A meteorological overview of the ORACLES (ObseRvations of Aerosols above CLouds and their intEractionS) campaign over the southeast Atlantic during 2016-2018: Part 1-Climatology

In 2016–2018, the ObsErvation of Aerosols above CLouds and their intEractionS (ORACLES) project undertook three month-long deployments to the Southeast (SE) Atlantic Ocean using research aircraft to better understand the impact of biomass burning (BB) aerosol transport to the SE Atlantic Ocean on climate. In this Part 1 of the meteorological overview paper, the climatological features at monthly time scale are investigated. The southern African easterly jet (AEJ-S), defined as 20 the zonal easterlies over 600–700 hPa exceeding 6 m s around 5–15° S, is a characteristic feature of the mid-level circulation over southern Africa that was also during the deployment months of August 2017, September 2016, and October 2018. Climatologically, the AEJ-S develops at lower altitudes (~3 km, 700 hPa) between 5–10° S in August, while it develops at around 4 km (~600 hPa) and further south (5–15° S) in September and October, largely driven by the strong sensible heating over the African plateau. 25


Introduction
The southeast (SE) Atlantic and the west coast of southern Africa is one of the key regions of the globe for understanding the interactions between the Earth's climate, weather, and pollution. It is characterized by a stratocumulus cloud deck associated with strong large-scale subsidence and the anticyclonic circulation of the semi-permanent south Atlantic sea level pressure 45 high (Klein and Hartman, 1993;Wood, 2015). The low-level stratocumulus clouds increase the net amount of outgoing radiation at the top of the atmosphere (TOA), inducing a negative radiative effect (a cooling).
The southern African Easterly Jet (AEJ-S) in the Southern Hemisphere (SH), one of two jets dominating the midtropospheric circulation over Africa, is an effective carrier of aerosols (Adebiyi and Zuidema, 2016). Adebiyi and Zuidema (2016) showed that about 55 percent of the biomass burning (BB) aerosols from southern Africa during September-October 50 is transported westward by the AEJ-S to the southern tropical Atlantic and beyond, and the remaining BB aerosols are either carried northwestward into the intertropical convergence zone or returned toward southern Africa. Additionally, a strong lowlevel wind, known as the "Benguela low-level jet" (Nicholson, 2010, hereafter LLJ), is also one of the characteristic features of the SE Atlantic circulation, which is related to the strength and location of the subtropical high (Nicholson, 2010). characteristic features over that region. The distinct difference in elevation between the Congo-Zaire basin (north of 10° S) and the Namibia-Kalahari dryland (south of 10° S, 15-21° S) is evident (Fig. 1a). At mid-levels (~600 hPa) during August-elevation of the coastal cloud-topped boundary layer. The meridional gradient in 600-850 hPa geopotential thickness (i.e. layer mean temperature) over the continent is large at ~10° S (Fig. 1c). This meridional temperature gradient over the land, sustained by a "heat low" over the Namibia-Kalahari dryland, is the dominant driver of the AEJ-S. After the onset of the rainy season around the end of October, precipitation reduces the local temperature gradient, and consequently weakens the AEJ-S (not shown). 75 To characterize the interaction between aerosols and clouds, the ORACLES (ObseRvation of Aerosols above Clouds and their IntEractionS) field deployments took place during 2016-2018 over the southeastern Atlantic Ocean immediately to the west of the southern African continent. The goal was to develop an understanding of the impacts of southern African BB 85 aerosol transport over the Atlantic Ocean on climate . Collaborative international deployment activities over the SE Atlantic such as U.K. CLARIFY 1 (September 2016, 16 August 2017-7 September 2017Haywood et al., 2021), DOE LASIC 2 (1 June 2016-31 October 2017; Zuidema et al., 2018), and the French AEROCLO-sA 3 (22 August 2017-12 September 2017; Formenti et al., 2019) have also advanced the understanding of aerosols and their interaction with clouds. A few results from ORACLES have elucidated the observed details of aerosol-cloud interactions (Kacarab et al., 2020;Gupta et 90 al., 2021), the combined direct aerosol radiative effect (Cochrane et al., 2020), and the impact of moisture outflow on midlevel clouds (Adebiyi et al., 2020;Pistone et al., 2021). However, while those studies highlight the detailed features of aerosolcloud interaction during the deployment period, they cannot tell us whether the observations were typical. Hence, understanding the meteorological characteristics during the ORACLES deployment, and how different they are compared to the climatological mean in various temporal and spatial scales is critical. This paper focuses on the climatological overview of 95 the meteorology reflecting the coupled land-ocean-atmosphere system and the representativeness of the deployment months.
Aerosol-cloud interactions are modulated by meteorology. For example, cloud cover changes with LTS, which is modified not only by surface temperature but also by absorption of solar radiation by aerosols residing above the cloud over the ocean in the African region (Gordon et al., 2018;Mallet et al., 2019;. Other recent studies also detail how the large-scale flow interacts with the entrainment of smoke into the boundary layer (Diamond et al., 2018;Zhang and Zuidema, 2019;Abel et al., 100 2020). Thus, it is important to identify the direct impact of the prevailing circulation on BB aerosol transport and stratocumulus decks, and to separate the meteorological impact on the stratocumulus deck from the aerosol impact on stratocumulus during the ORACLES deployment period.
The goal of this study is to describe the meteorological characteristics that directly impact aerosols and low clouds, particularly stratocumulus decks during the ORACLES campaign. In this Part 1 of the meteorological overview paper, we 105 focus on the climatological characteristics of meteorological variables during the deployment compared to the climatological mean at monthly time scale. To aid in the interpretation of airborne measurement during flight days in a more detailed manner, the key meteorological characteristics during the flight days at daily to weekly time scales will be separately presented in Part 2 of the meteorological overview paper. Atmosphere Monitoring Service (CAMS) reanalysis behaves along with the meteorological variables during the deployment months is discussed. The possible reasons for the different AEJ-S characteristics during the deployment month, particularly in 115 August 2017 are given in section 3.4. Finally, the summary and the conclusions are provided in section 4.

Data
The geographic domain of our study region is the SE Atlantic and southern Africa (30° S-5° N, 20° W-20° E) as shown in Fig. 1(a). Data and methods used to complete the relevant fields used in this study are described below. 120 • Meteorological fields such as 3-D wind (u, v, ω), temperature, geopotential height (Z), specific humidity (q), divergence, and potential vorticity (PV) come from the European Centre from Medium-Range Weather Forecasts (ERA-5, Hersbach et al., 2020). The analysis is primarily based on monthly-mean data, available on a 0.25 ° longitude x 0.25 ° latitude grid with 37 vertical levels ranging from 1000 hPa to 1 hPa. It is noted when hourly data is used in the analysis. The anomaly fields are computed by subtracting the climatological monthly-mean values (2000-2018) 125 from each monthly-mean value. Note that the monthly mean ERA5 BLH data are calculated based on the bulk Richardson number (ERA5 data description document).
• ERA5's depiction of the AEJ-S and its magnitude are also compared to those of the Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2, Gelaro et al., 2017), the Japanese 55-year Reanalysis (JRA55; Japan Meteorological Agency/Japan, 2013), and the NCEP/NCAR reanalaysis (Kalnay et al., 1996) over the 130 same time period using monthly-mean data. MERRA2 data are based on 0.6 ° longitude x 0.5 ° latitude with 42 pressure levels, and JRA55 data are based on 2.5 ° longitude x 2.5 ° latitude with pressure levels ranging 27 to 37. NCEP/NCAR data is based on 2.5 ° longitude x 2.5 ° latitude with 17 pressure levels. These comparison plots are provided in the supplementary materials.
• The Kalahari heat low is defined by the geopotential height thickness between 850 hPa and 600 hPa over a south 135 African plateau. We chose these levels rather than lower levels (e.g., the 700-925 hPa difference as used by Knippertz et al. (2017)) since the 925 hPa level is below the ground for much of our study region of interest. The sensitivity of the strength of the heat low during the deployment to the precise choice of levels is minimal.
• The low-level tropospheric stability (LTS) is defined as the potential temperature (θ) difference between 800 and 1000 hPa, below the aerosol layer at 700 hPa, following Adebiyi and Zuidema (2016). 140 • AEJ-S is defined as the zonal winds (zonal winds < -6 m s -1 , implying easterly winds) around 0-20° E, 5-15° S at 600 hPa for September and October and those at 700 hPa for August.
• LLJ is defined by a 925 hPa horizontal wind speed in excess of 5 m s -1 off of the coast of Namibia (0-10° E, 15-25° S).
• Monthly-mean Tropical Rainfall Measuring Mission (TRMM) product 3B43 with 0.25 ° grid spacing (Huffman et al. 2007) is used to characterize precipitation. This dataset is a combination of space-borne radar, microwave, and infrared channels with monthly calibration with surface rain gauges when available. Similar results were obtained 150 using monthly Global Precipitation Mission data (not shown).
• The Level 3 monthly cloud fraction product from the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard both Terra and Aqua (1° grid resolution) is used to calculate monthly mean low cloud fractions. The low cloud is defined when cloud-top height is below 2.5 km. Like Cermak et al. (2009), an extensive analysis to separate the low clouds into the detailed cloud types such as stratocumulus, stratus, and fog, will be desirable. However, 155 considering that the fraction of annual low cloud cover due to stratocumulus over the SE Atlantic Ocean is larger than 70% with a peak of 90% (Wood, 2012), we assume that the low clouds represent stratocumulus clouds in this study.
• ECMWF CAMS global reanalysis (EAC4) monthly mean data on a 0.75 ° longitude x 0.75 ° latitude grid with 25 160 vertical levels are analyzed to investigate aerosol transport features. Data during 2003-2020 are used to compute the climatological mean.

Seasonal mean and variability of the synoptic-scale circulation
We first examine the climatological mean and variability of the key meteorological factors directly affecting clouds and 165 aerosols during the ORACLES deployment. overlaid by the zonal wind, with the AEJ-S highlighted, for both the climatological mean and the individual deployment months for each of the three years. It is evident that there is month-to-month variability of the aerosol and cloud top height and their separation. The vertical extent of the aerosol layer and the depth of the separation layer between cloud and aerosols (Figs. 2(ac)) appear to be tied to the vertical extent of AEJ-S (Figs. 2(d-f)). That is, as the height of the AEJ-S increases from August to October, the aerosol layer top heights, and separation between the aerosol layer and the underlying cloud, increase as well. 180 The aerosol top height occurs around 4.5 km averaged over SE Atlantic in September, and is similar to the heights of the highest RH (> 70 %) and maximum AEJ-S wind speed, indicating that the large-scale circulation can directly affect local aerosol fields.
The AEJ-S core is located at near 3 km altitude in August, and 4 km altitude in September and October at around 8-10° S (Figs. 2(g-i)). The enhanced RH extends up to ~ 6 km just offshore at 10° E (Figs. 2(d-f)). The southerly LLJ off the Namibian 185 coast is also seen (~ 1-2 km, white wind vectors). The individual deployment months mimic the climatological mean values of RH and the zonal winds to some extent, but there are some differences. For example, an additional anomalously strong upper-level jet was observed at 6.5 km (5° S-15° N) during August 2017 (Fig. 2d). This jet is enhanced over a relatively dry region (RH < 30 %). The mid-latitude "dry tongue" at 1-2 km penetrates northward to around 20-10° S (Figs. 2(g-i)), with a high RH plume aloft at 3-5 km due to the AEJ-S. The reduced moisture at 1-2 km is tied to anomalous northward advection 190 of dry air originating from the southern oceans, while the dry air above 5 km reflects the Hadley circulation-driven large-scale subsidence (Wood, 2012;Myers and Norris, 2013;Adebiyi et al., 2015). The mid-level dryness above 4 km is stronger in August 2017 compared to the climatological mean (Fig. 2g). The dry intrusion along with southwesterly wind and moist plume above and south of 10° S during the deployment are both similar to their climatological mean in September 2016, while the free-tropospheric drying is reduced in October 2018 compared to the October climatology (Figs. 2(g-i)). The anomaly fields 195 in RH and horizontal wind speed show the features described above, which are provided in Fig. 1S in the supplementary material. Spatial features of the AEJ-S and RH are shown for the individual deployment months (August, September, and October) in Fig. 3. Since the core of AEJ-S in August is lower than in September and October (Fig. 2), the wind and RH are shown at 700 205 hPa (~ 3 km) for August. The southward progression of the regions of significant RH and continental rainfall is clearly apparent, as well as the strengthening of the easterly jet from August to September and October. The AEJ-S maximum wind speed is largest in September, confined over the coastal region over 5-15° S. The maximum wind speed is weaker in October than that in September, but the jet extends further westward over the tropical Atlantic Ocean. Note that the "recirculation" pattern (shown by reversing wind vectors around 15° W-10° E, 10-25° S) is present for all three months but is only about half 210 as strong in August as in the other two months. The zonal extent of the recirculation pattern appears to be associated with the strength of AEJ-S: the stronger the AEJ-S, the larger the radius of recirculation. The AEJ-S also exhibits substantial year-toyear variability. The distribution of RH, which is tied to the southward extent of precipitation, also varies from year to year.

Meteorological characteristics associated with the AEJ-S
The AEJ-S in August 2017 is significantly weaker than the climatological mean. The AEJ-S in September 2016 is similar to the climatological mean. Inland precipitation is also pronounced around 5° N, penetrating further south, with being wetter 215 in September 2016 compared to the climatological mean. In October 2018, AEJ-S is slightly weaker than the climatological mean around 0-10° E, 5-10° S. The weaker AEJ-S in August 2017 is also evident in the anomaly fields in Fig. 4. The horizontal wind speed is significantly weaker than the climatological mean in August 2017, and similar features were found when using zonal wind speed, reflecting 225 the weakening of AEJ-S (not shown). The region around 0-15° E, 5-10° S is drier in August 2017 than the August climatology ( Fig. 4a). The horizontal wind speed in September 2016 is similar to the climatological mean with more rain penetrating further south (Fig. 4b). In October 2018, the horizontal wind speed is slightly weaker than the climatological mean around 15° W-These meteorological features, including the location and intensity of the AEJ-S in ERA-5 during the deployment months, are also well observed in the other reanalyses. For instance, the weakening of AEJ-S in August 2017 is observed in other reanalysis data, such as ERA-interim (not shown), MERRA2, and JRA55, although the difference is small in JRA55 (see Figs. 2S and 3S in the supplementary material). However, the NCEP/NCAR reanalysis shows the greatest differences compared to the other three reanalyses during August-October. This may be related to the differences in the location and 235 intensity of the large-scale subsidence and the lack of enhancement in the local upper-level wind over the 5-10° S region during August 2017 in NCEP/NCAR.

250
To better understand how the AEJ-S is associated with RH, the heat low over the continent is examined in more detail in Fig.   5. The heat low is strongest in September and October, centered over 12-20° E between 10-20° S, encompassing the north of Namib-Kalahari dryland (12-18° E, 15-21° S), while the AEJ-S is observed near the border between Namibia-Kalahari dryland and Congo-Zaire basin (5-10° S). Both the AEJ-S and the heat low are strongest in September 2016, while the AEJ-S is weaker but more extensive, expanding to the ocean in October 2018. The heat low is weaker in August compared to other deployment 255 months, consistent with the thermal wind relationship, in which a stronger horizontal temperature gradient leads to stronger vertical wind shear, leading in turn to a stronger AEJ-S. Joint pdfs between the zonal wind at 600-700 hPa and RH (0-10° E, 5-10° S) indicate that RH and AEJ-S relationships in September and October are very similar, with a stronger jet advecting more moisture than a weaker jet. During August 2017, the AEJ-S is generally weak under dry conditions (RH < 20 %, Fig. 5d). In contrast, the AEJ-S is strong when RH was high 260 (> 60-70 %) in September and October (Figs. 5(e, f)). This can be also explained by moist convection over the continent migrating southward from September to October. The scatter plots show that the AEJ-S-heat low correlation is strongest in ). The correlation between AEJ-S and heat low in October and August is slightly lower than in September, suggesting that other factors may affect the strength of the jet besides the thermal wind relation at daily time scale.
The use of different times or daily mean data produces a similar result. The sensitivity of the correlation to the choice of the 265 heat low region between 10-20° S over the land is small. The correlation between AEJ-S and inland q is low: the AEJ-S may enhance convection through enhancing the vertical ascent at the jet entrance region (Jackson et al., 2009) but the additional advected moisture north of 10° S does not reduce the thermal contrast much (at most 10%, based on Adebiyi and Zuidema, 2016). Nonetheless, the question still remains whether the AEJ-S is also sustained by a meridional circulation driven by equatorial convection, similar to the AEJ-N ( Thorncroft and Blackburn, 1999). clear. We further look at the vertical motion along with the heat low and other meteorological variables during the deployment 280 months in Fig. 6.
The AEJ-S core is located further north (0-5° S; Fig. 6a, d) and lower in altitude (~3 km, 700 hPa; Fig. 6d) in August compared to September and October (5-10° S; ~ 4 km, 600 hPa (Fig. 2, Figs. 6(b-c, e-f)), and the jet cores are tied to the strength of the inland heat low (Figs. 6(a-c)), as shown in Fig. 5. Since the inland heat low is weaker in August, this explains why the jet core in August is located lower than that in September and October. This also accounts for why the jet strength at 285 around 4 km (~600 hPa) is weaker in August than in September and October (Fig. 6). Furthermore, the strong ascent inland, which is supported by a low-level convergence of surface winds, appears to be associated with a stronger AEJ-S over the SE Atlantic coast as well (Figs. 6(d-f)). This may also contribute to the aerosol transport efficiency, supported by the finding of Adebiyi and Zuidema (2016), that the AEJ-S also helps loft aerosol through ascent at the AEJ-S entrance region (10-15° S,

12-18° E). 290
Large-scale subsidence over the SE Atlantic is also associated with the strength of AEJ-S, especially over the subtropical region (15-25° S) (Figs. 6(a-c)). The weaker AEJ-S is associated with stronger subsidence, while the stronger mean AEJ-S is associated with suppressed subsidence, and this is most clearly shown in August (Fig. 6d). This result is consistent with the finding by Adebiyi and Zuidema (2016) that AEJ-S induces an ascending motion below the AEJ-S over the continent, including aerosol source regions while reducing the mean subsidence over the ocean north of 20° S. The vertical motion associated with 295 the AEJ-S during August, September, October in 2016-2018, are provided in Fig. 3S in the supplementary material. The stratocumulus deck, which accounts for most of the low cloud over the SE Atlantic, is affected by the BLH, LTS, and 305 large-scale subsidence (Wood and Bretherton, 2006;Wood et al., 2015;Der Dussen et al., 2016;Fuchs et al., 2018;Adebiyi and Zuidema, 2018). For all months, LTS is positively associated with low-CF (Figs. 7(a, b)). The lower troposphere is more stable in August 2017 and October 2018 near the south of offshore Namibia compared to climatology and the low-CF is increased there too. Climatologically, the strength of the LLJ increases from August to October, off the coast of Namibia around 15-25° S, 0-10° E, but LLJ is stronger in August 2017 and September 2016, while it is weaker in October 2018 310 compared to the climatological mean. The subsidence is stronger off the Namibian coast for all months than the climatological mean, especially in August. In particular, the LLJ is strong when subsidence off the Namibian coast is strong (Fig. 7) in August 2017 and September 2016. climatological mean for both ocean and land. The low-CF tends to get reduced when the BLH gets higher as shown in August 2017, but the exact relationship between low-CF and BLH at monthly mean time scale needs further investigation.  (-0.02, -0.07, -0.12 (from thick to thin)), and dark red (+0.02, +0.07, +0.12 (from thick to thin)), %), which is smoothed by averaging at 2° (longitude) by 5° (latitude) to reduce the noise. The purple square (yellow triangle) indicates shaded field (low-CF anomaly) data is significant at the 85% confidence level.

Meteorological characteristics associated with low-level cloud
To examine the impact of large-scale circulation on a low cloud during the ORACLES deployment months and how they 325 differ compared to the climatological mean more clearly, we computed the anomaly of each variable discussed in Fig. 7. In general, the high low-CF is strongly tied to high LTS (Fig. 8a). The LLJ strength is also negatively related to low CF, which is clearly shown in September 2016. (Fig. 8b). The strong subsidence tends to be positively associated with low-CF at monthly mean time scale, especially off the coast of southwestern African in August 2017 (5-12° E, 0-25° S, Fig 8c), but this association also appears to vary with regions and months (not shown). Furthermore, LLJ is positively associated with the 330 subsidence off the Namibian coast, clearly shown in August 2017. The relationship among AEJ-S, LLJ, subsidence, and low-CF, however, shows high variability due to factors such as 1) subsidence indirectly enhancing clouds by enhancing inversion strength or reducing clouds at the given inversion strength at monthly time scale (Myers and Norris, 2013), and 2) the different response of low clouds to subsidence as a function of timescale Adebiyi and Zuidema, 2018).
The relationship between low-CF and BLH is not clear here, partially because the BLH used in this study is not designed 335 to include the decoupled cloud-top layer. Although low-CF tends to be negatively associated with the BLH, especially over the ocean in August 2017 and September 2016, there is significant spatial variability (Fig. 8d). This demonstrates the need for use of BLH including the decoupled cloud-top layer to better understand its coupling with the low-CF. We will further investigate the relationship between cloud-topped BLH and low-CF in Part 2 of the meteorological overview paper. Similar

350
(latitude) to reduce the noise. The gray cross (yellow triangle) indicates SST anomaly (low-CF anomaly) data is significant at the 85% confidence level. The SST monthly mean data is obtained by averaging SST hourly data.
How the low-CF varies in association with the SST during the deployment years is examined in Fig. 9. The SST over the SE Atlantic is warmer during the deployment months compared to the climatological mean (Fig. 9c). Warmer SST tends to reduce the stratocumulus cloud fraction, by reducing the static stability (Wood et al., 2015). August 2017 has the warmest SST 355 anomaly in the 10-20° S region, and the smallest CF fraction is found over that region. However, in September 2016, the reduction of low-CF is found over the region of a cool SST anomaly in the 10-20° S region (Fig. 9b). There is also no robust relationship between SST and low-CF in October 2018, especially over the coastal region, although we notice the warmer SST is linked to the decreased low-CF anomaly especially over the tropical Atlantic (5° N-10° S, 20° W-0). Warming is more intense over the tropical Atlantic, although the warming trend is still reflected over the ORACLES flight region (5-20° S, a 360 boxed region in Fig. 9b).
The south Atlantic anticyclone associated with high SLP is the weakest in October 2018 compared to the other two deployment months (Fig. 9a). The south Atlantic anticyclone is stronger and closer to the coast in August 2017 and September 2016 than the climatological mean, but slightly weaker over the SE Atlantic and more southward in October 2018 compared to the climatological mean (shown in Fig. 8S in the supplementary materials). The monthly mean SST, SE Atlantic anticyclone 365 associated with SLP patterns for all August, September, and October during 2016-2018 are also provided in Fig. 8S in the supplementary material.

380
Considering the effect of the meteorological conditions on low-CF, we also examine how meteorology influences aerosol transport at regional and monthly time scales. Figure 10 shows the black carbon mixing ratios (BC) from CAMS and meteorological parameters from ERA5. The monthly mean BC is lower in August 2017 than the climatological mean, especially off the coast and over the ocean (Fig. 10a). This is consistent with a weaker AEJ-S during advection of these burning 385 constituents out over the SE Atlantic (Adebiyi and Zuidema, 2016;Pistone et al., 2021), which may help explain the anomalously low aerosol optical depth in August 2017 . The BC in both September 2016 and October 2018 is slightly higher than the climatological mean. The high values of BC are also found in the relatively high RH condition, which is mostly ranging from 20% at lower altitudes up to around 70% of RH. Similar features were found in other trace gas such as CO as well (not shown). 390 Another notable feature in the CAMS black carbon (and CO) analyses is that the peak mixing ratio occurs around 2-3 km in all deployment months (Fig. 10b). While the vertical motion clearly differs among three months, the vertical structures of BC do not covary with ascent strength. Shinozuka et al. (2020) and Doherty et al. (2021) both document that the aerosols in models tend to have more of their mass located at a lower altitude than is evident in the ORACLES observations, although their comparisons did not include CAMS reanalysis data. While more quantitative analysis needs to be done, this suggests the 395 reanalysis assimilation schemes still have room for improvement in how they distribute aerosols vertically. The difference between BC in August 2017 and its climatological mean is smaller over the land than the coast (e.g. when we average BC and zonal wind speed over 5-10° S, 12-18° E region), but the overall features are similar.

The influence of the upper-level disturbance on the AEJ-S 400
The primary focus of this paper is to provide a meteorological overview during the ORACLES deployment year. However, we seek, in addition, to understand why August 2017 was characterized by a lower and weaker AEJ-S compared to climatology. Previous studies have shown that upper-level mid-latitude disturbance can modulate the temperature gradient over SE Atlantic (Adebiyi and Zuidema, 2018;Kuete et al., 2020), and the different phases of upper-level waves play an important role in modulating the strength of the AEJ-S over the SE Atlantic as well (Kuete et al., 2020). Motivated by these 405 studies, we examine whether the upper-level disturbances can contribute to the weakening of AEJ-S in August 2017 and how they modulated the AEJ-S. To examine how the circulation associated with the upper-level disturbance can influence the AEJ-S further, we 420 computed the thickness (heat low over the continent) anomaly with the upper-level PV anomaly at 250 hPa. Indeed, the upperlevel disturbance is closely linked to the change in thickness, which is proportional to the layer mean temperature, shown in Fig. 11a. For example, in August 2017, the negative PV anomaly is linked to positive thickness anomaly (warm air), and the anticyclones associated with this advect the cool air from the mid-latitude ocean, reducing the heat low over the land (Fig.   11a). 425 Figure 11 shows the schematics of the anomalous 250 hPa circulation associate with the developing mid-latitude upperlevel wave (or disturbance). An anomalous ascent is found downstream of the troughs, and anomalous subsidence is found upstream of the trough. This subsidence and ascent drive an anomalous anticyclone and cyclone in the 900-600 hPa depth region up to the upper-tropospheric region (~ 250 hPa), leading to northward motion over Namibia. This advects air from higher mid-latitude latitudes (i.e. cooler air) to the Angolan highlands, reducing the meridional temperature contrast between 430 the Angolan highlands and the Congo Basin, leading to weaker AEJ-S over the land at the north of 10° S. Together with this, the subsidence over the ocean is much stronger in August 2017 especially off the Benguela coast up to around 3° S, preventing the AEJ-S from transporting to the ocean (Adebiyi and Zuidema, 2016), which is also closely tied to the phase of the upperlevel disturbances (Kuete et al., 2020). These upper-level waves and associated circulations also explain the variability and strength of the heat low and AEJ-S during the other deployment months as well. (Fig. 9S in the supplementary material). 435 As a remote driver, the Madden Julian Oscillation convection (MJO; Madden and Julian, 1994;Wheeler and Hendon, 2004), an intraseasonal convective variability in the equatorial troposphere with a periodicity of about 30-90 days, may contribute to the weakening of AEJ-S in August 2017, because this is weakened over Africa during this period (shown in Fig.   10S in the supplementary material). The MJO can affect the timing and intensity of convectively coupled Kelvin waves and convective activity over Africa (Guo et al., 2014), which can affect AEJ-S activity (Ventrice and Thorncroft, 2013;Zaitchik, 440 2017). However, this remote driver has been investigated for the AEJ-N, and the MJO's influence on the AEJ-S is less understood and remains unclear.

Summary and discussion
This paper describes the meteorological factors controlling aerosol transport and low cloud fraction during the August, 445 September, and October 2016-2018 deployments of ObsErvation of Aerosols above CLouds and their intEractionS (ORACLES) project, particularly focusing on the climatological aspect of the meteorological overview at monthly time scale.
The heat low was the lowest among deployment months as well. The region around 5-10° S, 0-15° E was drier than the climatology with the moisture mostly confined to near the equator (< 5° N) than the climatological mean. The LLJ was stronger in August 2017 (~ 2 m s -1 ) compared to the climatological mean. The well-mixed BLH in August 2017 was lower near the coast and higher over the SE Atlantic Ocean than the climatological mean. The subsidence near the Namibian 455 coast was considerably stronger in August 2017 than the climatological mean. The low-CF was slightly smaller than the August climatology. The negative PV anomalies and anomalous circulation associated with the mid-latitude upper-level disturbance over the SE Atlantic Ocean brought the cool mid-latitude airmass, resulting in the reduced heat low and the weaker AEJ-S. o September 2016: The AEJ-S in September 2016 was similar to the climatological mean. Interestingly, the 460 heat low (up to 5-10 m) over the north of Namibian dryland, and the ascending motion over the Namibian dryland was slightly stronger (up to 2 hPa day -1 ) in September 2016 than the climatological mean. The LLJ was stronger in September 2016 (~1 m s -1 ) compared to the climatological mean. In general, the BLH over the South Atlantic Ocean in September 2016 was higher (~ 100 m) than the climatological mean. The subsidence near the Namibian coast and South Atlantic in September 2016 was also slightly stronger than the climatological mean. The low-CF was slightly smaller than the 465 September climatology. o October 2018: The updraft over the continent was slightly weak, but the horizontal temperature gradient still maintained the AEJ-S and vertical motion associated with the jet. The AEJ-S was slightly weaker than the climatological mean, but the difference was small (less than ±1 m s -1 ) with the heat low weaker than the climatological mean. The LLJ was weaker in October 2018 (-2~ -3 m s -1 ) compared to the climatological mean. The BLH over the South Atlantic Ocean 470 in October 2018 was also lower (100-200 m) than the climatological mean. Furthermore, the BLH was lowest in October 2018 compared to other deployment months. The subsidence near the Namibian coast was also reduced in October 2018 by about 10~30 hPa day -1 off the Namibian coast (~13° E, ~22° S) compared to the climatological mean. The low-CF in October 2018 was the largest among deployment months, and larger than the October climatology. b) Key climatological characteristics that affect aerosol transport and low-level cloud: 475 • A weaker AEJ-S, weaker vertical motion over the land, a stronger LLJ, and stronger subsidence near the Benguela coast in August 2017 than the climatological mean was closely linked to a lower black carbon mixing ratio (BC) in August 2017 compared to climatology, in the CAMS reanalysis. Peak BC values were found in the high RH conditions, which is mostly resided between about 20%-70% of RH at the AEJ-S exit and entrance region at 600 -700 hPa. This indicates the possible association with the moist plume advected by AEJ-S from the continents. The 480 peak of maximum BC occurred between 2 -3km for all three months, without covarying with either the vertical motion or the AEJ-S. • The high low-level tropospheric stability (LTS) was closely associated with high low-CF. The LLJ strength is also negatively related to low-CF, most clearly shown over the subtropical Atlantic Ocean in September 2016. LLJ is positively associated with the subsidence off the Namibian coast, clearly shown in August 2017. Large-scale 485 subsidence off the coast of Namibia and the SE Atlantic Ocean and SST was also associated with the low-CF at monthly time scale, but the sign of relationship varied with time and space. The relationship between low-CF and BLH is not clear at monthly time scale, although low-CF tends to be negatively associated with the BLH, especially over the ocean in August 2017 and September 2016, 490 This paper provides meteorological context for interpreting the airborne aerosol measurements, particularly focusing on the climatological perspective. The goal of this Part 1 paper is primarily to describe how the large-scale meteorological factors are reflected in the aerosol transport and low cloud fractions during the deployment months, towards facilitating the interpretation of the field observation data. We will provide the detailed meteorological characteristics for the three deployment months at daily to weekly time scale in Part 2 of the meteorological overview paper. 495

Data availability
The analysis is based on open-source data.

Author contributions
RW and PZ envisioned the original ORACLES meteorological overview concept, and RU, JMR, and LP designed the manuscript structure. IC provided the climatological mean and monthly mean MODIS low-cloud data and assisted JMR to obtain the VIIRS daily mean cloud data product. LP and JMR developed the methodology of determining the BLH. JMR processed the data analysis and visualized the results. JMR, LP, and RU interpreted results and JMR wrote the manuscript. 515 LP, RU, PZ, RW, IC, and JR edited the manuscript. LP and RU led the meteorological forecast briefing during the 3 ORACLES field deployments.