In this study, high-frequency, multilevel measurements, performed from late October to mid-November of 2015 at a 80 m tall tower of the Amazon Tall Tower Observatory (ATTO) project in the central state of Amazonas, Brazil, were used to diagnose the evolution of thermodynamic and kinematic variables as well as scalar fluxes during the passage of outflows generated by deep moist convection (DMC). Outflow associated with DMC activity over or near the tall tower was identified through the analysis of storm echoes in base reflectivity data from an S-band weather radar at Manaus, combined with the detection of gust fronts and cold pools utilizing tower data. Four outflow events were selected, three of which took place during the early evening transition or nighttime hours and one during the early afternoon. Results show that the magnitude of the drop in virtual potential temperature and changes in wind velocity during outflow passages vary according to the type, organization, and life cycle of the convective storm. The nocturnal events had well-defined gust fronts with moderate decreases in virtual potential temperature and increases in wind speed. The early afternoon event lacked a sharp gust front and only a gradual drop in virtual potential temperature was observed, probably because of weak or undeveloped outflow. Sensible heat flux (
Deep moist convection (DMC) is a ubiquitous feature of the atmospheric environment of the Amazon rainforest. Because of intense diurnal solar heating in the moist planetary boundary layer (PBL), conditional instability builds up and convective storms form regularly in order to redistribute energy in the atmospheric column
Much of the initial knowledge on the effects of DMC on PBL evolution has been gained from research based on the GARP Atlantic Tropical Experiment (GATE) in 1974
The occurrence of DMC has significant impacts upon the evolution of surface scalar fluxes, since convective outflows are responsible for cooling and drying the PBL
While most studies have focused on the DMC–PBL interaction over the tropical oceans, the evolution of turbulent fluxes in DMC situations in forest environments has also been addressed, either observationally
In this context, high-frequency tower measurements performed at the Amazon Tall Tower Observatory (ATTO)
This paper is organized as follows: Sect. 2 provides information about the datasets employed in this study along with the methods utilized for identifying storm events and computing turbulent quantities at the tower; Sect. 3 presents an overview of the main aspects of the convective storms that were analyzed, in terms of radar features and meteorological changes detected at the tower; Sect. 4 is aimed at investigating the mechanisms by which the fluxes of sensible and latent heat are enhanced throughout the instrumented tower depth and how they relate to PBL evolution in the wake of storms; in Sect. 5 we investigate the TKE evolution during storm outflows using high-frequency tower observations from ATTO; finally, the conclusions are presented in Sect. 6.
The primary data source employed in this investigation consists of high-frequency (10 Hz) micrometeorological measurements performed at the 80 m tall walk-up tower, located 150 km northeast of Manaus, in the Uatumã Sustainable Development Reserve. The tower is situated at a base elevation of 130 m above sea level (a.s.l.). A detailed description of the site, instrumentation capabilities, underlying vegetation, and nearby topography, as well as other relevant features, can be found in
Computation of turbulent quantities from tower data such as mean flow, heat fluxes, and turbulent kinetic energy was accomplished by employing time averages over 1 min intervals. We have chosen such a short time window primarily because of the nonstationary nature of the events under study, but also to avoid contamination from low-frequency, nonturbulent processes, and, therefore, guarantee that the discussion refers to turbulent quantities alone. Most of the cases analyzed occurred when stable stratification was present at the site. This choice was based on the results of
Radar data used in this study came from the operational S-band Doppler radar located in Manaus (
In this study, the selection of DMC events was accomplished by following a two-step procedure relying on Doppler radar imagery and thermodynamic and kinematic changes associated with the storms as detected at the ATTO walk-up tower. The first step consisted of subjectively inspecting radar reflectivity fields using low-elevation plan position indicators (PPIs) to identify the passage of convective storms over or near the instrumentation site. Only storms that produced detectable impacts on the evolution of meteorological variables at the tower site were selected. To that end, time series of virtual potential temperature (
Studies addressing disturbances in the tropical PBL caused by DMC activity, such as
Following the aforementioned procedure, four DMC events were selected for investigation. Dates, duration, radar characteristics, and other relevant features of the storm events are presented in Table 1. The wind speed increase in the outflow (i.e., the gust front intensity) was measured as the bulk difference between the 1 min mean wind in the pre-storm environment and the maximum wind after storm arrival. Similarly, the maximum temperature drop (i.e., the cold pool intensity) was measured as the bulk difference between the 1 min mean
Main characteristics of the four storm events investigated in this study. (*) indicates Ponta Pelada Airport operational soundings in Manaus (SBMN; lower resolution). (**) indicates soundings taken at Campina site (
Previous studies of tropical DMC–PBL interaction have demonstrated the importance of characterizing morphological aspects of the convective activity that disturbs the PBL with the aid of radar imagery (e.g., SR98;
Storms struck the walk-up tower site at different times of the day. Two events (1 and 2) took place during the late afternoon or early evening transition (EET) while event 3 occurred during late morning hours. Event 4 occurred at dawn and was the longest-lived event. These differences in the time of storm occurrence are relevant as the convective outflows interact with the PBL during distinct stages of its evolution. In the following subsections, a description of each event is presented, focusing on their radar characteristics and the intensity of the thermodynamic and kinematic effects detected at the walk-up tower site.
At approximately 17:15 LST on 31 October 2015, the Manaus Doppler radar indicated a northeast–southwest-oriented band of convective cells advancing over the eastern-northeastern Amazon as part of a larger area of intense but disorganized convective activity (Fig.
0.9
Pre-storm measurements of winds and virtual potential temperature at the walk-up tower revealed a slight tendency of decreasing turbulence and temperature typical of pre-sunset conditions. However, as the outflow from the storm cluster arrived at ATTO, a sudden drop in
After the strongest cooling associated with the convective active stage of the system, minimum
Small clusters of short-lived thunderstorms were observed by the Manaus radar near the ATTO location during the late afternoon and early evening period on 2 November 2015 (Fig.
Temporal evolution of the mean horizontal wind speed at the different vertical levels, according to legend, for
The convectively active stage of the storm over the tower lasted approximately 20 min, being considerably less than what was observed with event 1, which was associated with a much larger storm and more easily detected in the
There are some clear differences between the recovery phase inside and above the canopy, as indicated by in-canopy measurements. The levels above the canopy show full recovery after 50 min, as stated above, while, below the canopy, cooler temperatures are maintained long after the above-forest air mass had attained a new steady state. In this scenario, it seems that the forest slowed down the recovery in its interior, thus fostering the establishment of a stable stratification next to the ground. Hence, some process(es) related to upward fluxes of heat and moisture must have occurred in order to warm and moisten the layers near the top and above the canopy. The mechanisms responsible for these processes will be described in detail in Sect. 3.
Around 10:20 LST, an unorganized cluster of convective cells rapidly formed around the ATTO site at the back side of a westward-moving MCS (Fig.
The same as in Fig.
This event clearly displayed a behavior that was quite different from the other cases studied, especially in light of the gradual nature of the potential temperature (wind speed) decreases (increases). The most probable explanation for this anomalous behavior is that the arrival of the outflow from the scattered storms at the ATTO site was not preceded by a sharp gust front, as was the case for the other events. Rather, it is plausible that merging of weak outflows from the incipient or decaying storm cells generated a slow-moving cold pool that gently spread over the site. If this was the case, it is safe to state that although the PBL was disturbed by the DMC outflow, the downdraft cores of the parent cells or the strongest portion of their gust fronts did not pass directly over the instruments.
The system with the largest convective core investigated was associated with a large southwestward moving cluster of strong storms with a trailing stratiform precipitation region (Fig.
Considering only the main convective system, the convectively active period in this episode was also longer compared to the other events. An “attempt” of a recovery phase was observed as a slight increase in
The longer recovery period observed in event 4, as well as that found in event 1, is in contrast with the short recovery observed in event 2, which points to the dependence on the spatial scale of the outflow-producing system. This observation is in line with the results of
Surface heat fluxes play a major role in the initiation process of convective storms in tropical regions, as intense diurnal heating drives thermals or plumes that grow upscale into large cumulonimbus clouds. On the other hand, when convective downdrafts introduce cool air from aloft into the PBL, the evolution of surface heat fluxes may also be affected significantly. Figure
The same as in Fig.
Prior to the occurrences of events 1, 2, and 4 (Fig.
During this early stage of the DMC activity over the tower, the
The same as in Fig.
Soon after the most intense DMC perturbations stage ends, a sudden transition to a prolonged period of negative
The strong persistent negative
In summary, positive, intense
A data quality analysis for the sensible heat fluxes in events 1 and 2 is presented in the Supplement, using multiresolution cospectra.
We now turn our attention to event 3 (Fig.
Latent heat fluxes are mainly controlled by the wind magnitude and vertical humidity gradients. SR98 showed that large enhancements of LE occur in the wake of oceanic gust fronts, which can sometimes be over 300 % stronger than in pre-storm conditions. Time series of LE for our case studies are shown in Fig.
The same as in Fig.
The occurrence of such enhancements of LE as a response to convective storm downdrafts has been demonstrated in previous studies
Cool outflows from convective storms tend to be very turbulent in nature. As discussed throughout the paper, many studies have shown significant enhancements in turbulent quantities, such as heat, moisture, and momentum fluxes, during and after the occurrence of convective outflows
Probably, the simplest way to analyze the intensity of turbulence in a given flow is to compute the turbulent kinetic energy (TKE) associated with it
Figure
The same as in Fig.
TKE peaks follow closely those seen in the
Multiresolution spectra of TKE for events 1 and 2 are presented in the supplementary material, showing that the turbulence data are consistent, in spite of the precipitation in the period.
Considering event 3 (Fig.
The time evolution of atmospheric variables and scalar fluxes during the occurrence of surface outflows produced by deep convective storms in a tropical rainforest was analyzed utilizing high-frequency, multilevel measurements performed at the 80 m walk-up tower of the Amazon Tall Tower Observatory (ATTO) located in northern Brazil. Four convective outflows that passed over ATTO from late October to mid-November of 2015 were studied, with three of them occurring during the early evening transition or nighttime hours and one during the early afternoon. The evening/nocturnal events were characterized by well-defined gust fronts associated with moderate decreases in virtual potential temperature and increases in wind speed. In contrast, the early afternoon event was a weak outflow, lacking a sharp gust front and producing only a slight drop in virtual potential temperature. With the gust front arrival, positive sensible heat flux (
Schematics illustrating the effects of a gust front passage over a tall tower in the forest during nighttime hours. Top: a gust front (blue line with triangles) from a convective storm (cold (warm) colors represent low (high]) radar reflectivity values) approaches the tall tower (gray tower symbol) at t1. At t2, the gust front has passed by the tower site which now is embedded in the turbulent wake of the cold pool (light purple shaded area, with turbulent eddies represented by curly, black arrows). Bottom: corresponding sensible heat flux response to gust front passage at tower levels above (blue) and within (pink) the canopy. The dark purple arrow indicates the storm motion vector (due southwest).
All data used in this study are kept in the ATTO Databases at Instituto de Pesquisas da Amazônia and Max Planck Institut Für Chemie. The overall project description can be found at
The supplement related to this article is available online at:
MIO, AOM, and MOA developed the scientific idea of the study and project. OCA, MS, PESO, and AT took part in data collecting and analysis and provided the scientific support on micrometeorological issues. ELN provided scientific support on severe weather concepts. MIO and DVB analyzed the data. All authors contributed to the discussion and interpretation of the results.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Amazon Tall Tower Observatory (ATTO) Special Issue”. It is not associated with a conference.
This work was developed during the participation of the first author in the ATTO project (December 2015–July 2017). For the operation of the ATTO site, we acknowledge the support of the German Federal Ministry of Education and Research (BMBF; contract 01LB1001A) and the Brazilian Ministry of Science, Technology and Innovation (MCTI/FINEP; contract 01.11.01248.00) as well as the Amazon State University (UEA), FAPEAM, LBA/INPA, and SDS/CEUC/RDS-Uatumã. Alessandro Araújo, Marta Sá, and the Micrometeorology group from LBA project in Manaus have been responsible for collecting, organizing, and quality controlling part of the data used in the study. We thank the Environmental Satellite Division from Brazil's Instituto Nacional de Pesquisas Espaciais and, particularly, meteorologist Thiago Bíscaro for providing the radar data used in this research project. Finally, we are grateful to Kathleen Schiro and two anonymous reviewers for their valuable comments that greatly improved the article. This work was, in particular, supported by the Max Planck Society (MPG), the Instituto Nacional de Pesquisas da Amazônia (INPA), CNPq, and CAPES.
This research has been supported by the ATTO project (grant no. 574009/2008-6; process: 381948/2009-9) of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).
This paper was edited by Gilberto Fisch and reviewed by Kathleen Schiro and two anonymous referees.