resolution is less than 1°. Along the vertical, they are provided over 60 levels, distributed from the ground up to 1 hPa. Higher spatial and temporal resolutions are necessary for our analysis and they are obtained by using the Weather Research and Forecasting Model (WRF; Skamarock et al. 2005) in a two-domain nested version. This model is a last-generation mesoscale code that can simulate regional weather systems, with initial and boundary conditions being provided by global models. WRF has been adapted to use data from ECMWF analyses, because its common interface is adapted to a few global models only, such as the model of the National Centers for Environmental Prediction (NCEP). In our WRF simulations we used a horizontal resolution of 21 km for the mother domain, and 7 km for the nested one. Vertically the model used 62 levels from the ground up to 10 hPa. The time step is 30 s for the mother domain, and 10 s for the nested one. An additional dynamical tool used in the paper is the “FLEXTRA” code, a Lagrangian kinematic trajectory model that computes several different types of trajectories (three-dimensional, isobaric, isentropic, etc.) and can help to define the origin of air masses (Stohl and Wotawa 1995; Stohl et al. 1995). The time step of the FLEXTRA code is continuously adjusted, convergence being assumed when the trajectory position between two subsequent iterations differ by less than 0.000 01 grid units.
To analyze the occurrence and characteristics of convective systems we also need to resort to other types of observations (i.e., satellite measurements). We determine the temperature at the top of clouds, a key diagnostics of convection intensity, from observations provided by the Service d’Archivage et de Traitement Météorologique des Observations Satellitaires (SATMOS) system (Sèze and Desbois 1987). The system was developed by Météo-France and the Institut National des Sciences de l’Univers (INSU) and provides an analysis of Geostationary Operational Environmental Satellite-8 (GOES-8) satellite data, which are calibrated at 0.01 K. The temperature at the top of clouds together with temperature profiles from WRF simulations are used to estimate the altitudes of clouds. We present here five situations in which convection is not triggered by the LLJ, both of which occur during the month of January 2001. Our approach can of course be extended to treat a larger number of cases, but this is beyond the objective of the present paper.
3. Convection associated with the LLJ: The case of 9 January 2001
In this section we have selected a test case that corresponds to a typical situation where moisture is being brought to northern Argentina by the LLJ and serves to trigger convection. This case occurred at 2100 UTC 9 January 2001. It is illustrated through different meteorological fields shown in Fig. 1. The cloud-top temperature distribution obtained from SATMOS is used to show the existence of several areas of cloud activity within the region under study. In particular, a small cloud in the form of a blob is present around 34°S, 64°W with a cloud-top temperature reaching 210 K throughout the blob. The cloud top therefore reaches about 12 km, a feature that is also confirmed by the distribution of the relative humidity simulated by WRF and presented in a longitude–altitude section at 34°S. This figure should be compared with the tropopause level at 16 km. They are characteristic of a deep convection case.
Figure 1b shows the pressure at the ground, and the wind at 850 hPa obtained from the ECMWF analysis on the same day. The 850-hPa level is very commonly chosen as a representative level to study the wind maximum within the LLJ (see the numerous references in section 1) and we also illustrate the events occurring in our test case by results at that level. In addition, we use surface pressure as an indicator of the anticyclone position. On 9 January, the highest surface pressure values are found within an anticyclone located over the Pacific Ocean. This anticyclone drifts eastward. The existence of eastward-propagating anticyclones (EAs) originating from the Pacific Ocean is a well-documented climatological feature (Jones and Simmonds 1994). These EAs arrive episodically at the coast of Chile, and cross the continent thereafter. In the situation under study, the residual of a former EA that has crossed the continent at an earlier time can be observed over the eastern part of the continent. On 9 January the eastern edge of the EA lying over the Pacific is getting close to the South American continent, but, at this early stage, the winds induced by the anticyclone have not crossed the mountains. They do not yet affect the LLJ, which is still characterized by northerly winds over the continent, at least in the regions located north of 40°S (as illustrated by Fig. 1).
We can also use the skew Tρ–logp diagram in Fig. 2a, from the associated WRF simulation, to get a hint at the mechanisms that trigger convection. We can determine the level of free convection (LFC) of the convective episode occurring at 34°S, 64°W: it is situated at 700 hPa, whereas the lifted condensation level is at around 800 hPa. The convective available potential energy (CAPE) has a value of 1064 J kg−1, which sets favorable conditions for very high convection, but its initiation requires a mechanism to lift the air parcels up to the LFC. Results from a WRF simulation (not shown here) clearly indicate that mountain waves may play this role. Figure 2b shows a longitude–altitude cross section of relative humidity at 35°S. At the same location, as indicated by Fig. 2a there is an occurrence of very deep convection as expected by the CAPE value. Furthermore, the deformation of the isolines of potential temperature above the cloud top can be interpreted as waves generated by the convection.
4. The impact of the anticyclone position on humidity transport
In this section, we consider a number of situations occurring just before and just after the case of 9 January previously described, where moisture is no longer being brought by the LLJ to northern Argentina.
On 3 January, for example, as shown in Fig. 3, the western edge of an anticyclone situated over the Atlantic Ocean, centered at about 35°S, induces a strong input of moisture from the Atlantic at about 25°S. If we compute the moisture flux across the two barriers that are shown in Fig. 3b by a dark line, we find that moisture entering the zone under study is 54.6 kg m−2 s−1 from the east while the contribution from the north is negative and equal to 2.1 kg m−2 s−1. The cloud-top temperature (Fig. 3a) has a minimum lower than 210 K. This low temperature corresponds to an altitude of 15 km when compared with a temperature profile provided by WRF. The fact that there is deep convection is confirmed by the longitude–altitude cross section of relative humidity (Fig. 3c).
At 0000 UTC 11 January, the situation is very different from the two cases of 3 and 9 January. An anticyclone has partially penetrated into the continent, as evidenced in Fig. 4a. The anticyclonic winds have crossed areas situated between 40° and 50°S, where mountains with a lower elevation are present. The LLJ has been strongly affected by the EA, and the northerly winds are now replaced by southerlies. At the same time, humidity over the Mendoza area is drastically reduced, and the region becomes cloud free.
The geopotential height shown in Fig. 4b, and the winds from ECMWF at 850 hPa indicate that they are very close to a geostrophic equilibrium. South of 20°S, the LLJ has not been removed but rather deviated toward the east. Indeed, a GOES image for the same day shows that cloud activity is also shifted toward the east, and is organized along the deviated path of the LLJ.
The cloud data from SATMOS (not shown) indicate that the Mendoza region remains subsequently dominated by clear sky until 2100 UTC 15 January, when some small clouds appear again.
By 0300 UTC 16 January, a deep convection event is observed again in the region as depicted in Fig. 5. The observed cloud-top temperatures constitute a clear indicator of the strength of the convection. The surface pressure and the wind at 850 hPa obtained from the ECMWF analysis show that the EA has crossed the continent and that its western edge is located over the Atlantic Ocean.
We note that, although convection is again active, the LLJ is not restarted in its usual location. The winds’ direction over the Atlantic coast suggests that humidity is transported from the east. Figure 5 also shows a diagnostic of CAPE computed from WRF simulations for the same day at 0300 UTC. High CAPE values are found within the Mendoza region. These high CAPE values and the deep convection that they produce are the result of warm and moist air getting to the region under study. When humidity is either transported from the north by the LLJ or from the east, the temperature over the Mendoza region increases. Otherwise, when the LLJ is deviated toward the east by the anticyclone, and when winds in the usual location of the LLJ are replaced by southerlies, temperature over the region decreases. The mean temperature at 850 hPa averaged over the region between 25°–40°S, 55°–70°W has been calculated for the 3-day period previously analyzed. The area mean temperature over the whole sequence illustrates this temperature variation. At 2100 UTC 9 January the mean temperature is about 290 K, the temperature then decreases and at 0000 UTC 11 January, it reaches a mean value of 284 K; then it increases again and reaches a mean value of 290 K at 0000 UTC 16 January. This temperature increase helps the development of convective phenomena.
The above time sequence shows that there are two possible sources for warm and humid air arriving into the Mendoza region and creating deep moist convective events. One of them is the transport of moist air from the Amazon basin by the LLJ, and the other is from regions situated much more to the east of the continent at around 30°S. Between these two episodes there is an intermediate period where humidity is absent over the Mendoza area, the sky being cloud free. The timing of these events is controlled by the position of the eastward-propagating anticyclone with respect to the continent.
At 2100 UTC 20 January, a cloud structure develops at latitudes south of 35°S. The situation of 20 January is shown in Fig. 6. Satellite imagery shows cloud-top temperatures that reach values as low as 205 K. Using the vertical profile of temperature obtained from WRF for the same day and location (not shown), the top of the associated clouds is estimated to be at a level higher than 15 km indicating again the presence of deep convection. The wind field from ECMWF shows that the anticyclonic winds are spread over an extended region between east of the continent and the Atlantic coast. The LLJ is not present over the region, and the only source that could provide the humidity required for the generation of convection is the advection of moist air from the Atlantic Ocean. We use the FLEXTRA kinematic trajectory model to confirm this hypothesis and present a set of 3D 2-day back trajectories, which reach the convective areas (marked by crosses) at around 850 hPa on day 20 at 2100 UTC. These trajectories clearly show that 2 days before, the air parcels were located over the Atlantic Ocean. A few days later at 0000 UTC 24 January (Fig. 7), the wind and the ground pressure show that most of the air comes from the east.
We can add another case of moist air arriving from the east at 1800 UTC 30 January. To further document the generality of situations where the transport appears to be from the east, we have computed the humidity fluxes entering the Mendoza region for five selected cases. It is extremely difficult to automate such diagnostics: the detailed history of each anticyclone crossing South America is distinct, and identifying the latitude or longitude where significant transport of moisture is occurring requires a precise identification of the location of the prevailing winds. We could carefully position barriers across which it would be meaningful to compute the moisture flux from the north or the east. Whenever these situations correspond to the situations that are described in detail in the present paper, with a map of the circulation, we have shown the corresponding barriers with a dark line. In total we have found five clear cases of easterly transport. The corresponding results are shown in Table 1, which clearly shows that all fives situations are characterized both by moisture transport from the east and cloudy conditions over the Mendoza area.
5. Discussion and conclusions
As emphasized in the introduction, previous studies have shown that the LLJ is the main cause responsible for the transport of warm and moist air from the Amazon basin into central South America, and more specifically into the Mendoza region. This situation should favor the development of a regional forecast system over the Mendoza region, a system urgently needed to prevent or limit the risks associated with storms on cultivated land.
In this paper we found additional elements of complexity, which will have to be taken into account for forecast processes. In particular we show that it is possible for moist convection to develop from moist air conditions associated with transport from the southeast of Brazil and Uruguay or even directly from the Atlantic Ocean.
Although the situation is more complex than anticipated, it remains controlled by changes in the large-scale flow pattern. More specifically the timing for convection over the Mendoza region appears to depend very strongly on the position of the eastward-moving southern Pacific anticyclone with respect to the continent. These processes have not been described earlier. The crucial role of the anticyclone is probably one of the major conclusions from this test case study. More specifically, when the anticyclone is located in the Pacific Ocean, convection over the Mendoza region is driven by the LLJ. As the anticyclone penetrates into the continent, the LLJ is deviated toward the east, and at the usual LLJ location winds tend to become southerly. As a result, convection is suppressed and clear skies dominate over the region. When the western part of the anticyclone has reached the eastern side of the continent, convection over Mendoza can be reactivated, but this time the presence of LLJ is not required, and moist air is originating from different regions, located at the east of the continent or even over the Atlantic Ocean.
By suggesting a link between the timing of deep convection development over the Mendoza region and the flow at a larger scale, which can be predicted by mesoscale models, we have identified a key mechanism for forecast exercises. Setting up a system to analyze in real time the risk of heavy precipitations or hailstorm events, is probably beyond reach.
Acknowledgments
We are thankful to ECMWF for the free access to their archives. The WRF simulations were calculated with the computing resources of IDRIS (Institut du Developement et des Ressources en Informatique Scientifique). The SATMOS cloud-top temperature was provided by J. P. Olry from Météo-France.
REFERENCES
Berbery, E. H., and E. A. Collini, 2000: Springtime precipitation and water vapor flux over southeastern South America. Mon. Wea. Rev., 128 , 1328–1346.
Berbery, E. H., and V. Barros, 2002: The hydrological cycle of the La Plata basin in South America. J. Hydrometeor., 3 , 630–645.
Berri, G. J., and J. Inzunza, 1993: The effect of the low-level jet on the poleward water vapour transport in the central region of South America. Atmos. Environ., 27A , 335–341.
Douglas, M. W., M. Pena, and W. R. Villarpando, 2000: Special observations of the low-level flow over eastern Bolivia during the 1999 Atmospheric Mesoscale Campaign. Preprints. Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Santiago, Chile, Amer. Meteor. Soc., 157–158.
Jones, D. A., and I. Simmonds, 1994: A climatology of Southern Hemisphere anticyclones. Climate Dyn., 10 , 333–348.
Li, Z. X., and H. Le Treut, 1999: Transient behavior of the meridional moisture transport across South America and its relation to atmospheric circulation patterns. Geophys. Res. Lett., 26 , 1409–1412.
Liebmann, B., G. N. Kiladis, C. S. Vera, A. C. Saulo, and L. M. V. Carvalho, 2004: Subseasonal variations of rainfall in South America in the vicinity of the low-level jet east of the Andes and comparison to those in the South Atlantic convergence zone. J. Climate, 17 , 3829–3842.
Marengo, J. A., and W. Soares, 2002: Episodio de Jatosde baixos niveis durante 13 a 19 de Abril de 1999. Rev. Bras. Meteor., 17 , 35–52.
Marengo, J. A., M. W. Douglas, and P. L. Silva Diaz, 2002: The South American low level jet east of the Andes during the 1999 LBA-TRMM and LBA-WET AMC campaign. J. Geophys. Res., 107 .8079, doi:10.1029/2001JD001188.
Marengo, J. A., W. R. Soares, C. Saulo, and M. Nicolini, 2004: Climatology of the low-level jet east of the Andes as derived from the NCEP–NCAR reanalyses: Characteristics and temporal variability. J. Climate, 17 , 2261–2280.
Misra, V., P. Dirmeyer, B. Kirtman, H. Huang, and M. Kanamitsu, 2000: Regional simulation of interannual variability over South America. COLA Tech. Rep. 85, 42 pp.
Mo, K. C., and R. W. Higgins, 1996: Large-scale atmospheric moisture transport as evaluated in the NCEP/NCAR and the NASA/DAO reanalyses. J. Climate, 9 , 1531–1545.
Nicolini, M., and A. C. Saulo, 2000: Eta characterization of the 1997–1998 warm season Chaco jet cases. Preprints. Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Santiago, Chile, Amer. Meteor. Soc., 330–331.
Nogues-Paegle, J., and K. C. Mo, 1997: Alternating wet and dry condition over South America during summer. Mon. Wea. Rev., 125 , 279–291.
Paegle, J., 2000: American low-level jets in observation and theory: The ALLS project. Preprints. Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Santiago, Chile, Amer. Meteor. Soc., 161–162.
Salio, P., M. Nicolini, and C. Saulo, 2002: Chaco low-level jet events characterization during the austral summer season. J. Geophys. Res., 107 .4816, doi:10.1029/2001JD001315.
Saulo, A. C., and M. Nicolini, 2000: The atmospheric conditions preceding the occurrence of strong low level jets east of the Andes during January 1998. Preprints. Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Santiago, Chile, Amer. Meteor. Soc., 336–337.
Saulo, A. C., M. Nicolini, and S. C. Chou, 2000: Model characterization of the South American low-level flow during the 1997–98 spring-summer season. Climate Dyn., 16 , 867–881.
Sèze, G., and M. Desbois, 1987: Cloud cover analysis from satellite imagery using spatial and temporal characteristics of the data. J. Climate Appl. Meteor., 26 , 287–303.
Silva Dias, P. L., 2000: The role of latent heat release in the dynamics of the LLJ’s along the Andes. Preprints. Sixth Int. Conf. on Southern Hemisphere Meteorology and Oceanography, Santiago, Chile, Amer. Meteor. Soc., 163–164.
Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers, 2005: A description of the Advanced Research WRF Version 2. NCAR Tech. Note NCAR/TN-468+STR, 88 pp.
Stensrud, D., 1996: Importance of low-level jets to climate: A review. J. Climate, 9 , 1698–1711.
Stohl, A., and G. Wotawa, 1995: A method for computing single trajectories representing boundary layer transport. Atmos. Environ., 29 , 3235–3238.
Stohl, A., G. Wotawa, P. Seibert, and H. Kromp-Kolb, 1995: Interpolation errors in wind fields as a function of spatial and temporal resolution and their impact on different types of kinematic trajectories. J. Appl. Meteor., 34 , 2149–2165.
Vernekar, A., B. Kirtman, and M. Fennessy, 2003: Low-level jets and their effects on the South American summer climate as simulated by the NCEP Eta Model. J. Climate, 16 , 297–311.
Wang, H., and R. Fu, 2004: Influence of cross-Andes flow on the South American low-level jet. J. Climate, 17 , 1247–1262.
Illustration of the meteorological conditions at 2100 UTC 9 Jan 2001: (a) cloud-top temperature and (b) surface pressure and 850-hPa wind.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Results of the WRF Model forced with the conditions at 2100 UTC 9 Jan 2001: (a) skew Tρ–logp diagram at 34°S, 64°W, which provides a measure of the CAPE; and (b) longitude–altitude cross section of the relative humidity at 34°S.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Illustration of the meteorological conditions at 0000 UTC 3 Jan 2001: (a) cloud-top temperature, (b) surface pressure and 850-hPa wind, and (c) longitude–altitude cross section of the relative humidity at 35.5°S.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Illustration of the meteorological conditions at 0000 UTC 11 Jan 2001: (a) cloud-top temperature, (b) surface pressure and 850-hPa wind, and (c) cloud imagery from geostationary satellite GOES-8.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Illustration of the meteorological conditions at 0300 UTC 16 Jan 2001: (a) cloud-top temperature, (b) surface pressure and 850-hPa wind, and (c) aerial distribution of the CAPE as computed from the WRF Model.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Illustration of the meteorological conditions at 2100 UTC 20 Jan 2001: (a) cloud-top temperature, (b) surface pressure and 850-hPa wind, and (c) back trajectories from FLEXTRA.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Illustration of the meteorological conditions at 0000 UTC 24 Jan 2001: surface pressure and 850-hPa wind.
Citation: Monthly Weather Review 136, 6; 10.1175/2007MWR2088.1
Water vapor flux (kg m−2 s−1) across barriers that are 10° wide and extend from 1000 to 700 hPa. The flux from the east is computed across a barrier at a fixed longitude, whose position in latitude is determined from the observation of the wind at 850 hPa. The flux from the north is computed though a barrier at a fixed latitude, whose position in longitude is determined from the observation of the wind at 850 hPa. Values are positive for fluxes from the north or the east and, thus, for fluxes entering the area of northern Argentina. The total cloud cover over an area of 30°–40°S, 60°–70°W is added for reference.