Assessing the Tropical South Atlantic atmosphere thermodynamics under distinct Sea Surface Temperature patterns

Northeast Brazil (NEB) is a susceptible region to the occurrence of extreme rainfall events. Sea surface temperature (SST) is used as an indicator for predicting intense weather events in this region. The westernmost Tropical South Atlantic region, also called Southern Atlantic Warm Pool (SAWP), is characterized by a source of heat and humidity which creates atmospheric instability for the NEB. In June 2010 the eastern coast of NEB (ENEB)was influenced by heavy rainfall, causing flash floods and landslides. On the other hand, 2012 marked the beginning of a period of droughts that affected the whole NEB area. The SAWP temperature in turn recorded anomalous values of + 1ºC (-0.5ºC) in 2010 (2012), respectively, although in June 2012 intense rainfall was recorded in ENEB, even with intense negative SST anomalies. With the Coupled-Ocean-Atmosphere-Wave and Sediment Transport (COAWST) model, simulations were made to characterize 2010 and 2012 atmospheric conditions, modifying the SST input data in both situations. The goal of this work is to assess the meteorological systems that occurred in 2010 and 2012 using observational, reanalysis, and simulated data, as well as to identify changes in atmospheric instability patterns, which are under influences of different SST conditions. We performed four cases, including: a) SST measured in 2010 with 2010 atmospheric conditions; b) SST measured in 2012 with weather conditions of 2010; c) SST measured in 2012 with 2012 weather conditions; d) measured in2010 with atmospheric conditions from 2012. The results showed that SAWP temperature significantly influenced the instability of meteorological systems. The impacts were more significant in the lower layer of the (04-24-2020) in Recife, (02-13-2020) in João Pessoa and (01-09-2020) in Natal, which caused transport, health, human, economic, and social security problems. The results obtained in this work and others (Debortoli et al. 2017; Rodrigues et al. 2020) show that the climate for ENEB is in warning pathways, with more intense precipitations in highly vulnerable urban sites, leading to more constant social impacts cases.

3 by easterly waves over the SAWP region, but with different rainfall variability. In 2010, persistent SST anomaly up to 1ºC records preceded extreme rainfall events at ENEB. On the other hand, in 2012, negative SST anomalies were commonly recorded in the Tropical South Atlantic and were the beginning of the most prolonged periods of drought in the NEB (Silva and Guedes 2012; Marengo et al. 2017). The drought event affected regions which were not commonly affected by this severe rainfall deficit, such as ENEB.
Due to the lack of observational data, and the need for environmental management, civil defense and strategic planning, terrestrial system models fill this gap. The ocean-atmosphere coupled models allow highfrequency turbulence and radiative exchange processes between the model's components with a defined small time-step. Hypothesis tests enable us to identify climatic patterns that might influence future climates. These adverse weather scenarios should serve as a resource for urban planning to mitigate the effects of natural disasters and minimize social impacts in a situation of potential climate change.
We performed simulations with Weather Research and Forecasting (WRF) and the Regional Oceanic Model System (ROMS) in a two-way coupled model with Coupled Ocean-Atmosphere-Wave and Sediment Transport (COAWST) modeling system (Warner et al. 2008(Warner et al. , 2010Renault et al. 2012). The experiments consisted of changing the SST input data in the model, making the SST an independent variable for these simulations. We performed four cases, including i) SST measured in 2010 with 2010 atmospheric conditions; ii) SST measured in 2012 with weather conditions of 2010; iii) SST measured in 2012 with 2012 weather conditions; iv) SST measured in2010 with atmospheric conditions from 2012.
The objective of this work is to evaluate the 2010 and 2012 meteorological events in the ENEB under different SST anomalies over the SAWP region, simulating the possible impact of the SAWP on the extreme rainfall at the ENEB, regarding climate change scenarios.

Methods
The study area (50ºW -20ºW; 20ºS -5ºN) covers part of the Western South Tropical Atlantic, including the South Atlantic Warm Pool (SAWP, 34ºW -25ºW; 12.5ºS -4ºS) as well as the Eastern Northeast Brazil (ENEB,. Figure 1 shows the topography and bathymetry of the simulation domain. We performed trends in the number of rainy days in precipitation intervals for the states' capitals located in the ENEB following Greene et al. (2019), being: i) Prec. Event <1mm; ii) 1 mm <Prec. Event <30 mm; iii) 30 mm <Prec. Event <100 mm; iv) Prec. Event> 100 mm. The dataset consists of daily observed rainfall from 1979 to 2019, and was provided by INMET.
Radiosonde data from the University of Wyoming database was collected for both weather events at specific points in the ENEB. The goal is to investigate atmospheric instability during the period of the highest intensity of the meteorological systems. Thus, the radiossonde data collected for this analysis were: Fernando de Noronha b. Reanalysis and Regional Modeling In order to determine the precipitation response of the ENEB region to SAWP SSTs, we used precipitation data from the Climate Prediction Center (CPC) Global Unified Gauge-Based Analysis of Daily Precipitation (Xie et al. 2007) in the period from 1979 to 2019, as well as data from the NOAA Extended Reconstructed SST V5 (Huang et al. 2017)for the same period. We performed treatment for daily precipitation, selecting case events according to the day with or without rainfall and the intervals recorded similar to ENEB state's capital. After the registered event filter, we applied the sum of annual cases, totaling the annual rainfall number of events. After treatment, daily rainfall, as well as SST over the Tropical South Atlantic trends were performed, based on the algorithms of Greene et al. (2019), indicating the trends for spatial rainfall episode in NEB and Western Tropical Atlantic SST trends for the period 1979-2019.
For the WRF component, the initial and boundary conditions were from the NCEP Final Operational Global Analysis Data (FNL) (NCEP 2000), with a horizontal resolution of 1.0° x 1.0°. For the ROMS model, the variables of temperature, salinity, and currents (u, v) for the oceanic model were given by the Hybrid Coordinate Ocean Model (HYCOM), with a horizontal resolution of 0.08° x 0.08° (Wallcraft et al. 2008 The climate projections data are from the fifth and sixth phase of the Coupled Model Intercomparison Project (CMIP5 and CMIP6) on the World Climate Research Program (WCRP) platform. It used the Representative Concentration Pathways (RCP) 4.5 for climate scenario, corresponding to the radiative forces of +4.5 W.m -2 until the end of the 21st century, as well as several climatic models consisting of boundary conditions (r), physical of the model (i) as well as disturbances in the initial state of the data (p). The same was valid for CMIP6, using the SSP245 experiment, an update of RCP4.5 project. The variables used consisted of projected SST, as well as monthly precipitation, using climatology for (a) 2020 to 2040, (b) 2040 to 2070, and (c) 2070 to 2100, from SAWP area as well as from ENEB area.
The analysis in this work was performed analysis from CMIP5 and CMIP6 multi-model experiments. It performed collections of the simulated values and the respective intervals of maximum and minimum of the simulations, leading to an analysis of the probable scenario. The chosen experiment for projections characterizes the "middle of the road," that is, a scenario that is neither as optimistic nor pessimistic.

a. Climate Overview
The ENEB is a narrow coastal strip that differsin terms of biome and annual total rainfall from the neighbor semiarid areas of the NEB. The predominant characteristic of ENEB region is the milder climate, with smaller thermal amplitudes (on average) and a growing topography towards the continent, besides holding the most populous municipalities located in the coastal areas.The maximum precipitation occurs in June, duringthe rainy season, between April and July, mainly due to the easterly waves disturbances. The rainy season is well correlated with high SST in the SAWP region with a lag of approximately three months, as seen in Figure  The rainfall regime in the ENEB responds to the SST both in the Atlantic and Pacific oceans (Moura and Shukla 1981;Kayano 2006, 2007;Silva and Guedes 2012;Silva et al. 2018;Moura et al. 2020). The release of latent heat, water vapor transport, and moisture convergence ispositively correlated tothe rainfall regime of the ENEB (Foltz and McPhaden 2006). The most important meteorological systems in the period, such as the EWD and disturbances in the trade winds, also respond directly to the SST patterns (Moura et al. 2009;Silva et al. 2018).
The definition of warm pools considers SST values exceeding 28.5ºC in the near closed oceanic region, weak trade winds, resultant deep convection, and has a distinct seasonal cycle (Wang et al. 2006). The SST climatology for the SAWP region shows that such values occur predominantly in March. However, in positively anomaly years, the SAWP may appear months before and vanish months later (e.g., 2010 and 2011 years), in the same way, in negatively years, the SAWP may not even be characterized, as between 2012 and 2018 years.
The SAWP has an important influence on the meteorological events that occur in the region. During meteorological events susceptible to rain, the warm pools intensifythe latent heat flow, buoyancy and instabilities at low levels, which increases the probability of extreme events. Figure 3 shows the coherence in wavelet between SAWP Area SST and the total monthly ENEB rainfall. The results show a high lagged coherence in all years in the period between 8 to 16 months. Also, in the period near to 32 months, we record a high coherence between 1990 to 1996 and 2007 to 2012. The result shown by the wavelets coherence indicates the lagged linear relationship between SST and the region's precipitations. With this, it can project that the increase in SST intensity in this area could lead to more significant volumes of rainfall in coastal regions in the ENEB.   Figure 4a shows an increase daily trends without rainfall for all state capitals in the ENEB. However, the occurrence of rainfall events of intervals between 1 and 30 mm.day -1 and from 30 to 100 mm.day -1 , in general, are decreased, except for Natal, where rainfall events above 30mm.dia -1 increased. For rainfall above 100mm.day -1 , there is an increase in the capitals at north of the ENEB (Natal, João Pessoa, and Recife) and a decrease in Sergipe, Salvador and Maceió ( Figure 4d). 8 FIG. 4. Normalized trends for: days without rainfall (a), rainfall between 1mm to 30mm (b), rainfall between 30mm to 100mm (c), and above 100mm (d). The lines represent the rainfall trends for ENEB state capitals, from INMET rainfall gauges from 1979 to 2019. Figure 5 shows the spatial trend of daily NEB rainfall based on GCPC dataset (1979 -2019). Several regions have showed trends of increasing rainfall above 30mm.day -1 and 100 mm.day -1 . These results agree with previous results, which showed more significant periods with a lack of rainfall and increasing intense rainfall events in Northeastern Brazil, corroborating with climate projections (Alves et al. 2021). 9 FIG. 5. Spatial trends for: days without rainfall (a), daily rainfall between 1mm to 30mm (b), daily rainfall between 30mm to 100mm (c), and above 100mm. The data was provided by Global Daily Unified Gauge-Based Analysis of Precipitation Project from 1979 to 2019. The contour lines represent the neutral trend, showing the transition between positive trend (red) and negative trend (blue). Positive (negative) trend means that quantity of days, with precipitation in a specific threshold, are increasing (decreasing). The dotted area is the 95% significance.
Similarly, Figure 6 shows the spatial trend of SST anomaly in the WTA, using monthly data from NOAA V3 reanalysis, 2 x 2 of spatial resolution. The northernmost WTA region is warming, with a rate of 0.03ºC per year, from 1979 to 2019. However, the region over the southern branch of the South Equatorial Current (sSEC) has a cooling trend up to 0.05ºC. The SST trends show systematic warming of the regions adjacent to the coast of the NEB, including the SAWP region, which is favorable to more intense precipitation regime and extreme events at the ENEB.
FIG. 6. Sea surface temperature trends in the Western Tropical South Atlantic. The period comprehend data from 1979 to 2019 and the trend is the total for period. The dotted area is the 95% significance.
The analysis for RCP4.5 and SSP245 projections shows a significant increase in the SAWP SSTs for the next decades ( Figure 7). The SST until 2040 will increase above 28.5ºC remaining throughout the year. Over the decades, the values of 28.5ºC have continually exceed the threshold, having only averaged values below in the decade from 2040 to 2070, during the austral winter. In the decades from 2070 to 2100, even the minimum threshold exceeds the values of 28.5º in the SAWP Area. Therefore, the climatic characteristics of the STA must undergo a significant change directly influencing the ENEB rainfall climatology.  Wang et al. (2006). The 2010 SST, with positive anomalies about 1ºC, is equivalent to the temperatures for the decade of 2020 to 2040; which means, the COAWST simulations with SAWP can represent a real specific scenario of climate change that might be already influencing us.
In 2010 (Figure 8a), there was a southeast wind flow with a magnitude of about 15 m.s -1 over the near ocean coast, presenting slight disturbances at 850 hPa. At 700 hPa (Figure 8c), a trough is observed over the SAWP region, indicating a wave disturbance that moved from the ocean towards the continent. Then, at 500 hPa ( Figure  8e), the high-intensity eastward flow was associated with a pre-existing post-frontal high pressure, leading to a large amount of water vapor transport also in middle levels of the atmosphere.
In counterpart, in 2012, it was observed southeastern disturbances with weaker winds magnitudes at 850 hPa The meteorological systems occurred in 2010 and 2012 had similar wave disturbances in atmosphere lowlevels. The simulated period comprised the months with the highest accumulated rainfall for ENEB (May and June). The EWDs are the weather systems with the largest impact on rainfall accumulation in the region during austral winter (Ramos 1975;Torres and Ferreira 2011;Gomes 2012;Kouadio et al. 2012). During 2010, the meteorological event between June 17 th and 18 th presented low-level instabilities, reaching 700 hPa, and a high pressure in the South Atlantic, which favored the strong easterly winds in 500 hPa. This circulation pattern favors the transport of water vapor from adjacent regions of the Southern Tropical Atlantic to the continent (Cavalcanti et al. 2002), supporting atmospheric convections (Brown and Zhang 1997). However, on June 14 th , 2012, the atmosphere showed disturbances at low-levels, but a low wind magnitude with a ridge circulation at 700 hPa, while at 500 hPa an eastward flow, showing water vapor transport from the continent to the ocean (Cavalcanti et al. 2002).
The results from radiosondes showed dry atmospheric conditions from mid to high layers. However, there were high humidity patterns at the Fernando de Noronha and Recife surroundings in 2010, with strong subsidence in Salvador, the southernmost ENEB locality. This pattern shows that the weather system was restricted in the northernmost part of ENEB. The radiosonde for Fernando de Noronha shows a wet layer from the surface near to 700 hPa level, as well as in Recife.
For 2012 radiosondes, the results showed the presence of a significant amount of humidity restricted to the 800 hPa level for Fernando de Noronha and Natal (located between Recife and Fernando de Noronha). However, in Recife, there was a similar characteristic compared to 2010, with low humidity at 700 hPa and a moisten layer above 700hPa and 600 hPa. The instability index values coming from the radiosonde at 06/17/2010 12Z measurements allow the comparison of the probability for atmospheric conditions for deep convection. In 2010, the values indicated median K index values, as well as CAPE and Total Totals Index values, but a large amount of precipitable water mainly in Fernando de Noronha and Recife. It is noteworthy that the radiosonde shown is that recorded on the 17 June at 12Z, while the most substantial convection occurred on the 18th at 00Z, which may underestimate the index values.
In 2012, there are low values of all instability indices, but a large amount of precipitable water, near to 2010 measured, except for Fernando de Noronha which presented a difference of 8mm between 2010 and 2012. Radiosonde measurements for 14 June 2012 coincide with the period of more activity weather systems, showing the atmospheric conditions during the event.
When comparingthe instability indices between the years studied, it is observed that the values in 2010 presented higher intensification for convective buoyancy than in the 2012 event. Thus, the phenomena differs in the cloud formation characteristics, lying deep convective clouds in 2010 ( Figure 10a) and shallow convection (Figure 10b) according to the GOES satellite image. Figure 10a shows the GOES IR satellite images with cloud top temperature enhancement, where the colder values show intense vertical cloud development due to the weather system. It is justified by the high mid-level's atmosphere instability, a through configuration at 700 hPa and a large amount of water vapor transport to the region due to the high-intensity wind speed from ocean to land.
However, Figure 10b shows that there was no intense vertical development. The pattern of satellite images with grey tones of clouds indicate less convection activity than the 2010 weather system. It distinguishes from the instability indices, which were lower in 2012, as well as the precipitable water and water vapor transport, as the winds were less intense, including a ridge at 700 hPa and a eastward flow at 500hPa. Although the characteristics of warm convective clouds, in the event of 2012 the precipitation reached the level of 100 mm.day -1 in isolated areas of the Metropolitan Recife Region. The radiosonde data complemented the analysis by showing the profiles of the atmosphere during the systems. Atmospheric instabilities associated with the June 2010 event had higher values than June 2012, indicating an increased probability of occurrence of convective events in the region. The atmospheric instabilities in 2010 culminated in an intense convective event with large vertical development that can be seen in satellite imagery. Temperature enhancement at the top of the cloud indicates the presence of a mesoscale convective system acting during the peak of the meteorological system's development. However, according to the synoptic analyzes and radiosonde measurements, cloud development over the period of June 14 th , 2012 had no intense convective activities. Cloud characteristics indicate the development of shallow cumulus clouds with slightly colder small cores. However, warm rainfall (Liu and Zipser 2009) were observed in ENEB mainly in Recife weather stations, which record values close to 100 mm.day -1 .

c. Simulations of the 2010 and 2012 Events
An average over the SAWP Area (34ºW -25ºW; 12.5ºS -4ºS) was performed to calculate the difference between SST and air temperature at 2m (T 2m ) (Figure 11a The warm pools are either created or vanished according to atmospheric characteristics. A dry and cold atmosphere over a warm SST, for example, is associated with increased latent heat flux from ocean to atmosphere, generating the decease of the warm pool. The opposite occurs in a wet and warm atmosphere, when the latent heat flux losses through evaporation are weakened, and the warm pool intensifies. Hence, process as cloudiness, convective activity and outgoing longwave radiation, are important factor for warm pool formation, maintenance and demise (Wang and Enfield 2003;Clement et al. 2005).
The warm SST maintenance is associated with the warmer and wetter troposphere, reducing not only sea level pressure (SLP) and weaker easterly winds, but also lowering vertical shear and weakening subsidence in the air. During years with a well-developed warm pool, these conditions lead to wetter summers in the Intra American Sea (IAS), Central America (Wang and Enfield 2003) and rainfall regimes in the Western Pacific (Brown and Zhang 1997). The opposite also occurred (Brown and Zhang 1997;Wang and Enfield 2003), with cold water generating dry summers in these regions.
The cooling pattern of the TSA western basin is due to latent heat losses (Foltz and McPhaden 2006) and coincides with the highest rainfall period at ENEB (Cintra et al. 2015), which are also directly related to rainfall anomalies in these areas (Moura and Shukla 1981;Rao et al. 1993;Aragão 1998;Kayano 2006, 2007;Moura et al. 2009;Silva and Guedes 2012;Silva et al. 2018). Figure 11e,f shows the averaged atmospheric surface pressure for the SAWP Area. The simulated surface pressure presented different patterns, responding to different SST inputs (2010 and 2012). Figure 11e shows the surface pressure for 2010 simulations. The results evidence a lower average surface pressure at beginning of the simulation periods with an increasing from the middle to the end of simulations. In the period with activity of the meteorological system (16 June 2010), the surface pressure decreased. In 2012 ( Figure 11f) the surface pressure showed a pattern with weaker variability, and a decreasing surface pressure between 10 th and 15 th June. Comparing 2010 with 2012, the lower pressure preceeded the rainfall events over ENEB. Also, the simulations shows slightly low values when using high SST input. The ENEB virtual temperature profile (Figure 12a,b) were plotted to identify MABL thermal instabilities during both weather events (18 June 2010 and 14 June 2012). The results showed warmer MABL when holding the 2010 SST as a boundary condition. Differently, cold SST conditions simulated a small warming of the lower vertical profile. Thus, under high SST conditions, the lower troposphere increases buoyancy at low levels. Figure 12c,d shows the vertical motions based on omega for the case studies. Omega (ω) represents the rate of change of pressure in a parcel over time (dp/dt), which is proportional to the vertical motion under hydrostatic balance. There were upward movements for all simulations in low-mid levels. The highest values were at low levels up to around 850 hPa for all scenarios. However, the most intense upward movement occurred when the model had the 2010 SST as input. It is noteworthy that, at medium-high levels, the vertical movements were close to zero. However, the simulation of 2012 under the SST of 2010 as boundary conditions, gived rise to upward movements throughout the atmosphere. The zonal and meridional water vapor transport over ENEB shows the moisture convergence from the ocean. When observing the zonal transport (Figure 13a), it is possible to identify a displacement from east to west at low-mid levels, with higher intensity of around 850 hPa for simulations of both 2010 and 2012 atmosphere conditions. However, the most significant transport occurred in 2010 with higher SST, without any considerable alteration for both simulations of 2012 patterns. The meridional water vapor transport (Figure 13b) showed higher values in lower layers of the atmosphere. In the simulations of 2010, a slight increase in the vapor transport with the SST of 2012 could be verified. However, meridional transport values had a significant increase over warmer SST. The result indicates that in 2012, the highest water vapor transport contribution was from the southernmost regions of the Atlantic, which means the higher SST in the south contributed to the increase of moisture in the ENEB.
The water vapor content in the atmospheric layer is considerably increased in simulations with warmer SSTs (Figure 13c). The vapor content in the ENEB region has a close relationship with the moisture transport from the South Atlantic, where warmest waters favored the increase or the water vapor content. Both 2010 runned simulations presented higher water vapor content than the 2012 simulations (Figure 13d). This is a response to the higher SSTs reported in 2010, favoring more storage of water vapor in the lower atmosphere. Therefore, the result implies that the warmer environment, with large transport of moisture, will generate a region with higher water vapor content, leading to higher volumes of rainfall. Thus the simulations with the warmer SSTs favored higher precipitation, which leads to significant increases in risks of extreme eventsin the ENEB region. The simulations for accumulated rainfall (Figure 14) during the day of the event showed that the SST of 2010 contributed to intensify the total precipitation. In contrast, the SST of 2012 inhibited significant accumulated rainfall during the events. As a result, the influences of the SAWP simulated with the 2010 SST conditions indicate that even meteorological systems with the weak potential to cause heavy rainfall, such as the 2012 event, may lead to urban and social adversities. The SAWP SST played a crucial role in intensifying pre-existing weather systems. However, studies like those of Gomes et al. (2015) and Silva et al. (2018) showed that SST does not directly cause meteorological systems, which shows the non-linearity of SST on the formation of disturbances, although it intensifies those extreme rainfall events. The results show that the heat and humidity losses to the atmosphere were larger when there were higher SSTs. However, the loss of temperature from the ocean to the atmosphere was faster in the 2012 run with 2010 SSTs. It indicated that the atmosphere in 2012 favored a stronger cooling of the surface ocean waters, while in 2010 the temperature loss was weaker, which shows that the maintenance of high SST anomalies in 2010 was due to the air-sea dynamics as suggested by (Wang and Enfield 2003;Clement et al. 2005).
In addition, the lower surface pressure was associated with a higher SST, agreeing with Wang and Enfield (2001). The high SAWP SST values also favored a warmer atmosphere over ENEB coast than those with colder SST. Also, the SST changes directly influenced the lower layers of the atmosphere, increasing the vertical velocity. The 2012 omega with the 2010 SSTs showed a higher upward velocity compared to the 2012 run with 2012 SSTs, including upward movement throughout the atmosphere. In the 2010 run with 2012 SSTs, the upward velocity decreased, reducing the deep convection.
The results confirm the presence of higher instabilities when SAWP is present. It suggests that the oceanic waters in the SAWP region are dominant drivers in the intensification of the meteorological phenomena over the ENEB. The results also suggest that SAWP acted similar to other warm pools in other oceanic regions, as shown in works such as Brown and Zhang (1997), Enfield (2001, 2003), Wang et al. (2006) and Wang and Lee (2007). The increase in SST favored the warming of the atmosphere at low levels by the sensible and latent heat fluxes, increasing the vertical motion, and increasing the instability of the atmosphere.

20
The assessment of SAWP SSTs impacts on the extreme rainfall events on ENEB is a crucial factor to future decision-making policies in mitigating vulnerabilities to climate change. Even though the region is a narrow coastal strip, the number of inhabitants sheltered in this region is high. In fact, the total of the metropolitan regions of the ENEB combined is close to 13.1 million, which corresponds to approximately 25% of the total population of NEB (IGBE, 2019).
Climate projections expect a substantial increase in ocean temperatures in the SAWP region. In fact, the SST over Atlantic regions shows that SAWP only occurs in a short period between February and April, or when positive anomalies are higher, reaching from January to May, and with the decrease starting from late May. However, the climate projections show that SAWP regions are shifted from being a possible appearance, to become a constant presence in the near-future scenario, and usual act in a long-future scenario.
As shown in previous results, the SAWP region leads to an increase in convective activity over the east coast of the NEB. Even when the synoptic conditions are not favorable to precipitation, as in 2012, positive SST anomalies in the SAWP region results in rainfall increase over the ENEB. Negative anomalies simulated a reduction in the total precipitation. Therefore, it might be expected an increase in dynamical behavior that implies in a high amount of rainfall intensity. However, the results found in CPC Global Daily Unified Gauge-Based Analysis of Precipitation and measured data trends also show an increase in days without rain. This result is also worrisome due to the hydrological cycle in populous regions. The water supply in the metropolises may be impaired due to increased population and decreased total accumulated rainfall.
The relationship found in the COAWST simulation results, between SAWP SST and extreme rainfall over the ENEB, shows a higher potential virtual temperature, higher omega and stronger latent and sensible heat loss to the atmosphere. The key factor for increasing probability for an extreme rainfall event is the capability of the atmosphere in retaining more water vapor in a warmer environment, which can be represented by the potential virtual temperature. Furthermore, the positive feedback effects can contribute to the warming of the SAWP and the deepening of the ocean's surface layers, also increasing the amount of heat from the ocean's surface in these areas.
The air-sea thermodynamics also explains that the more significant warming of the atmosphere, a high amount of water vapor near the surface, and the decrease in surface wind speed, may inhibit the loss of sensible and latent heat fluxes. During periods when ocean-atmosphere temperature gradient are small, sensible and latent heat fluxes decrease causes an increase in SAWP SST. However, when gradient becomes higher, implying in a higher latent and sensitive heat loss, which is most present in projected scenarios than in the current patterns, it suggests more activity of SAWP for the ENEB rainfalls.
Moreover, climate models coarser resolutions could lead to a miss characterization of the future rainfall climatology in the east of the Brazilian northeast, essentially by the ENEB held in a narrow coastal strip with strong oceanic influence in the rains. Also, the general meteorological systems in the rainy season, as for example the easterlies wave disturbances showed in this work, are phenomena with a mesoscale extension which could be not well resolved with climate models with coarse resolutions. Finally, we suggest for future work a dynamical downscaling of climate models to understand the dynamical impact of climate projections scenarios in a highresolution overview.

Conclusion
This work evaluates the 2010 and 2012 rainfall events in the ENEB under different SST anomalies over the South Atlantic Warm Pool region. Through air-sea coupled simulations and climate projection scenarios, it was possible to quantify the anomalous warm SST role during two events and under climate change scenarios.
This work reveals a shift in rainfall regime climatology at the eastern part of the Brazilian Northeast, characterized by an increase in days without rainfall and the frequency of intense precipitation above 100 mm.day-1. At the same time, there is a decrease in weak rainfall < 30 mm.day-1 occurrence. The atmospheric instabilities at the ENEB, consequently the extreme rainfall are forced by the positive anomalous SAWP temperature.
Moreover, this study indicates that the SAWP temperature significantly influences the instability of meteorological systems in the ENEB. The impacts were more significant in the lower atmosphere, especially in the variables that lead to low-level instabilities. Also, the atmospheric conditions favor the ocean environment to keep warmer, retaining the unstable conditions over the SAWP. The COAWST coupled-model simulations evidence that meteorological events that present a small potential for rainfall events, as in 2012, can cause extreme rainfall events when combined with anomalous warming conditions of the SAWP, due to this instability potential induced by this region.
Climate projections indicate warm trends over the SAWP region. According to CMPI5 and CMPI6, this warming can lead from a condition of probable SAWP formation observed today to a state of permanent SAWP from 2070 to 2100, considering the models projected the "middle-of-road".
The climate variability in Brazilian Northeastern is strongly related to anomalies in the Pacific and Atlantic oceans. Further, 25% (about 13.5 million) of the population of this region are at the state capitals in the ENEB, which also presents high social inequality and susceptibility to natural disasters, such as flash floods, landslides, and sea-level rise. Therefore, it is essential to evaluate the climate change impacts concerning atmospheric events occurring in the entire Brazilian Northeast. Due to the ENEB inherent vulnerabilities, we highlight that it is crucial to formulate and guarantee public policy to deal with the impacts of these extreme events in this area, especially under this climate change scenario, as indicate the future projections for this region.