Michael Douglas

National Severe Storms Laboratory

Norman, Oklahoma


Matilde Nicolini and Celeste Saulo

Departamento de Ciencias de la Atmósfera (UBA)

Centro de Investigaciones del Mar y la Atmósfera (UBA/CONICET)

Universidad de Buenos Aires

Buenos Aires, Argentina




Routine global analyses suggest the existence of a lower-tropospheric southward current from the Amazon Basin that extends into the Argentine plains, east of the Andes (see Fig. 1). These analyses also indicate that the strongest winds are between 15ûS and 20ûS, near the eastern flank of the Andes. This would, in general terms, suggest that the flow is similar to the much more extensively studied low-level jet (LLJ) seen in many daily analyses over the south central part of North America during the warm season (Bonner, 1968). Despite the depiction of a maximum wind speed in the global analyses over central South America, in-situ observations have been scarce, due to the sparse sounding network in this region, and to the lack of field programs to investigate the existence of a low-level jet.


As part of the 1997-8 Pan American Climate Studies (PACS) Sounding Network (SONET) project, special observations were made at Santa Cruz, Bolivia to determine the intensity and diurnal variability of a LLJ east of the Andes. These twice-daily observations, made over a two month period (January - early March 1998), have clearly shown the existence of a strong jet in the region on many days. In this paper we describe the results of a short field campaign to explore the low-level flow at Santa Cruz, Bolivia. We then compare operationally-run mesoscale model output with the Santa Cruz observations.


FIG 1. The NCEP long-term mean 850 mb wind vectors for the months of June-August (North America, left figure) and December-February for South America (right figure above). Solid dots are radiosonde stations(historical) nearest to the jet core, open circle is Santa Cruz, Bolivia - pilot balloon site in 1998.




During January - March 1998 special pilot balloon observations were made at Santa Cruz, Bolivia. Approximately 90 soundings were made over about a 70 day period. Ideally, the intent was to have soundings made in the early morning, and then in the late afternoon. In this manner the diurnal variation could be roughly captured. Although this plan was attempted, unusual cloudiness and frequent morning rainstorms prevented obtaining many morning soundings. The cloudiness also prevented tracking many of the balloons to high levels. Nonetheless, the data set is unique, and it shows unambiguously that there exists a strong LLJ in the region.

The flat observing site, at ~410 m above sea level, was about 25 km to the northeast of closest Andean foothills. In a general sense the low-level flow was usually parallel to the orientation of the topography in the first several km above ground level. A site farther to the east might have been preferable, but was not logistically feasible.

2.1 The mean sounding

The mean wind speed and direction profile, based on all observations, is shown in Fig. 2. Most apparent from the profiles is a northerly wind maximum between


FIG 2. Plot of mean wind direction and wind speed as function of height at Santa Cruz, based on all pilot balloon observations made during Jan-Mar 1998.


1 and 2 km above the surface. The winds approach calm near 5 km, where the direction changes from northwesterly flow through northerly to southeast flow. There is reasonable agreement between Fig. 2 and the NCEP analyses at 850 mb (Fig. 1) in wind speed at the grid point nearest Santa Cruz. The profiles of meridional and zonal wind (Fig. 3) reveal that at 5 km both components nearly vanish; this is about 1 km above the level of the Andean high terrain (altiplano) to the west and southwest.


2.2 Diurnal variations

The diurnal variation of the speed and direction profiles are slightly easier to interpret than those of the components and are shown in Fig. 4. The relatively


FIG 3. Profiles of the meridional wind, zonal wind, and wind speed of the mean Santa Cruz sounding.



small number of observations made during the morning suggests that we cannot put high confidence in the results. However, it appears that there is a major difference between the two profiles. The morning profile (made about 30 min after sunrise) shows weak (~2-3 ms-1) winds to about 2.5 km. These winds are from a westerly to almost southerly direction, from the direction of higher terrain. The afternoon sounding shows a strong northwesterly LLJ, with maximum speeds near 10 ms-1. Above 2.5 km both profiles are similar, and the directions are both from the northwest.


2.3 Temporal evolution

A time series of the meridional wind at, or close to the 1 km level, composed of observations made at all observing times, shows the large variability of the lower tropospheric flow at Santa Cruz during the observing program (Fig. 5). There are clearly fluctuations of a synoptic time scale in the record, with around 9-10 peaks in either the southerly or northerly winds. This gives an average periodicity of near 7 days. It is clear that, although northerly jets are more common (in fact northerly jets were approximately 3 times as common as southerly jets), there are strong southerly wind events even at this relatively low-latitude during the height of the warm season. Examination of the wind profiles associated with these events (not shown) indicates that the flow is somewhat shallower than the northwesterly flow regime.


Fig. 4. Mean wind speed and direction profiles for both morning and afternoon soundings at Santa Cruz.



Fig. 5. Time series of v-component of wind at or near the 1 km level at Santa Cruz.




In this section we discuss the structure of the LLJ in central South America from output that was available from the Brazilian Center for Weather Forecasts and Climate Studies (Centro de Previsão de Tempo e Estudos Climáticos, CPTEC). The Eta model (Mesinger et al. 1988; Black 1994) is used at CPTEC to produce short-range weather forecasts. The domain of the model covers most of South America and part of the Atlantic and Pacific Oceans. The resolution used at CPTEC is 40 km, with 38 layers. The operational outputs were available for this study every 6 hours during the 97-98 warm season.

3.1 Mean meridional wind component

The mean diurnal variation of the v-wind (Fig. 6a) shows a maximum at 0000 UTC and a minimum around 1800 UTC. Although northerly winds are present up to 500 mb, the highest speeds are located near 850 mb, with maximum values just over 10 ms-1. This intensity is also apparent in the core immediately to the east of the Andes (Fig. 6b). Figure 6b also shows the spatial dimensions of this core, which extends vertically to 650 mb, and horizontally about 300 km in the cross wind direction. A strong west-east gradient in the v-wind is apparent on the western side of the jet core, associated with the steep slope of the Andes.

3.2 Model comparisons

Output from the CPTEC Eta model have been compared with the pilot balloon data from Santa Cruz for 1998. Bonner's (1968) criterion 1 was applied to both data sets for the period January 6 to February 28. This criterion specifies that the wind speed profile have a maximum of at least 12 ms-1 within 1.5 km of the ground, and that the wind speed decrease by at least 6 ms-1 at the lowest minimum located above the maximum, at or below the 3-km level. This criterion was


Fig. 6. a) time-height section of the mean (Jan-Feb 98) v-wind at the grid point nearest to Santa Cruz. From Eta output, corresponding to forecast cycle beginning at 1200 UTC. b) vertical E-W section of the mean (Jan-Feb) v-wind at 17.8ûS and 0000 UTC from Eta output. Vertical axis is pressure in mb. Contours in ms-1.


applied each day at every forecast time for the entire warm season. Only the afternoon soundings were included as only one morning sounding verified this criterion. Twelve events were identified as LLJ's from the pilot balloon sample. The mean wind profile for this subset of soundings is shown in Fig. 7a. These same 12 cases were extracted from the Eta model 0000 UTC output at the nearest grid-point to Santa Cruz to construct a mean LLJ sounding (Fig. 7b).

From the afternoon pilot balloon sample to which Bonner's criterion was applied (20 observations during January and 12 during February) more than 30% of the soundings corresponded to LLJ cases. This compared with a LLJ frequency diagnosed from the Eta model output at the nearest grid-point to Santa Cruz of around 80% for the complete period of January-February. Thus the model indicates a significantly higher frequency of LLJ's. However, a daily comparison of the 32 available soundings is in good agreement with Eta profiles in terms of wind direction (north or south component) and LLJ occurrence. The frequency difference can be explained in part by the difference in temporal completeness of the two data sets - many balloons were not tracked high enough to apply the LLJ criterion.

Fig. 7a. Observed mean wind speed (ms-1) profile during LLJ events (around 1800 local time) at Santa Cruz. Heights are in meters above ground level (AGL).


Fig. 7b. Eta model mean wind speed profile during LLJ events (at 0000 UTC) at Santa Cruz. Vertical coordinate is pressure in mb.

Compositing the pilot balloon soundings that satisfy Bonner's criterion 1 clearly depicts the strength of these LLJ's (Fig. 7a). The mean maximum wind speed is greater than 19 ms-1 near 1.7 km AGL, with a vertical shear above the maximum of ~ 5 ms-1 per km. This is in good agreement with the composite of the Eta model soundings (Fig. 7b).


The results of this study have encouraged us to use the output of the CPTEC Eta model to characterize the structure and diurnal variability of the central South American LLJ. Although inadequate observations still exist to compare the diurnal variability of the LLJ with the model simulations, the Eta output should be valuable in helping to design future field observation programs in the region that might be able to better provide the required observations. The model output should also help to determine what spatial and temporal resolution are needed for observations if we are to adequately describe the fluctuations of the LLJ on both weather and climate time scales.


One of the authors (MD) wishes to thank the NOAA Office of Global Programs for supporting the PACS-SONET observation program, one component of which was operating the Santa Cruz, Bolivia sounding site. The PACS-SONET project was supported by many individuals in NOAA and especially at NSSL, where innumerable shipping and financial transactions were handled efficiently. Rosario Douglas assisted in handling at-times complicated logistical arrangements in Santa Cruz and elsewhere during PACS-SONET. Malaquias Peña manages the PACS-SONET Homepage helped in many aspects of the work. A more complete acknowledgments, together with the Santa Cruz data and other material on PACS-SONET can be found on the PACS-SONET homepage which can be located at:

The UBA co-authors thank CPTEC for the use of the ETA model products and the assistance given to Celeste Saulo during her stay in that Center. Their work has been partially funded by Projects TX30 of the University of Buenos Aires, and PICT 01757 of the Agencia Nacional de Promoción Científica y Tecnológica.


Berri, G. and J. C. Inzunza B., 1993: The effect of the

low-level jet on the poleward water vapour transport in the central region of South America. Atmos. Env. Part A, 27A(3), 335-341.

Black T.L., 1994: NMC Notes: The New NMC

mesoscale Eta model: Description and forecast examples. Weather and Forecasting, 9, 256-278.

Blackadar, A. K., 1957: Boundary layer wind maxima

and their significance for the growth of nocturnal inversions. Bull. Amer. Meteor. Soc., 38, 283-290.

Bonner, W. D., 1968: Climatology of the low level jet.

Mon. Wea. Rev., 96, 833-850.

Bonner, W. D., and J. Paegle, 1970: Diurnal variations

in boundary layer winds over thesouth-central United States in summer.Mon. Wea. Rev., 98, 735-744.

Holton, J. R., 1967: The diurnal boundary layer wind

oscillation above sloping terrain. Tellus, 19, 199-205.

Mesinger, F., Z. I. Janjic, S. Nickovic, D. Gavrilov and

D.G. Deaven, 1988: The step-mountain coordinate:

Model description and performance for cases of

Alpine lee cyclogenesis and for a case of Appalachian

redevelopment. Mon. Wea. Rev., 116, 1493-1518.

1 Corresponding author address: 1313 Halley Circle, Norman,

OK 73069 email: