Universidade Federal de Santa Maria
Ci. e Nat., Santa Maria v.42, Special Edition: Micrometeorologia, e2, 2020
DOI:10.5902/2179460X45217
ISSN 2179-460X
Received: 27/05/20 Accepted: 27/05/20 Published: 28/08/20
Special Edition
Convective boundary layer characterization in the Amazon rainforest before and after the passage of Mesoscale convective systems
Caracterização da camada limite convectiva na floresta amazônica antes e depois da passagem de sistemas convectivos de mesoescala
Vanessa C. Monteiro I
I Pennsylvania State University, University Park, USA – vvm5136@psu.edu
RESUMO
Neste estudo são descritas variáveis termodinâmicas, como temperatura e humidade, da camada limite atmosférica convectiva (CLC) e sua taxa de crescimento antecedente e posterior a passagem de sistemas convectivos de mesoescala (SCMs) na floresta Amazônica. Utilizando os dados obtidos através do experimento GoAmazon 2014/15, são analisados dois casos de estudo e uma amostra de dias nos quais ocorreram a passagem de SCMs. Os resultados mostram que a temperatura potencial equivalente da CLC sofre redução de 2 a 8K e a umidade específica é reduzida em até 2 g kg-1 após a passagem de um SCM, devido ao ar seco e frio trazido para a superfície através de correntes de ar descendentes (downdrafts). Essas duas variáveis em adição a outras, como fluxos de energia, são responsáveis por baixas taxas de crescimento da CLC, taxas que são reduzidas a 100 m h-1 nas horas seguintes após a precipitação cessar, quando comparado com dias sem precipitação. Este trabalho mostra uma quantificação de variáveis termodinâmicas da CLC durante condições prévias e posteriores à precipitação, que podem servir de base e complemento para futuros estudos nessa região.
Palavras-chave: Floresta amazônica; Camada limite convectiva; Sistemas convectivos de mesoescala; Termodinâmica.
ABSTRACT
This study describes thermodynamic variables (e.g., temperature and humidity) of the atmospheric convective boundary layer (CBL) and its growth rates preceding and following the passage of mesoscale convective systems (MCSs) in the Amazon rainforest. Using the data set provided by GoAmazon 2014/15, this study will address its objectives through the evaluation of case studies and an ensemble of days when there was the passage of MCSs. The results show that the convective boundary layer experiences reductions in the equivalent potential temperature within 2 to 8K and in the specific humidity up to 2 g kg-1 after the passage of a MCS, due to the cold and dry air brought to the surface by storms downdrafts. These two variables in addition to others (e.g., energy fluxes) are responsible for the low growth rates of the convective boundary layer, that were reduced by 100 m h-1 in the following two hours after the rainfall ceases, when compared to undisturbed conditions. Nonetheless, this work provides a quantitative evaluation of the thermodynamic features of the convective boundary layer under the passage of mesoscale convective systems in the Amazon rainforest.
Keywords: Amazon rainforest; Convective boundary layer; Mesoscale convective systems; Thermodynamics.
1 Introduction
The Amazon rainforest has an important role in climate regulation due to its complex interaction among forest, water resources, and the atmosphere. During 2014 and 2015, the GoAmazon (Green Ocean Amazon) experiment (MARTIN et al., 2017) was set up to measure meteorological variables at the surface and the atmospheric column (e.g., temperature, relative humidity, air pressure, wind speed), and chemicals mixing ratios (e.g., ozone, isoprene, monoterpenes), to better understand the cloud formation in pristine and polluted environments. Since then, many results were already reported in atmospheric chemistry and aerosol particles (LIU et al., 2018; FRAUND et al., 2017; SÁ et al., 2017; GU et al., 2017; FREIRE et al., 2017; THALMAN et al., 2017; CECCHINI et al., 2017; FUENTES et al., 2016; PÖHLKER et al., 2016; MARTIN et al., 2017; GERKEN et al., 2016); convective parameterizations (SCHIRO et al., 2018; SONG; ZHANG, 2017); buoyancy was deeply investigated (AHMED; NEELIN, 2018; ZHUANG; FU; WANG, 2018), as well precipitation processes (MACHADO et al., 2018; AHMED; NEELIN, 2018; KUO; SCHIRO; NEELIN, 2018; SCHIRO; NEELIN, 2018; GIANGRANDE et al., 2017; MARENGO et al., 2017; DIAS-JUNIOR et al., 2017; COLLOW; MILLER; TRABACHINO, 2016).
However, mesoscale convective systems (MCSs) and its impacts in the convective boundary layer (CBL), also known as mixed layer (ML), are not well represented in climate models (SCHIRO et al., 2018). A study derived from GoAmazon 2014/15 campaign (SCHIRO; NEELIN, 2018) showed the variations of the thermodynamics before and after the passage of MCSs, but studies showing the CBL evolution after and before such events are not found in the literature. Thus, this work intends to examine the thermodynamic features of the CBL (e.g., temperature and humidity) to characterize the CBL growth rates in the Amazon rainforest before and after MCSs, using the data set provided on GoAmazon 2014/15 experiment (MARTIN et al., 2017). Case studies representing the precipitation events (MCSs), called disturbed days, will be explored. This work aims to provide a deeper understanding of the unique features resulting from the passage of convective systems over a forested area.
2 Material and Methods
The Green Ocean Amazon (GoAmazon) was an experiment set up to obtain datasets to better understand the coupled atmosphere-cloud-terrestrial tropical systems (MARTIN et al., 2017). From January 2014 to December 2015, the Office of Biological and Environmental Research’s Climate and Environmental Sciences Division in collaboration with Brazilian and German organizations, collected datasets to study the interactions and impacts of pollutants coming from megacities into the pristine forest, and to try to understand how it can interfere in cloud formation (MARTIN et al., 2017).
The experiment consisted of a deployment of nine ground sites and two onboard aircraft (MARTIN et al., 2017). For the current study, was selected one site, called T3, with geographical coordinates 3⁰ 12’ 46.70” S and 60⁰ 35’ 53.0” W. It is located 70 km southeast (downwind) from the city of Manaus (Brazil) and receives the pollutants transported from Manaus. During the period between 1 January 2014 and 31 November 2015, the Atmospheric Radiation Measurement (ARM) Climate Research Facility of the United States Department of Energy (MARTIN et al., 2017; MATHER; VOYLES, 2013) operated a set of equipment to measure energy fluxes, wind speed and direction, radiation components, humidity, temperature, chemical fluxes, precipitation, and convective available potential energy (CAPE). Upper air measurements, such as air pressure, temperature, relative humidity, wind speed and direction, were obtained through radiosondes launched over the T3 site at 6-hours intervals (01:30, 07:30, 13:30 and 19:30 LT) and occasionally 10:30 LT over the wet season. The reason to choose this site is the data availability that makes possible to conduct the aimed analysis.
In order to evaluate the thermodynamic
attributes of the CBL, the mixed layer depth () was obtained from the
maximum value of the second derivative of the virtual potential temperature (
) (STRONG et al., 2005).
Subsequently, averaged mixed layer specific humidity (
) and virtual potential
temperature (
) were obtained by
integrating the quantities from the surface to the top of the mixed layer (
) as
(1)
(2)
The strength of virtual potential temperature () and specific humidity
(
) inversions above the
mixed layer (
;
) were estimated to
determine how boundary layer attributes changed over the course of the day.
These quantities are obtained from the measurements observed at the top of the
mixed layer (
) and at height
, where
is the thickness of the entrainment zone. Above the mixed layer
(e.g.,
), variations of virtual
potential temperature (
) and specific humidity
(
) with altitude were
also estimated.
At the surface, sensible () and latent (
) heat fluxes densities
were obtained from kinematic quantities computed from eddy covariance
measurements. For humid environments, such as the Amazon rainforest, is defined
the virtual sensible flux (
) (BETTS, 1992).
For the case studies, when CAPE (convective available potential energy) is missing, it is estimated for the specific times when the soundings were launched, according:
(3)
where is the environment
temperature,
is the parcel
temperature,
is the height,
is the level of free convection,
is the equilibrium level, and
is the acceleration due
to the gravity.
The equivalent potential temperature () at the surface was
also derived. Relatively high equivalent potential temperatures in a region
with instability is an ingredient to induce mesoscale convective systems.
2 Results and Discussion
2.1 Case Studies
On 15 October 2014 was identified an isolated mesoscale convective system moving westward over the site of study at 04:00 LT (local time). The MCS passage resulted in a total rainfall of 30.8 mm within 6 hours of continuous precipitation, being the first three hours with the most intense rainfall.
As the air is advected due to the storm formation, the wind speed at the surface increased by 6 m s-1. In addition, the cold and dry air from the storm downdrafts reduced the equivalent potential temperature by approximately 7 K and the specific humidity by 2.12 g kg-1. As the surface started warming up after the sunrise and the evapotranspiration started taking place, the air temperature increased at a rate of 1.47 K h-1 and humidified at a rate of 0.74 g kg-1 h-1 (figures not shown). Some of the results are summarized in table 1.
High incoming solar radiation characterized this day, reaching maximum of 1208.6 W m-2 at 13:30 LT. This incoming solar radiation contributed to a high virtual sensible heat flux, which maximum reached 87.7 W m-2 (13:30 LT), and to a high latent heat flux, with a maximum of 433.0 W m-2 (13:00 LT). The convective potential available energy attained the maximum of 2578 J kg-1. The high incoming energy warms the air parcels increasing its buoyancy, therefore producing thermal turbulence, which will contribute to the formation of a deep CBL. These conditions led to a mixed layer growth rate of 155 m h-1 between 08:00 LT and 14:00 LT. Evaluation of growth rates for undisturbed days (not presented here) can be more than 2 times larger when compared to the presented case.
The temperature inversion at the entrainment
zone varied from 1.3 K (08:00 LT) to 0.3 K (14:00 LT). The capping inversion
became weaker over the day and the lapse rate above the mixed layer varied from
2.4 K km-1 (08:00 LT) to 4.1 K km-1 (14:00 LT). The mean virtual
potential temperature of the CBL () varied from 296 K to
301.2 K and the mean specific humidity (
) varied from 15.8 to
16.4 g kg-1, respectively at 08:00 and 14:00 LT.
The storm passage, during the nighttime, caused an additional cooling in the lower layer of the troposphere which took extra time to warm up. Also, the cloudy conditions in the early morning decreased the incoming solar radiation, slowing the formation of an even deeper mixed layer. Which agrees with the observation made by Garstang et al. (1990) that says that early morning precipitation can delay the growth of the mixed layer.
Table 1 – Thermodynamic attributes for 15 October 2014, derived from the measurements at T3 site
|
Attributes |
08:00 LT |
14:00 LT |
|
|
223 |
1150 |
|
|
296 |
301.2 |
|
|
15.8 |
16.4 |
|
|
-0.2 |
-0.7 |
|
|
1.3 |
0.3 |
|
|
2.4 |
4.1 |
|
|
-4.3 |
-1.4 |
Even though the passage of storms over the night can slow down the development of the mixed layer, the main impacts in its development can be seen when storms occur during the day. To illustrate this situation was taken a case of study when there was a passage of a MCS over the site of study in the morning (09 November 2014, 09:00 LT) which resulted in an accumulated rainfall of 43.6 mm. The storm formation led to an increase in the surface wind speed of 10.6 m s-1.
The high CAPE (2895 J kg-1) is another indicator of the strength of the storm. During the passage of the storm, the equivalent potential temperature was reduced about 22 K and the specific humidity by 5.76 g kg-1, due to the storm downdrafts carrying cold and dry air from aloft to the CBL.
After the rainfall, the CBL air temperature
increased at a rate of 0.78 K h-1, almost half of the rate observed
for 15 October 2014, and the humidification rate was four times smaller than on
15 October 2014, 0.18 g kg-1 (figures not shown). These warming and
humidification rates are up to 7 times lower than the average for typical
undisturbed days. The mean virtual potential temperature of the CBL (), for this day, varied
from 300.6 K to 297.3 K and the mean specific humidity (
) varied from 19.0 to
16.4 g kg-1, respectively at 08:00 and 14:00 LT, which quantifies the effect of
cold and dry air brought from storm downdrafts to the ABL. The daily average
solar radiation on 09 November 2014 was equivalent to 21% of the average for 15
October 2014, for example, and virtual sensible heat flux (daily average of
15.4 W m-2), resulted in weak surface heating. Consequently, low growth rates
of the ML are observed, which attained only 50 m h-1, with a ML depth of 312 m at 14:00
LT. This value represents approximately a quarter of typical values for
undisturbed days. The temperature inversion strength at the top of the ML
varied between 1.6 K (08:00 LT) and 3.2 K (14:00 LT). The passage of storms
during the day also led to a more stable entrainment zone, limiting the growth
of the CBL. Some of the results are summarized in table 2.
Table 2 – Thermodynamic attributes for 09 November 2014, derived from the measurements at T3 site
|
Attributes |
08:00 LT |
14:00 LT |
|
|
64 |
312 |
|
|
300.6 |
297.3 |
|
|
19.0 |
16.4 |
|
|
-0.2 |
-0.4 |
|
|
1.6 |
3.2 |
|
|
1.6 |
5.3 |
|
|
-1.7 |
-3.0 |
2.2 Averaged convective boundary layer growth under disturbed conditions
Taking an ensemble of 10 days with precipitation in the nighttime and an ensemble of 27 days with precipitation during the daytime we can see how long the mixed layer takes to grow from the time of the precipitation, and how it was before the precipitation. The criteria for defining the ensemble was to consider only events when the rainfall rates exceeded 20 mm h-1, which is a characteristic rainfall rate for MCSs over the site of study (FUENTES et al., 2016). For the nighttime period, the CBL in the previous day of the ensemble (6 to 20 hours before the rain event) had a high variability but was considerably deep (up to an average of 1067 m) (figure 1a). About 8 hours after the precipitation, the ML reaches, on average, growth rates of 202.7 m h-1, which is comparable to undisturbed days.
For the days with daytime precipitation, a shallow ML is observed up to 5 hours before the rain event, with an average depth of 170 m (figure 1b). However, after the passage of the MCS, the CBL depth was very variable. For some periods, growth rates were, on average, as high as 835 m h-1 (9 hours after the storm), and as low as 50 m h-1 (2 hours after the storm).
Some of the surface variables are summarized in figure 2 and shows the behavior of the variables according to the time of the precipitation event, using the same ensemble as presented in the figure 1b, except when there was no measurement for a specific day or variable, then the day was omitted from the ensemble. The averaged variables show a common feature prior to the precipitation event. The sensible heat flux, latent heat flux, equivalent potential temperature, specific humidity and convective available potential energy all decrease drastically at the moment antecedent the precipitation, and even after about 3 hours after the end of the rainfall do not return to its initial state.
Figure 1 – Convective boundary layer growth for disturbed days. (a) Nighttime occurrence of a MCS and (b) daytime occurrence of a MCS. The box has 50% of the data and the redline indicates the median (25% of the data over the median and 25% under the median), the other quartiles are represented by the upper (25%) and lower (25%) whiskers. The red symbols are the outliers
The sensible heat flux is reduced by an average of 38.0 W m-2 and latent heat flux by 223.0 W m-2 (figure 2a). The diminish of solar radiation due to the increasing cloudiness prevents the warming of the lower layers of the atmosphere, causing the gradual decrease in the intensity of these variables. Interestingly, the latent heat flux has a peak half-hour before the precipitation (from an average of 70.0 W m-2 to 223.0 W m-2 (figure 2b), what is explained by the fact that there is an intense transport of moisture from the surface to the higher levels of the atmosphere at this moment.
The averaged CAPE was approximately 1893.0 J kg-1 (figure 2e) before the rainfall starts, which is a compatible value expected for the occurrence of storms in the Amazon rainforest. The specific humidity and the equivalent potential temperature increased, for these cases, respectively, by 1.27 g kg-1 and 5.2 K (figure 2c,d). The importance of these variables is linked to the transport of moisture and temperature across the atmospheric boundary layer. The decrease in these variables is the result of the downdrafts associated to the storm passage, that brings dry and cold air from the layers above the CBL. Thus, these changes in moisture content and temperature at the surface are directly linked to the strength of the downdrafts, in contrast to the CAPE value that is related to the strength of the updrafts, since it is the amount of energy available for convection.
Figure 2 – Variation of thermodynamic variables at surface for mesoscale convective systems passage (time 0): 3 hours before the system’s passage (-3) up to 3 hours after (+3). (a) Sensible heat flux, (b) latent heat flux, (c) equivalent potential temperature, (d) specific humidity, and (e) convective available potential energy. The shadow represents the standard deviation of the data
The diurnal pattern of the variables explored above has a similar pattern comparable in “shape” to undisturbed days (not presented here), but different in intensity. While for sensible heat flux undisturbed days can have on average more than 100.0 W m-2 around noon, the disturbed days did not exceed the average of 55.0 W m-2 in the same period (figure 3a). The latent heat flux, however, exhibited a slightly higher value for disturbed days than undisturbed days. The maximum observed during the dry season (wet season) for the undisturbed condition is the average of 268.0 W m-2 (253.5 W m-2), and for the condition of mesoscale convective systems has recorded the average of 270.9 W m-2 (figure 3b). It can be attributed to the more intense activity of transport of moisture in disturbed conditions, as mentioned previously for the analysis focused on the time of the precipitation. Equivalent potential temperature has a mean diurnal value of 346.7 K (figure 3c), and the specific humidity, 19.4 g kg-1 (figure 3d), and CAPE, 1107.0 J kg-1 (figure 3e).
Figure 3 – Diurnal variation of thermodynamic variables at surface for days when there is a MCS passage. (a) Sensible heat flux, (b) latent heat flux, (c) equivalent potential temperature, (d) specific humidity, and (e) convective available potential energy. The shadow represents the standard deviation of the mean
3 Conclusion
The case studies illustrated thermodynamic variables for different times of the MCS occurrence. Comparing the two cases, when the precipitation occurred in the nighttime, the convective boundary layer (CBL) had a growth rate of 3.7 times larger than when the precipitation happened in the morning. The averaged virtual potential temperature and specific humidity changes in the layer also differ in both conditions. The precipitation during the morning delays the layer heating and dries the layer. These are ingredients that contribute to the lower CBL depth for such conditions.
The averaged variables obtained at the surface level for the ensemble of days when mesoscale convective systems occurred also reaffirm the drastic changes in the thermodynamic variables in the CBL. On average, virtual potential temperature dropped 5.2 K, but was as large as 22 K, as observed in the case studies and specific humidity could drop 2 g kg-1 on average. The combined changes in the CBL, due to the MCSs passage, resulted in a delay of its development, which growth rates were observed to be reduced up 100 m h-1 in the two hours following the precipitation when compared to averaged undisturbed conditions, not presented here.
This the work could provide some thermodynamic variables quantification regarding the passage of mesoscale convective systems over the Amazon rainforest. For future studies, it is needed to evaluate the dynamics involved in the process of the CBL growth, which is another relevant component that can prevents deep convective boundary layer depth.
Acknowledgments
This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001. VCM acknowledge support received from the Office of Biological and Environmental Research of the U.S. Department of Energy (under grant SC0011075) as part of the Atmospheric Radiation Measurement Climate Research Facility deployment as part of the GoAmazon project. VCM also thanks Dr. Jose D. Fuentes for useful discussions.
References
AHMED, F.; NEELIN, J. D. Reverse engineering the tropical precipitation–buoyancy relationship. Journal of the Atmospheric Sciences, v. 75, n. 5, p. 1587–1608, 2018.
BETTS, A. K. FIFE atmospheric boundary layer budget methods. J. Geophys. Res., v. 97, n. D17, p.18523–18531, 1992.
CECCHINI, M. A.; MACHADO, L. A.; ANDREAE, M. O.; MARTIN, S. T.; ALBRECHT, R. I.; ARTAXO, P.; BARBOSA, H. M.; BORRMANN, S.; FÜTTERER, D.; JURKAT, T. et al. Sensitivities of Amazonian clouds to aerosols and updraft speed. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 17, n. 16, p. 10037–10050, 2017.
COLLOW, A. B. M.; MILLER, M. A.; TRABACHINO, L. C. Cloudiness over the Amazon rainforest: Meteorology and thermodynamics. Journal of Geophysical Research: Atmospheres, Wiley Online Library, v. 121, n. 13, p. 7990–8005, 2016.
DIAS-JUNIOR, C. Q.; DIAS, N. L.; FUENTES, J. D.; CHAMECKI, M. Convective storms and nonclassical low-level jets during high ozone level episodes in the Amazon region: An ARM/GoAmazon case study. Atmospheric environment, Elsevier, v. 155, p. 199–209, 2017.
FRAUND, M.; PHAM, D. Q.; BONANNO, D.; HARDER, T. H.; WANG, B.; BRITO, J.; SÁ, S. S. de; CARBONE, S.; CHINA, S.; ARTAXO, P. et al. Elemental mixing state of aerosol particles collected in central Amazonia during GoAmazon2014/15. Atmosphere, Multidisciplinary Digital Publishing Institute, v. 8, n. 9, p. 173, 2017.
FREIRE, L.; GERKEN, T.; RUIZ-PLANCARTE, J.; WEI, D.; FUENTES, J.; KATUL, G.; DIAS, N.; ACEVEDO, O.; CHAMECKI, M. Turbulent mixing and removal of ozone within an Amazon rainforest canopy. Journal of Geophysical Research: Atmospheres, Wiley Online Library, v. 122, n. 5, p.2791–2811, 2017.
FUENTES, J. D.; CHAMECKI, M.; SANTOS, R. M. Nascimento dos; RANDOW, C. V.; STOY, P. C.; KATUL, G.; FITZJARRALD, D.; MANZI, A.; GERKEN, T.; TROWBRIDGE, A. et al. Linking meteorology, turbulence, and air chemistry in the amazon rain forest. Bulletin of the American Meteorological Society, v. 97, n. 12, p. 2329–2342, 2016.
GARSTANG, M.; ULANSKI, S.; GRECO, S.; SCALA, J.; SWAP, R.; FITZJARRALD, D.; MARTIN, D.; BROWELL, E.; SHIPMAN, M.; CONNORS, V. et al. The Amazon boundary-layer experiment (ABLE 2B): A meteorological perspective. Bulletin of the American Meteorological Society, American Meteorological Society, v. 71, n. 1, p. 19–32, 1990.
GERKEN, T.; WEI, D.; CHASE, R. J.; FUENTES, J. D.; SCHUMACHER, C.; MACHADO, L. A.; ANDREOLI, R. V.; CHAMECKI, M.; SOUZA, R. A. F. de; FREIRE, L. S. et al. Downward transport of ozone rich air and implications for atmospheric chemistry in the Amazon rainforest. Atmospheric Environment, Elsevier, v. 124, p. 64–76, 2016.
GIANGRANDE, S. E.; FENG, Z.; JENSEN, M. P.; COMSTOCK, J. M.; JOHNSON, K. L.; TOTO, T.; WANG, M.; BURLEYSON, C.; BHARADWAJ, N.; MEI, F. et al. Cloud characteristics, thermodynamic controls and radiative impacts during the observations and modeling of the Green Ocean Amazon (GoAmazon2014/5) experiment. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 17, n. 23, p. 14519–14541, 2017.
GU, D.; GUENTHER, A. B.; SHILLING, J. E.; YU, H.; HUANG, M.; ZHAO, C.; YANG, Q.; MARTIN, S. T.; ARTAXO, P.; KIM, S. et al. Airborne observations reveal elevational gradient in tropical forest isoprene emissions. Nature communications, Nature Publishing Group, v. 8, p. 15541, 2017.
KUO, Y.-H.; SCHIRO, K. A.; NEELIN, J. D. Convective transition statistics over tropical oceans for climate model diagnostics: Observational baseline. Journal of the Atmospheric Sciences, v. 75, n. 5, p. 1553–1570, 2018.
LIU, Y.; SECO, R.; KIM, S.; GUENTHER, A. B.; GOLDSTEIN, A. H.; KEUTSCH, F. N.; SPRINGSTON, S. R.; WATSON, T. B.; ARTAXO, P.; SOUZA, R. A. et al. Isoprene photo-oxidation products quantify the effect of pollution on hydroxyl radicals over Amazonia. Science advances, American Association for the Advancement of Science, v. 4, n. 4, p. 2547, 2018.
MACHADO, L. A.; CALHEIROS, A. J.; BISCARO, T.; GIANGRANDE, S.; DIAS, M. A. S.; CECCHINI, M. A.; ALBRECHT, R.; ANDREAE, M. O.; ARAUJO, W. F.; ARTAXO, P. et al. Overview: Precipitation characteristics and sensitivities to environmental conditions during GoAmazon2014/5 and ACRIDICON-CHUVA. Atmospheric Chemistry and Physics, opernicus GmbH, v. 18, n. 9, p. 6461–6482, 2018.
MARENGO, J. A.; FISCH, G. F.; ALVES, L. M.; SOUSA, N. V.; FU, R.; ZHUANG, Y. Meteorological context of the onset and end of the rainy season in central Amazonia during the GoAmazon2014/5. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 17, n. 12, p. 7671, 2017.
MARTIN, S.; ARTAXO, P.; MACHADO, L.; MANZI, A.; SOUZA, R.; SCHUMACHER, C.; WANG, J.; BISCARO, T.; BRITO, J.; CALHEIROS, A. et al. The Green Ocean Amazon experiment (GoAmazon2014/5) observes pollution affecting gases, aerosols, clouds, and rainfall over the rain forest. Bulletin of the American Meteorological Society, v. 98, n. 5, p. 981–997, 2017.
MATHER, J. H.; VOYLES, J. W. The ARM climate research facility: A review of structure and capabilities. Bull. Am. Meteorol. Soc., v. 94, n. 3, p. 377–392, 2013.
PÖHLKER, M. L.; PÖHLKER, C.; DITAS, F.; KLIMACH, T.; ANGELIS, I. Hrabe de; ARAÚJO, A.; BRITO, J.; CARBONE, S.; CHENG, Y.; CHI, X. et al. Long-term observations of cloud condensation nuclei in the Amazon rain forest – Part 1: Aerosol size distribution, hygroscopicity, and new model parametrizations for CCN prediction. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 16, n. 24, p. 15709–15740, 2016.
SÁ, S. S. d.; PALM, B. B.; CAMPUZANO-JOST, P.; DAY, D. A.; NEWBURN, M. K.; HU, W.; ISAACMAN-VANWERTZ, G.; YEE, L. D.; THALMAN, R.; BRITO, J. et al. Influence of urban pollution on the production of organic particulate matter from isoprene epoxydiols in central Amazonia. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 17, n. 11, p. 6611–6629, 2017.
SCHIRO, K. A.; AHMED, F.; GIANGRANDE, S. E.; NEELIN, J. D. GoAmazon2014/5 campaign points to deep-inflow approach to deep convection across scales. Proceedings of the National Academy of Sciences, National Acad. Sciences, v. 115, n. 18, p. 4577–4582, 2018.
SCHIRO, K. A.; NEELIN, J. D. Tropical continental downdraft characteristics: mesoscale systems versus unorganized convection. Atmos Chem Phys, v. 18, p. 1997–2010, 2018.
SONG, F.; ZHANG, G. J. Improving trigger functions for convective parameterization schemes using GoAmazon observations. Journal of Climate, v. 30, n. 21, p. 8711–8726, 2017.
STRONG, C.; FUENTES, J.; GARSTANG, M.; BETTS, A. Daytime cycle of low-level clouds and the tropical convective boundary layer in southwestern Amazonia. Journal of Applied Meteorology, v. 44, n. 10, p. 1607–1619, 2005.
THALMAN, R.; SÁ, S. S. d.; PALM, B. B.; BARBOSA, H. M.; PÖHLKER, M. L.; ALEXANDER, M. L.; BRITO, J.; CARBONE, S.; CASTILLO, P.; DAY, D. A. et al. CCN activity and organic hygroscopicity of aerosols downwind of an urban region in central Amazonia: seasonal and diel variations and impact of anthropogenic emissions. Atmospheric Chemistry and Physics, Copernicus GmbH, v. 17, n. 19, p. 11779–11801, 2017.
ZHUANG, Y.; FU, R.; WANG, H. How do environmental conditions influence vertical buoyancy structure and shallow-to-deep convection transition across different climate regimes? Journal of the Atmospheric Sciences, v. 75, n. 6, p. 1909–1932, 2018.