1. Introduction
Low-level jets (LLJs) occur throughout the tropics and extratropics causing horizontal advection of moisture and temperature and an increase in low-level wind shear and convergence, which can help support deep convection and the development of mesoscale convective systems and complexes (e.g., Velasco and Fritsch 1987; Monaghan et al. 2010). The physical mechanisms that promote these low-level wind speed maxima can be fairly diverse in accordance with each locality (Stensrud 1996), but the inertial oscillation due the diurnal variation of eddy viscosity (Blackadar 1957) and the terrain effects that produce slope and valley winds or forced wind deflection (e.g., Holton 1967; Paegle et al. 1984) are among the main dynamical mechanisms associated with LLJs.
The most prominent LLJ in South America is the South America LLJ (SALLJ). Virji (1981) was the first to observe the northerly SALLJ along the east side of the Andes Mountains. Marengo et al. (2004) and Vera et al. (2006), among others, have since described this jet in more detail. Its nocturnal intensification and humidity transport from the Amazon to the La Plata basin supports the development of some of the most severe thunderstorms in the world over Argentina and other parts of southeast South America (e.g., Zipser et al. 2006).
Studies from individual locations have hinted at a zonally oriented low-level jet over the Amazon rainforest. Greco et al. (1992) showed a nocturnal LLJ over Manaus in the central Amazon during the Amazon Boundary Layer Experiments (ABLE 2A and 2B) and Alcântara et al. (2011) analyzed a multiyear climatology of a LLJ over Belem, which is located near the coast. However, there is no current information in the literature specific to a more large-scale LLJ across the Amazon. The existence of a large-scale LLJ could have implications for convective cloud system development and evolution across the Amazon basin.
Alcântara et al. (2011) found that the LLJ at Belem was often associated with the formation of convective systems that would then propagate inland. We refer to these propagating convective systems as Amazonian coastal squall lines (ACSLs) after Garstang et al. (1994), who analyzed a subset of the ACSLs that occurred during ABLE 2B. Garreaud and Wallace (1997) more broadly described the southwestward march of convection across the Amazon and its diurnal variability based on nine years of infrared satellite imagery. They showed that the convective cloudiness maximized in two bands: one near the coast and the other 1600 km inland. They suggested that the coastal band was associated with the sea-breeze front (Kousky 1980), while the inland band was from the reactivation of the propagating coastal convection (i.e., the ACSLs) from the day before. Farther inland propagation of the band was argued to be related to enhancement of the prevailing low-level winds northwest of the mouth of the Amazon (i.e., the oceanic trade winds).
Burleyson et al. (2016) used 15 years of higher temporal and spatial resolution infrared satellite data to postulate that the sea-breeze front and associated ASCLs extend well into the Amazon basin, with a greater impact on the convection in the central Amazon near Manaus when the timing of the sea-breeze front arrival coincides with daytime heating of the land surface. However, from a theoretical and observational standpoint, sea-breeze fronts typically only extend a few hundred kilometers inland from the northern and northeastern Brazilian coast (Dalu and Pielke 1989; Planchon et al. 2006; Souza and Oyama 2017). ACSLs that are triggered by the sea-breeze front occasionally cross the Amazon basin but most decay within a few hundred kilometers of the coast and very few move far into the Amazon basin (Cohen 1989), leaving a minimum in rain and cloudiness in between the coast and central Amazon during the transition and rainy seasons (e.g., Garreaud and Wallace 1997).
The original impetus of this research was to understand the relationship between organized convection and the large-scale environment during the Observations and Modeling of the Green Ocean Amazon 2014–15 (GoAmazon2014/5) field campaign (Martin et al. 2016). GoAmazon2014/5 was a Department of Energy (DOE) sponsored deployment that took place in the central Amazon from 2014 to 2015 to study aerosols, clouds, and precipitation and their interactions over a pristine rainforest occasionally impacted by pollution from the large city of Manaus. In the process, we discovered the existence of a large-scale nocturnal Amazonian LLJ (ALLJ) that appears distinct from the northeasterly trade winds and sea-breeze front. We evaluate the ALLJ areal extent during the diurnal cycle together with cloud cluster density and moisture fluxes. The ALLJ structure is also examined during days that were considered having high or weak convective activity.
2. Methodology
a. MAM winds over the Amazon
The convection in the Amazon is highly active during March–April–May (MAM), with precipitation and the number and inland extent of ACSLs maximizing in this season (Cohen et al. 1995; Alcântara et al. 2011). We also assume that two years is a long enough period to characterize a phenomenon with a regular diurnal cycle. Therefore, this study focuses on analyzing the ALLJ and its connection to convection during MAM of 2014 and 2015, the two years of the GoAmazon2014/5 deployment. Burleyson et al. (2016) showed that the diurnal cycle of cold cloudiness during 2014 was typical of their 15-yr satellite climatology and that the austral fall diurnal cycle over the central Amazon was consistent from year-to-year, further bolstering the use of our 2-yr dataset. We should note that 2015 was an El Niño year and the Amazon was drier than normal. While some variability in diurnal characteristics may occur because of this interannual forcing, the nocturnal ALLJ signatures are very clear in both years.
Horizontal winds were analyzed from two reanalysis datasets: the European Centre for Medium-Range Weather Forecasts interim reanalysis (ERA-I) with a resolution of 0.75° (latitude) × 0.75° (longitude) × 37 pressure levels every 6 h (Dee et al. 2011), and the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), with a resolution of 0.500° (latitude) × 0.625° (longitude) × 42 pressure levels every 3 h (Gelaro et al. 2017). Figure 1a shows the MERRA-2 900 hPa winds over northern South America for MAM during 2014–15. Strong (~10 m s−1) northeasterly trade winds exist north of the equator with weaker (~7 m s−1) easterly and southeasterly trade winds south of the equator. Moderately strong (5–10 m s−1) northeasterlies are present across much of the Amazon basin. Winds at 200 hPa (Fig. 1b) show weak (~5 m s−1) southerly flow across the Amazon, transitioning to strong (~20 m s−1) westerlies over the intertropical convergence zone (ITCZ). A cross section was drawn in Fig. 1 to represent flow along the main core of the low-level jet from the coast through T3, the location of the DOE Atmospheric Radiation Measurement (ARM) mobile facility and sounding site during GoAmazon2014/5.
MAM horizontal winds (m s−1) from MERRA-2 at (a) 900 and (b) 200 hPa for 2014–15. Line A is used to demarcate the vertical cross section of the horizontal wind along the ALLJ. T3 is the location of the DOE ARM mobile facility in Manacapuru during GoAmazon2014/5.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
MAM horizontal winds (m s−1) from MERRA-2 at (a) 900 and (b) 200 hPa for 2014–15. Line A is used to demarcate the vertical cross section of the horizontal wind along the ALLJ. T3 is the location of the DOE ARM mobile facility in Manacapuru during GoAmazon2014/5.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
MAM horizontal winds (m s−1) from MERRA-2 at (a) 900 and (b) 200 hPa for 2014–15. Line A is used to demarcate the vertical cross section of the horizontal wind along the ALLJ. T3 is the location of the DOE ARM mobile facility in Manacapuru during GoAmazon2014/5.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
b. Convective cell tracking
The convective system tracking was developed by automatic overlapping of successive images (Woodley et al. 1980; Williams and Houze 1987; Chen et al. 1996; Machado et al. 1998; Mathon and Laurent 2001) using the Geostationary Operational Environmental Satellite 13 (GOES-13) infrared imagery (channel 4, 10.2–11.2 μm). The images have a temporal resolution of 30 min and a horizontal resolution of 4 km × 4 km at nadir. Cloud clusters were defined as areas greater than or equal to 2500 km2 and brightness temperatures colder than or equal to 235 K (Vila et al. 2008). The trajectory area is the total area covered by the tracked convective cloud clusters over their life cycles. With respect to time gaps in the GOES-13 satellite imagery, the tracking algorithm allows two missing images before starting new systems.
Cell tracking examples from two days during the study period are shown in Fig. 2. The top panel contains a satellite snapshot from each day and the colored areas represent cloud clusters. The bottom panel shows the diurnal evolution of each day’s cloud systems across the Amazon from 15°S to 13°N. The area of the tracked cloud clusters are represented by ellipses. Figure 2 indicates apparent differences in the overall convective activity and diurnal cycle of clusters between the two days; these differences will be discussed in more detail in section 3f.
GOES-13 infrared images from (a) 0030 UTC 2 Apr and (c) 2030 UTC 11 Apr 2014 over the Amazon. Tracked cloud cluster IR temperatures (in K) are indicated by color. (b),(d) Time–longitude diagrams from 15°S to 13°N of cloud clusters tracked on the days shown in (a) and (c). The horizontal dashed lines in (b),(d) indicate the time of the GOES-13 image. The size of the tracked cloud clusters is shown by the ellipses. The time 1500 UTC represents approximately midday across the study domain.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
GOES-13 infrared images from (a) 0030 UTC 2 Apr and (c) 2030 UTC 11 Apr 2014 over the Amazon. Tracked cloud cluster IR temperatures (in K) are indicated by color. (b),(d) Time–longitude diagrams from 15°S to 13°N of cloud clusters tracked on the days shown in (a) and (c). The horizontal dashed lines in (b),(d) indicate the time of the GOES-13 image. The size of the tracked cloud clusters is shown by the ellipses. The time 1500 UTC represents approximately midday across the study domain.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
GOES-13 infrared images from (a) 0030 UTC 2 Apr and (c) 2030 UTC 11 Apr 2014 over the Amazon. Tracked cloud cluster IR temperatures (in K) are indicated by color. (b),(d) Time–longitude diagrams from 15°S to 13°N of cloud clusters tracked on the days shown in (a) and (c). The horizontal dashed lines in (b),(d) indicate the time of the GOES-13 image. The size of the tracked cloud clusters is shown by the ellipses. The time 1500 UTC represents approximately midday across the study domain.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Figure 2 is in UTC, however, the Amazon encompasses three time zones with local time in the Amazon basin being UTC − 4 in central Amazonia and local time in the eastern Amazon and along the Brazilian coast being UTC − 3. There is an additional time zone over the western Amazon (UTC − 5), but our cross section from Fig. 1 does not extend into this time zone. For the general purposes of this study we will assume the central Amazon UTC time conversion, but the reader should keep in mind that there is a 1-h difference over the eastern portion of the domain. However, this should not affect our general results since we are only showing 3-h averages in our analysis.
Days of highly active convection were separated from days of weakly active convection using the convective system tracking statistics. We created cumulative distributions of convective system trajectory area and lifetime using the tracking database for all of 2014 and 2015 over the Amazon basin (a total of 72 649 systems) to determine values for the largest and longest-lasting convective systems. The 90th percentile trajectory area and lifetime based on the cumulative distributions were 81 158 km2 and 6.5 h, respectively. Any system that achieved both of these attributes was considered extreme. Highly active days were defined as days with the highest number of extreme systems over the Amazon. Conversely, weakly active days were defined as having the fewest number of extreme systems.
The days classified as highly and weakly active are listed in Table 1, as well as the number of extreme convective systems tracked each day (ne). We only considered days that had at most three successive missing GOES-13 images (i.e., one and a half hours), thus excluding 60 days from any analysis concerning convectively active days. We then used the top and bottom 15% of the remaining 124 days in MAM of 2014–15 to represent highly and weakly active convective conditions. The top 18 days had 19–23 extreme systems per day with 10 days from 2014 and 8 days from 2015. The bottom 18 days had 8–12 extreme systems per day, also approximately half from each year. An interesting point is that there are no days without very large and long-lived convective systems over the Amazon during MAM. Reanalysis winds were evaluated during the highly and weakly active days to determine if there was any relationship between the degree of convective activity and the ALLJ. Itterly et al. (2018) performed an analogous separation using a satellite-based daily maximum rain rate threshold of 0.5 mm h−1 to compare diurnal changes in the large-scale circulation on convective and nonconvective days over the Amazon in reanalysis data from 2002 to 2016.
Days classified as highly active and weakly active with the respective number ne of extreme (i.e., trajectory > 81 158 km2 and lifetimes ≥ 6.5 h) convective systems tracked.
3. Results
a. Vertical structure of MAM winds along the cross section
Figure 3 shows the vertical cross section of the daily average MAM horizontal wind speed projected along cross section A in Fig. 1 (i.e., from the coast into the central Amazon) for MERRA-2 and ERA-I during 2014–15. Winds increase up to 800 hPa along the cross section in both reanalyses, with maximum wind speeds reaching 8 m s−1. Each tick mark on the x axis is about 67 km, so the low-level wind maximum extends more than a thousand kilometers into the Amazon basin, including T3, which is approximately 1300 km from the coast. The MERRA-2 and ERA-I cross sections exhibit only minor differences between each other. Itterly et al. (2018) also found that the vertical structure of the horizontal winds of MERRA-2 and ERA-I, including diurnal variations, were comparable over the Amazon. Thus, we will use MERRA-2 instead of ERA-I in the rest of the analysis because MERRA-2 has 3-h temporal resolution compared to the 6-h resolution of ERA-I.
Mean MAM horizontal wind speeds along cross section A from Fig. 1 with (a) MERRA-2 and (b) ERA-I reanalysis data during 2014 and 2015. Contour lines are plotted every 0.5 m s−1 and the colored lines indicate positive values (i.e., flow toward the Andes). Negative wind values (or flow toward the Atlantic Ocean) are plotted in black.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM horizontal wind speeds along cross section A from Fig. 1 with (a) MERRA-2 and (b) ERA-I reanalysis data during 2014 and 2015. Contour lines are plotted every 0.5 m s−1 and the colored lines indicate positive values (i.e., flow toward the Andes). Negative wind values (or flow toward the Atlantic Ocean) are plotted in black.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM horizontal wind speeds along cross section A from Fig. 1 with (a) MERRA-2 and (b) ERA-I reanalysis data during 2014 and 2015. Contour lines are plotted every 0.5 m s−1 and the colored lines indicate positive values (i.e., flow toward the Andes). Negative wind values (or flow toward the Atlantic Ocean) are plotted in black.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
b. Diurnal low-level wind variations
While it could be argued that the enhanced low-level winds in Fig. 3 are simply the extension of the oceanic trade winds into the Amazon basin, a diurnal decomposition of the reanalysis wind fields provides a more nuanced view. We build on the results of Itterly et al. (2018) and others to show that there is considerable diurnal variability in low-level Amazonian winds produced by the complex interaction between the trade winds and boundary layer evolution. Figures 4 and 5 present the diurnal cycle of flow and its anomaly from the daily average along cross section A in MERRA-2 for MAM.
As in Fig. 3, but images are mean flow every 3 h along cross section A for MERRA-2.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 3, but images are mean flow every 3 h along cross section A for MERRA-2.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 3, but images are mean flow every 3 h along cross section A for MERRA-2.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 4, but representing the anomalous diurnal flow from the MAM daily mean wind (m s−1). Solid lines are positive wind anomalies and dashed lines are negative wind anomalies.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 4, but representing the anomalous diurnal flow from the MAM daily mean wind (m s−1). Solid lines are positive wind anomalies and dashed lines are negative wind anomalies.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 4, but representing the anomalous diurnal flow from the MAM daily mean wind (m s−1). Solid lines are positive wind anomalies and dashed lines are negative wind anomalies.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Figure 4 shows that the low-level winds undergo a strong diurnal oscillation over the Amazon basin along cross section A. There is a dynamic evolution in the wind field below 900 hPa that starts near the coast in the early evening at 1700 LT. The enhanced low-level wind appears to propagate inland and reaches its maximum extension in the central Amazon at 0800 LT with wind speeds of 8 to 10 m s−1. After 0800 LT, the low-level winds weaken rapidly, especially near the coast.
The anomalous diurnal winds in Fig. 5 show that the low-level increase near the coast becomes especially evident at 2000 LT. This increase in wind speed is consistent with the decrease of inland eddy viscosity due to stabilization of the planetary boundary layer near sunset (Blackadar 1957), which allows the trade winds to penetrate farther inland because of the reduced friction. At 2300 LT, the coastal wind anomaly has penetrated farther inland and strengthened to values over 2 m s−1. The rest of the Amazon basin shows only positive anomalies along the cross section (i.e., enhanced northeasterlies) with a small maximum centered inland at 57°W or about 850 km from the coast (each large tick mark on Fig. 5 represents approximately a 400 km distance). After 2300 LT, the stabilization of the planetary boundary layer in the central Amazon is most pronounced (Carneiro and Fisch 2020) and at 0200 LT, the low-level wind anomaly that began at the coast appears to combine with the inland anomaly based on the 1 m s−1 contour. There is an elevation decrease from ~500 to 50 m between 53.5° and 56.6°W that may contribute to the apparent propagation of the 1 m s−1 contour wind anomaly at this time. By 0500 LT, the 1 m s−1 contour line has reached almost 60°W and maximum wind anomalies exceed 3 m s−1 400–800 km from the coast. At 0800 LT, the low-level wind anomalies are at their strongest and extend farthest inland with the 1 m s−1 contour past 63°W. A remnant of enhanced low-level flow toward the Andes persists in the far west Amazon until 1100 LT, but the reestablishment of the daytime planetary boundary layer rapidly weakens the low-level northeasterlies between 1100 and 1400 LT as evidenced by the large negative wind anomalies along the cross section.
The positive low-level wind anomaly represented by the 1 m s−1 contour line in Fig. 5 travels along the cross section from the coast to 63°W, a distance of about 1600 km, from 1700 to 0800 LT. This is an apparent propagation speed of 29 m s−1, which is much faster than the mean flow in Fig. 4 and the diurnal propagation speed of the IR cold cloudiness observed by Garreaud and Wallace (1997) and Burleyson et al. (2016). The 2 m s−1 contour line travels 1200 km in 12 h (i.e., from 53° to 59°W between 2300 and 0800 LT), which is an apparent propagation speed of 25 m s−1, consistent with the 1 m s−1 contour line propagation speed. This apparent propagation could simply be a result of the longer time it takes the weaker winds in the central Amazon to reach a 1 m s−1 anomaly (or stronger winds closer to the coast to reach a 2 m s−1 anomaly) after nocturnal decoupling from the surface. It has also been shown that the diurnal oscillation of the boundary layer at low latitudes can be a source of mesoscale internal gravity waves that can propagate rapidly for hundreds of kilometers (Orlanski 1973; Sun and Orlanski 1981), which may be an additional explanation for the observed propagation of the low-level wind anomaly. Regardless of the mechanism, the maximum inland extent of the low-level wind anomaly is much farther inland than one typically observes sea breezes and is farther than a sea breeze would travel when friction is considered based on theoretical expectations near the equator (Rotunno 1983; Dalu and Pielke 1989). Based on the above analysis, we argue that there exists a nocturnal Amazonian low-level jet (ALLJ) that is distinct from the northeasterly trade winds because of its fast propagation speed and from the coastal sea breeze because of its large spatial extent inland.
c. Sounding comparisons over the GoAmazon2014/5 T3 site
To show that the low-level wind features observed by reanalysis are reliable over the central Amazon, we compare the GoAmazon2014/5 T3 sounding wind profiles to the MERRA2 wind profiles over the T3 site (Fig. 6). T3 was located at 3°S, 60.6°W in Manacapuru, about 65 km downwind of Manaus. While the twice per day operational soundings at Manaus (0800 and 2000 LT) were regularly reporting during 2014–15 and were likely assimilated into the MERRA-2 system, the four to five per day GoAmazon2014/5 soundings at T3 were not assimilated by MERRA-2 (M. Bosilovich 2019, personal communication).
Mean wind speed profiles along cross section A over the T3 site from (a) DOE ARM soundings and (b) MERRA-2 during MAM of 2014 and 2015. (c),(d) The DOE ARM sounding and MERRA-2 winds over T3 from 9 Mar 2014.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean wind speed profiles along cross section A over the T3 site from (a) DOE ARM soundings and (b) MERRA-2 during MAM of 2014 and 2015. (c),(d) The DOE ARM sounding and MERRA-2 winds over T3 from 9 Mar 2014.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean wind speed profiles along cross section A over the T3 site from (a) DOE ARM soundings and (b) MERRA-2 during MAM of 2014 and 2015. (c),(d) The DOE ARM sounding and MERRA-2 winds over T3 from 9 Mar 2014.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Figures 6a and 6b show the mean MAM diurnal wind profiles during 2014 and 2015 for the soundings and MERRA-2, respectively. Low-level winds are weakest over Manacapuru in the afternoon and early evening (1400–2000 LT) and increase by about 2 m s−1 below 900 hPa in the late evening and early morning (0200–0800 LT) in both datasets. The mean low-level wind maximizes at 0800 LT with a wind speed between 7 and 8 m s−1.
The diurnal evolution of the ALLJ is more evident when considering an individual day. Figures 6c and 6d show the radiosonde and MERRA2 wind profiles over T3 for 9 March 2014. There was an additional 1100 LT sounding taken during the first intensive operation period (IOP-1) of the field campaign that is shown as a dashed gray line. Maximum winds of 11 and 13 m s−1 were observed at 950 hPa by the T3 soundings between 0200 and 0800 LT. MERRA-2 also showed maximum 950-hPa wind speeds at these times, although the magnitudes were slightly weaker (8 and 11 m s−1). The jet appeared to rise over the next 3 h as evidenced by the 1100 LT profiles in both datasets, collapsing at low levels by 1400 LT. Other days show strong diurnal jet evolution that is reasonably captured by MERRA-2, but it is clear that the ALLJ is not present every day. While the focus of this study is not on the day-to-day variability of the ALLJ, the next section provides an objective definition of the jet and quantifies how often it occurs during MAM.
d. Objectively identified jet
Bonner (1968) originally created objective criteria to quantify the occurrence of the Great Plains low-level jet (GPLLJ). We use their objective criteria with thresholds corresponding to the LLJ-0 (or “weaker jet”) category in Whiteman et al. (1997). The following criteria were applied to the 3-hourly MERRA-2 wind profiles between 1000 and 650 hPa to identify the occurrence of a low-level jet over the Amazon: (i) a maximum speed (υmax) ≥ 10 m s−1 (representing the nose of the jet), (ii) the difference between υmax and the minimum wind speed above the υmax level (Δυ) ≥ 5 m s−1, and (iii) υmax between 0° and 90° in azimuth.
Figure 7 shows the diurnal cycle of LLJ occurrence across the Amazon based on the above criteria. The contours are the probability of jet occurrence (%) each day at each hour. The vectors are the median of the probability density function of υmax (intensity and direction) in each grid point. Only grid points with more than 10% jet occurrence are shown. The cloud cluster occurrence in relation to the jet will be discussed in the next section.
GOES-13 cloud cluster density (×10−3%, shading), ALLJ occurrence frequency (%, contours), and the 50th percentile of ALLJ υmax (arrows) based on MERRA-2 winds every 3 h during MAM 2014 and 2015. Cross section line A is shown at 1700 LT.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
GOES-13 cloud cluster density (×10−3%, shading), ALLJ occurrence frequency (%, contours), and the 50th percentile of ALLJ υmax (arrows) based on MERRA-2 winds every 3 h during MAM 2014 and 2015. Cross section line A is shown at 1700 LT.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
GOES-13 cloud cluster density (×10−3%, shading), ALLJ occurrence frequency (%, contours), and the 50th percentile of ALLJ υmax (arrows) based on MERRA-2 winds every 3 h during MAM 2014 and 2015. Cross section line A is shown at 1700 LT.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
While there are two distinct nocturnal LLJs over Venezuela and Guyana that reach peak occurrences between 40% and 70% in the late evening/early morning hours, this study is focused on the somewhat broader area of LLJ occurrence across the central Amazon. The Venezuelan and Guyana jets are narrower because they occur between regions of elevated topography. The Amazonian jet experiences only modest elevation along its progression and thus goes over the terrain rather than being channeled by it.
Consistent with the diurnal timing of the mean MAM winds along cross section A in Figs. 4 and 5, there is no evidence of a LLJ over the Amazon between 1400 and 1700 LT. At 2000 LT, the 10% occurrence line has moved inland a few hundred km along the coast, possibly linked to the sea breeze. By 2300 LT, the 10% occurrence contour is well inland, with possible influence from downslope processes, and spread across much of the eastern Amazon. The jet reaches the central Amazon by 0200 LT and persists there through 1100 LT. We note that even though the ALLJ is present only about a quarter of the days during MAM, it still significantly impacts the mean MAM wind climatology (cf. Fig. 3). In addition, the oceanic trade winds are identified as a low-level jet feature based on our criteria, but are present throughout the diurnal cycle. It is possible that the establishment of the nocturnal boundary layer over land allows the trade winds to penetrate inland more easily due to the reduction of drag.
The probability of ALLJ occurrence is quite similar to the probability of SALLJ occurrence obtained in Salio et al. (2007) where υmax and Δυ was 12 and 6 m s−1, respectively, in accordance with Bonner (1968). In this paper υmax and Δυ is 10 and 5 m s−1 because there is no major mountain range that channels the wind as with the GPLLJ and the SALLJ. Rife et al. (2010) have more recently created a nocturnal LLJ index that they applied globally. Their results confirm the possibility of a nocturnal LLJ along the Amazonian coast in austral summer, but they did not further analyze this region.
e. Diurnal cloud cluster and moisture flux variations
In addition to depicting the objectively identified jet, Fig. 7 shows the geographic density of the GOES-13 cloud cluster area. Cloud cluster occurrence corresponds to the fraction of the 255 113 cloud clusters tracked during MAM of 2014 and 2015. The diurnal variation in cloud cluster density across northern South America is complex, but appears to be linked to both local and large-scale wind and humidity features.
Before discussing cloud cluster occurrence over land, we note a nocturnal offshore propagation of cloud clusters from the north and northeast coast of Brazil starting at 2300 LT that persists offshore until 1100 LT (Fig. 7). This pattern is consistent with Negri et al. (1994) and is likely associated with the local land breeze. A similar nocturnal offshore propagation can be seen off the coast of Venezuela, but the cloud clusters last longer and extend farther offshore, ostensibly because of the added forcing by topography (Mapes et al. 2003).
Figure 7 shows a robust inland propagation of cloud clusters from the northeast coast of Brazil (i.e., south of the Amazon River and not part of cross section A) starting at 1400 LT and ending around 0200 LT a few hundred kilometers from the coast. The initiation of convection at the coast in the afternoon and its inland propagation into the evening is typically ascribed to the local sea breeze (Garreaud and Wallace 1997; Rickenbach 2004; Burleyson et al. 2016). There is also inland propagation of cloud clusters from the north coast of Brazil (i.e., north of the Amazon River and encompassed by one end of cross section A). However, the cloud clusters do not propagate as far as the convection associated with the sea breeze on the northeast coast, potentially because of stronger topographic variations near the north coast of Brazil. Notably, the propagation of the clusters inland from the north and northeast coast of Brazil is not associated with the timing or spatial extent of the ALLJ contour.
There is a secondary enhancement in cloud cluster occurrence in Fig. 7 close to 1°S, 56°W (north and west of the coastal maximum) that begins around 2000 LT, maximizes at 0200 LT, and persists through 1100 LT. This cloud cluster enhancement is coincident with the development of the ALLJ across the Amazon as depicted by the 10% jet occurrence contour line. These cloud clusters travel farther inland than those presumably associated with the coastal sea breeze and faster than the prevailing winds (Anselmo et al. 2020, manuscript submitted to Int. J. Climatol.). Silva Dias and Ferreira (1992) ran a linear spectral model with varying basic state winds and forced by parameterized cumulus heating. Their results show that stronger variations in low-level winds produce more unstable modes that better mimic the propagation speeds of the cloud cluster speeds observed here and by Cohen (1989). More recently, Tulich and Kiladis (2012) showed that squall lines coupled to inertia-gravity waves can move faster than the background flow and that low-level shear assists in determining the westward motion. Thus, the ALLJ could be playing a role in the propagation of cloud clusters and convectively coupled inertia-gravity into the central Amazon in the evening hours.
During the day and well away from the coast, Fig. 7 shows large-scale enhanced cloud cluster occurrence in the central Amazon that is distinct from the coastal cloud clusters. Cloud cluster occurrence increases rapidly at 1400 LT, likely associated with daytime heating, and reaches a maximum at 1700 LT. These systems propagate westward with the prevailing winds and decay into the evening hours. Garreaud and Wallace (1997) defined a cross section very similar to cross section A in this study to evaluate the propagation of convective cloudiness across the Amazon using satellite IR data (e.g., see their Fig. 5). They observed diurnal cloud system evolution in agreement with what is found here and in particular, a break between the coastal convection and central Amazonian convection. In Burleyson et al. (2016), the propagation of cloud clusters by thousands of kilometers inland over the larger Amazon basin during MAM was understood as an effect of the sea breeze, however, sea-breeze fronts in northeastern South America typically only extend a few hundred kilometers inland (Planchon et al. 2006; Souza and Oyama 2017). There is also no indication of an objectively identified low-level wind feature during this time. Thus, we argue that these daytime central Amazonian cloud clusters are different from the late afternoon cloud clusters associated with the coastal sea breeze and the nocturnal clusters associated with the ALLJ.
However, the ALLJ may still play a role in central Amazon cloud cluster evolution during the day. Figure 8 shows the vertically integrated moisture flux (Q) from 925 to 600 hPa derived from MERRA-2 and the mean MAM winds between 950 and 700 hPa. While the low-level winds in this layer do not show a strong diurnal variation in direction or magnitude, the moisture flux from the Atlantic Ocean to the central Amazon exhibits a strong diurnal cycle. In particular, moisture flux values exceed 225 kg m−1 s−1 during the times of day associated with the development of the ALLJ and the spatial pattern of the moisture flux shows a strong similarity to the objectively defined jet occurrence indicated by the red contours.
As in Fig. 7, but showing the diurnal cycle of vertically integrated moisture flux (Q; kg m−1 s−1) between 925 and 600 hPa based on MERRA-2 data. Wind vectors represent the mean MAM wind between 950 and 700 hPa. The red contour lines indicate the ALLJ occurrence frequency at 10%, 30%, 50%, and 70%.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 7, but showing the diurnal cycle of vertically integrated moisture flux (Q; kg m−1 s−1) between 925 and 600 hPa based on MERRA-2 data. Wind vectors represent the mean MAM wind between 950 and 700 hPa. The red contour lines indicate the ALLJ occurrence frequency at 10%, 30%, 50%, and 70%.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
As in Fig. 7, but showing the diurnal cycle of vertically integrated moisture flux (Q; kg m−1 s−1) between 925 and 600 hPa based on MERRA-2 data. Wind vectors represent the mean MAM wind between 950 and 700 hPa. The red contour lines indicate the ALLJ occurrence frequency at 10%, 30%, 50%, and 70%.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
We postulate that the ALLJ enhances moisture flux from the Atlantic Ocean overnight that remains present to assist the onset of daytime convection in the central Amazon in the late morning hours (i.e., around 1100 LT). At 1400 and 1700 LT, the times associated with the beginning of the sea-breeze penetration and the greater development of the convective planetary boundary layer, the moisture flux is weaker and daytime heating is the main driver of convective growth. We also observe that the Amazon River is an important channel of humidity transport for central Amazonia from 1400 to 1700 LT.
f. Highly active and weakly active convective periods
Reanalysis winds were further evaluated during the highly and weakly active convective days in Table 1 to determine if there were any wind profile variations associated with the level of convective activity over the Amazon basin during GoAmazon2014/5. During weakly active days, convective systems preferentially form in the afternoon (e.g., starting around 1700 UTC in Fig. 2d, which is 1300 LT in the central Amazon) and are active into the early evening; however, they are much less active in the late evening/early morning hours. On highly active days, convective systems also preferentially form in the afternoon, but can persist through the evening and propagate farther (e.g., Fig. 2b). An exceptional sequence of large and long-lived systems is evident in Fig. 2b starting near the coast (50°W) before 0000 UTC and propagating westward to the central Amazon near Manaus (60°W) by 2000 UTC (or 1600 LT). These large, long-lasting systems are analogous to the ACSLs observed during ABLE-2 B (Garstang et al. 1994) and represent a subset of the convection that occurs on highly active days. The mean winds at 900 and 200 hPa (Fig. 9) and along cross section A (Fig. 10) are shown from MERRA-2 for highly and weakly active days and the difference between them.
Mean MAM wind (m s−1) from MERRA-2 during the (a),(d) highly active periods and (b),(e) weakly active periods at 900 and 200 hPa. (c),(f) The difference between the highly and weakly active period winds at 900 and 200 hPa.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM wind (m s−1) from MERRA-2 during the (a),(d) highly active periods and (b),(e) weakly active periods at 900 and 200 hPa. (c),(f) The difference between the highly and weakly active period winds at 900 and 200 hPa.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM wind (m s−1) from MERRA-2 during the (a),(d) highly active periods and (b),(e) weakly active periods at 900 and 200 hPa. (c),(f) The difference between the highly and weakly active period winds at 900 and 200 hPa.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM wind speed (m s−1) from MERRA-2 along cross section A for (a) highly active and (b) weakly active convective days. (c) The wind speed difference between the highly active and weakly active days.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM wind speed (m s−1) from MERRA-2 along cross section A for (a) highly active and (b) weakly active convective days. (c) The wind speed difference between the highly active and weakly active days.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Mean MAM wind speed (m s−1) from MERRA-2 along cross section A for (a) highly active and (b) weakly active convective days. (c) The wind speed difference between the highly active and weakly active days.
Citation: Monthly Weather Review 148, 10; 10.1175/MWR-D-19-0414.1
Figures 9a and 9b show that the northeasterly winds do not extend as far inland during highly active days compared to weakly active days. However, low-level winds are slightly stronger over most of north and northeast Brazil on highly active days (Fig. 9c) potentially advecting more moisture into the central Amazon. During the CHUVA-Belem field campaign (Machado et al. 2014), Adams et al. (2015) observed more moisture transport from Belem to Manaus associated with the sea breeze on convective days compared to nonconvective days.
The 900-hPa winds during weakly active days are much stronger to the west and along the Andes and are reminiscent of the westerly phase 850-hPa wind composites during the Tropical Rainfall Measuring Mission–Large-Scale Biosphere–Atmosphere Experiment (TRMM-LBA; Halverson et al. 2002). Convective systems were considered more oceanic (i.e., weaker, but larger) over the southwest Amazon during the westerly phases of TRMM-LBA so it would be interesting to study convective interactions with the ALLJ in DJF, which is when TRMM-LBA took place. Regardless, it appears that on weakly active convective days, the low-level winds from the Amazon more directly couple to the SALLJ.
At upper levels the main differences are found in the position and intensity of the Bolivian high, which has been linked to convective activity over the Amazon (Kousky and Kagano 1981; Molion and Kousky 1985; Gandu and Silva Dias 1998). During the highly active convective days identified in this study, the Bolivian high is clearly present at 200 hPa (Fig. 9d), as well as strong southerly flow over the Atlantic. During weakly active days (Fig. 9e), the Bolivian high is not present and southwesterly flow predominates over Amazonia at upper levels. The difference between the highly active and weakly active periods at 200 hPa (Fig. 9f) shows anticyclonic anomalies reaching 12–15 m s−1 and an anomalous cyclonic vortex can be seen over the Atlantic. Thus, the large-scale synoptic field in MAM appears to assist in the raininess and organization of convection over the Amazon via upper-level anticyclonic circulations over South America (Gan et al. 2004) and upper-level cyclonic vortices over the tropical South Atlantic (Kousky and Gan 1981).
The mean MAM MERRA-2 winds along cross section A show that the low-level wind enhancement associated with the ALLJ is deeper as well as stronger near the coast during highly active days (cf. Figs. 10a,b), again suggesting stronger moisture flux inland and consistent with Alcântara et al. (2011). About 600 km from the coast, the midtropospheric wind anomaly (Fig. 10c) is stronger for the highly active days. This strengthening (and the coincident weakening of winds closer to the surface) could potentially be due to increased vertical momentum transport by the highly organized convective systems across the region on highly active days (Moncrieff 1992, 2004).
4. Conclusions
Prompted by our goal to understand the observed large-scale variations of MCS propagation and life cycle over the Amazon basin during GoAmazon2014/5, we discovered the existence of an Amazonian low-level jet that spans much of the Amazon basin in MAM. The ALLJ has a vertical structure, horizontal extent and diurnal cycle distinct from the coastal sea breeze and northeasterly trades and unifies previous observations of more local jet features across the Amazon. For example, the diurnal oscillation of low-level winds observed during ABLE 2A and 2B by Greco et al. (1992) and de Oliveira and Fitzjarrald (1993) was interpreted as a local phenomenon related to nocturnal eddy viscosity reduction and local horizontal pressure gradients between the forest, river, topography and city. However, the wind cross sections shown in this work highlight diurnal jet variations across a large extension of Amazonia.
During MAM, the mean low-level wind speeds over the Amazon are enhanced at heights near 900 hPa in the late evening and early morning hours. This structure was similar between the ERA-I and MERRA-2 reanalyses and consistent with GoAmazon2014/5 soundings over the T3 site at Manacaparu. The low-level wind enhancement begins over the coast at 2000 LT and grows ~1600 km inland by 0800 LT. Objective wind criteria were applied to the MERRA-2 data and showed that a low-level jet is present across the Amazon 10%–40% of the time during the evening and early morning hours. Even though the ALLJ does not occur every day, it still impacts the climatological low-level winds over the Amazon. However, better understanding of what controls the occurrence of the ALLJ should lead to more insight on the dynamical controls of the jet. At 1400 and 1700 LT, there is a break period related to the increase of eddy viscosity due to the daytime convective boundary layer development and the ALLJ is no longer present.
Enhanced cloud cluster occurrence and significant moisture flux from the Atlantic into the central Amazon is associated with the ALLJ. The ALLJ acts as a source of moisture for both nighttime convective systems near the coast and afternoon convective systems in the central Amazon. The ALLJ is stronger in the central Amazon on weakly convective days, which may suggest intensification of the SALLJ when the Amazon is less convectively active. A positive midtropospheric wind anomaly occurs in the central Amazon near 600 hPa on highly active days, potentially due to vertical momentum transport from the ALLJ by deep moist convection.
This work shows that the inland propagation and evolution of cloud clusters as observed by Garreaud and Wallace (1997) and Burleyson et al. (2016) over the Amazon region are associated with two diurnal wind features: the coastal afternoon sea breeze and the nocturnal activity of the ALLJ farther inland. The sea breeze enhances cloud cluster occurrence and propagation at the coast in the early afternoon (e.g., 1400–1700 LT). After this, the sea breeze loses strength and there is a reduction in coastal cloud cluster activity. However, during the late evening and early morning hours, cloud clusters increase slightly inland and to the northwest of the coastal convection and propagate into the central Amazon with the development of the ALLJ. These geographical and temporal differences in cloud cluster occurrence and propagation indicate that the ACSLs that weaken and then regenerate as described in Garstang et al. (1994) are likely associated with the demise of the sea breeze and the beginning of ALLJ, respectively. The majority of daytime convection in the central Amazon is separate from the late afternoon coastal sea breeze and nocturnal ALLJ.
There is much left to analyze concerning the dynamics of the ALLJ and its interaction with convection over the Amazon. Separate analysis suggests that the ALLJ is present in other seasons (except JJA) and that climate models can capture its characteristics and links to convective development to varying degrees. It is also of possible interest to compare and contrast the ALLJ to the GPLLJ and SALLJ. For example, in the GPLLJ there is a clockwise wind direction rotation during the diurnal cycle and the highest speed occurs between 0000 and 0300 LT (Bonner 1968; Bonner and Paegle 1970; Jiang et al. 2007; Du and Rotunno 2014). In the ALLJ, there is no wind direction change during the diurnal oscillation letting the jet configuration last longer (from 2000 to 0800 LT). The most pronounced ALLJ activity is between 0500 and 0800 LT, which is later than the GPLLJ. While the GPLLJ has been associated with topographic forcing (e.g., Holton 1967), the ALLJ has only moderate elevated topographic forcing along its main path. The nocturnal circulations due to topography are more important along the east side of Andes in the SALLJ, where Marengo et al. (2002, 2004) observed the highest jet velocities mainly between 0200 and 0800 LT. One of the most important mechanisms associated with the SALLJ is the deflection of the trade wind circulation along the Andes (Vera et al. 2006), and the ALLJ potentially plays an important role in the SALLJ by influencing the winds and moisture flux before this deflection occurs. In future work, it would be interesting to compare characteristics of the ALLJ with the GPLLJ and SALLJ in more detail.
Acknowledgments
We acknowledge financial support from the Coordination for the Improvement of Higher Education Personnel (CAPES) Grant 99999.000481/2016-05, FAPESP 2015/14497-0, and from the U.S. Department of Energy (DOE) Grant DE-SC0016245. We would also like to thank John Nielsen-Gammon, Craig Epifanio, and Dale Durran for helpful discussions. We also thank David Adams and the anonymous reviewers for their suggestions and recommendations.
REFERENCES
Adams, D. K., and Coauthors, 2015: The Amazon dense GNSS meteorological network: A new approach for examining water vapor and deep convection interactions in the tropics. Bull. Amer. Meteor. Soc., 96, 2151–2165, https://doi.org/10.1175/BAMS-D-13-00171.1.
Alcântara, C. R., M. A. S. Dias, E. P. Souza, and J. C. Cohen, 2011: Verification of the role of the low level jets in Amazon squall lines. Atmos. Res., 100, 36–44, https://doi.org/10.1016/j.atmosres.2010.12.023.
Blackadar, A. K., 1957: Boundary layer wind maxima and their significance for growth of nocturnal inversions. Bull. Amer. Meteor. Soc., 38, 283–290, https://doi.org/10.1175/1520-0477-38.5.283.
Bonner, W. D., 1968: Climatology of the low level jet. Mon. Wea. Rev., 96, 833–850, https://doi.org/10.1175/1520-0493(1968)096<0833:COTLLJ>2.0.CO;2.
Bonner, W. D., and J. Paegle, 1970: Diurnal variations in boundary layer winds over the south-central United States in summer. Mon. Wea. Rev., 98, 735–744, https://doi.org/10.1175/1520-0493(1970)098<0735:DVIBLW>2.3.CO;2.
Burleyson, C. D., Z. Feng, S. M. Hagos, J. Fast, L. A. T. Machado, and S. T. Martin, 2016: Spatial variability of the background diurnal cycle of deep convection around the GoAmazon2014/5 field campaign sites. J. Appl. Meteor. Climatol., 55, 1579–1598, https://doi.org/10.1175/JAMC-D-15-0229.1.
Carneiro, R. G., and G. Fisch, 2020: Observational analysis of the daily cycle of the planetary boundary layer in the central Amazon during a non-El Niño year and El Niño year (GoAmazon project 2014/5). Atmos. Chem. Phys., 20, 5547–5558, https://doi.org/10.5194/acp-20-5547-2020.
Chen, S. S., R. A. Houze Jr., and B. E. Mapes, 1996: Multiscale variability of deep convection in realation to large-scale circulation in TOGA COARE. J. Atmos. Sci., 53, 1380–1409, https://doi.org/10.1175/1520-0469(1996)053<1380:MVODCI>2.0.CO;2.
Cohen, J. C. P., 1989: Um estudo observacional de linhas de instabiblidade na Amazônia. Ph.D. thesis, National Institute for Space Research, 160 pp.
Cohen, J. C. P., M. A. Silva Dias, and C. A. Nobre, 1995: Environmental conditions associated with Amazonian squall lines: A case study. Mon. Wea. Rev., 123, 3163–3174, https://doi.org/10.1175/1520-0493(1995)123<3163:ECAWAS>2.0.CO;2.
Dalu, G. A., and R. A. Pielke, 1989: An analytical study of the sea breeze. J. Atmos. Sci., 46, 1815–1825, https://doi.org/10.1175/1520-0469(1989)046<1815:AASOTS>2.0.CO;2.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
de Oliveira, A. P., and D. R. Fitzjarrald, 1993: The Amazon River breeze and the local boundary layer: I. Observations. Bound.-Layer Meteor., 63, 141–162, https://doi.org/10.1007/BF00705380.
Du, Y., and R. Rotunno, 2014: A simple analytical model of the nocturnal low-level jet over the Great Plains of the United States. J. Atmos. Sci., 71, 3674–3683, https://doi.org/10.1175/JAS-D-14-0060.1.
Gan, M. A., V. E. Kousky, and C. F. Ropelewski, 2004: The South America monsoon circulation and its relationship to rainfall over west-central Brazil. J. Climate, 17, 47–66, https://doi.org/10.1175/1520-0442(2004)017<0047:TSAMCA>2.0.CO;2.
Gandu, A. W., and P. L. Silva Dias, 1998: Impact of tropical heat sources on the South American tropospheric upper circulation and subsidence. J. Geophys. Res., 103, 6001–6015, https://doi.org/10.1029/97JD03114.
Garreaud, R., and J. M. Wallace, 1997: The diurnal march of convective cloudiness over the Americas. Mon. Wea. Rev., 125, 3157–3171, https://doi.org/10.1175/1520-0493(1997)125<3157:TDMOCC>2.0.CO;2.
Garstang, M., H. L. Massie Jr., J. Halverson, S. Greco, and J. Scala, 1994: Amazon coastal squall lines. Part I: Structure and kinematics. Mon. Wea. Rev., 122, 608–622, https://doi.org/10.1175/1520-0493(1994)122<0608:ACSLPI>2.0.CO;2.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Greco, S., S. Ulanski, M. Garstang, and S. Houston, 1992: Low-level nocturnal wind maximum over the central Amazon basin. Bound.-Layer Meteor., 58, 91–115, https://doi.org/10.1007/BF00120753.
Halverson, J. B., T. Rickenbach, B. Roy, H. Pierce, and E. Williams, 2002: Environmental characteristics of convective systems during TRMM-LBA. Mon. Wea. Rev., 130, 1493–1509, https://doi.org/10.1175/1520-0493(2002)130<1493:ECOCSD>2.0.CO;2.
Holton, J. R., 1967: The diurnal boundary layer wind oscillation above sloping terrain. Tellus, 19, 200–205, https://doi.org/10.3402/tellusa.v19i2.9766.
Itterly, K. F., P. C. Taylor, and J. B. Dodson, 2018: Sensitivity of the Amazonian convective diurnal cycle to its environment in observations and reanalysis. J. Geophys. Res. Atmos., 123, 12 621–12 646, https://doi.org/10.1029/2018JD029251.
Jiang, X., N.-C. Lau, I. M. Held, and J. J. Ploshay, 2007: Mechanisms of the Great Plains low-level jet as simulated in an AGCM. J. Atmos. Sci., 64, 532–547, https://doi.org/10.1175/JAS3847.1.
Kousky, V. E., 1980: Diurnal rainfall variation in Northeast Brazil. Mon. Wea. Rev., 108, 488–498, https://doi.org/10.1175/1520-0493(1980)108<0488:DRVINB>2.0.CO;2.
Kousky, V. E., and M. A. Gan, 1981: Upper tropospheric cyclonic vortices in the tropical South Atlantic. Tellus, 33, 538–551, https://doi.org/10.3402/tellusa.v33i6.10775.
Kousky, V. E., and M. T. Kagano, 1981: A climatological study of the trospospheric circulation over the Amazon region. Acta Amazonica, 11, 743–758, https://doi.org/10.1590/1809-43921981114743.
Machado, L. A. T., W. Rossow, R. Guedes, and A. Walker, 1998: Life cycle variations of mesoscale convective systems over the Americas. Mon. Wea. Rev., 126, 1630–1654, https://doi.org/10.1175/1520-0493(1998)126<1630:LCVOMC>2.0.CO;2.
Machado, L. A. T., and Coauthors, 2014: The CHUVA project: How does convection vary across Brazil? Bull. Amer. Meteor. Soc., 95, 1365–1380, https://doi.org/10.1175/BAMS-D-13-00084.1.
Mapes, B. E., T. T. Warner, and M. Xu, 2003: Diurnal patterns of rainfall in northwestern South America. Part III: Diurnal gravity waves and nocturnal convection offshore. Mon. Wea. Rev., 131, 830–844, https://doi.org/10.1175/1520-0493(2003)131<0830:DPORIN>2.0.CO;2.
Marengo, J. A., M. W. Douglas, and P. L. Silva Dias, 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, https://doi.org/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, https://doi.org/10.1175/1520-0442(2004)017<2261:COTLJE>2.0.CO;2.
Martin, S. T., and Coauthors, 2016: Introduction: Observations and modeling of the Green Ocean Amazon (GoAmazon2014/5). Atmos. Chem. Phys., 16, 4785–4797, https://doi.org/10.5194/acp-16-4785-2016.
Mathon, V., and H. Laurent, 2001: Life cycle of Sahelian mesoscale convective cloud systems. Quart. J. Roy. Meteor. Soc., 127, 377–406, https://doi.org/10.1002/qj.49712757208.
Molion, L. C. B., and V. E. Kousky, 1985: Climatologia da dinâmica da troposfera sobre a Amazônia. INPE Tech. Rep. INPE-3560-RPE/480, INPE, São José dos Campos, São Paulo, 28 pp.
Monaghan, A. J., D. L. Rife, J. O. Pinto, C. A. Davis, and J. Hannan, 2010: Global precipitation extremes associated with diurnally varying low-level jets. J. Climate, 23, 5065–5084, https://doi.org/10.1175/2010JCLI3515.1.
Moncrieff, M. W., 1992: Organized convective systems: Archetypal dynamical models, mass and momentum flux theory, and parametrization. Quart. J. Roy. Meteor. Soc., 118, 819–850, https://doi.org/10.1002/qj.49711850703.
Moncrieff, M. W., 2004: Analytic representation of the large-scale organization of tropical convection. J. Atmos. Sci., 61, 1521–1538, https://doi.org/10.1175/1520-0469(2004)061<1521:AROTLO>2.0.CO;2.
Negri, A. J., R. F. Adler, E. J. Nelkin, and G. J. Huffman, 1994: Regional rainfall climatologies derived from Special Sensor Microwave Imager (SSM/I) data. Bull. Amer. Meteor. Soc., 75, 1165–1182, https://doi.org/10.1175/1520-0477(1994)075<1165:RRCDFS>2.0.CO;2.
Orlanski, I., 1973: Trapeze instability as a source of internal gravity waves. Part I. J. Atmos. Sci., 30, 1007–1016, https://doi.org/10.1175/1520-0469(1973)030<1007:TIAASO>2.0.CO;2.
Paegle, J., M. McCorcle, and E. Miller, 1984: Diagnoses and numerical simulation of a low-level jet during ALPEX. Beitr. Phys. Atmos., 57, 419–430.
Planchon, O., F. Damato, V. Dubreuil, and P. Gouéry, 2006: A method of identifying and locating sea-breeze fronts in north-eastern Brazil by remote sensing. Meteor. Appl., 13, 225–234, https://doi.org/10.1017/S1350482706002283.
Rickenbach, T. M., 2004: Nocturnal cloud systems and the diurnal variation of clouds and rainfall in southwestern Amazonia. Mon. Wea. Rev., 132, 1201–1219, https://doi.org/10.1175/1520-0493(2004)132<1201:NCSATD>2.0.CO;2.
Rife, D. L., J. O. Pinto, A. J. Monaghan, C. A. Davis, and J. R. Hannan, 2010: Global distribution and characteristics of diurnally varying low-level jets. J. Climate, 23, 5041–5064, https://doi.org/10.1175/2010JCLI3514.1.
Rotunno, R., 1983: On the linear theory of the land and sea breeze. J. Atmos. Sci., 40, 1999–2009, https://doi.org/10.1175/1520-0469(1983)040<1999:OTLTOT>2.0.CO;2.
Salio, P., M. Nicolini, and E. J. Zipser, 2007: Mesoscale convective systems over southeastern South America and their relationship with the South American low-level jet. Mon. Wea. Rev., 135, 1290–1309, https://doi.org/10.1175/MWR3305.1.
Silva Dias, M. A. F., and R. N. Ferreira, 1992: Application of a linear spectral model to the study of Amazonian squall lines during GTE/ABLE 2B. J. Geophys. Res., 97, 20 405–20 419, https://doi.org/10.1029/92JD01333.
Souza, D. C., and M. D. Oyama, 2017: Breeze potential along the Brazilian northern and northeastern coast. J. Aerosp. Technol. Manage., 9, 368–378, https://doi.org/10.5028/jatm.v9i3.787.
Stensrud, D. J., 1996: Importance of low-level jets to climate: A review. J. Climate, 9, 1698–1711, https://doi.org/10.1175/1520-0442(1996)009<1698:IOLLJT>2.0.CO;2.
Sun, W.-Y., and I. Orlanski, 1981: Large mesoscale convection and sea breeze circulation. Part I: Linear stability analysis. J. Atmos. Sci., 38, 1675–1693, https://doi.org/10.1175/1520-0469(1981)038<1675:LMCASB>2.0.CO;2.
Tulich, S. N., and G. N. Kiladis, 2012: Squall lines and convectively coupled gravity waves in the tropics: Why do most cloud systems propagate westward? J. Atmos. Sci., 69, 2995–3012, https://doi.org/10.1175/JAS-D-11-0297.1.
Velasco, I., and J. M. Fritsch, 1987: Mesoscale convective complexes in the Americas. J. Geophys. Res., 92, 9591–9613, https://doi.org/10.1029/JD092iD08p09591.
Vera, C., and Coauthors, 2006: The South American low-level jet experiment. Bull. Amer. Meteor. Soc., 87, 63–78, https://doi.org/10.1175/BAMS-87-1-63.
Vila, D. A., L. A. T. Machado, H. Laurent, and I. Velasco, 2008: Forecast and Tracking the Evolution of Cloud Clusters (ForTraCC) using satellite infrared imagery: Methodology and validation. Wea. Forecasting, 23, 233–245, https://doi.org/10.1175/2007WAF2006121.1.
Virji, H., 1981: A preliminary study of summertime tropospheric circulation patterns over South America estimated from cloud winds. Mon. Wea. Rev., 109, 599–610, https://doi.org/10.1175/1520-0493(1981)109<0599:APSOST>2.0.CO;2.
Whiteman, C. D., X. Bian, and S. Zhong, 1997: Low-level jet climatology from enhanced rawinsonde observations at a site in the southern Great Plains. J. Appl. Meteor., 36, 1363–1376, https://doi.org/10.1175/1520-0450(1997)036<1363:LLJCFE>2.0.CO;2.
Williams, M., and R. A. Houze Jr., 1987: Satellite-observed characteristics of winter monsoon cloud clusters. Mon. Wea. Rev., 115, 505–519, https://doi.org/10.1175/1520-0493(1987)115<0505:SOCOWM>2.0.CO;2.
Woodley, W. L., C. G. Griffith, J. S. Griffin, and S. C. Stromatt, 1980: The inference of GATE convective rainfall from SMS-1 imagery. J. Appl. Meteor., 19, 388–408, https://doi.org/10.1175/1520-0450(1980)019<0388:TIOGCR>2.0.CO;2.
Zipser, E. J., C. Liu, D. J. Cecil, S. W. Nesbitt, and D. P. Yorty, 2006: Where are the most intense thunderstorms on Earth? Bull. Amer. Meteor. Soc., 87, 1057–1072, https://doi.org/10.1175/BAMS-87-8-1057.