1. Introduction
The timing of early summer rainfall onset over southern Africa exerts a profoundly important control on socioeconomic viability from the community to the regional scale (Reason et al. 2005; Kniveton et al. 2009; Thurlow et al. 2012; Conway et al. 2015). Agricultural productivity (Kanemasu et al. 1990), food security (Tadross et al. 2005) and power generation (Conway et al. 2015) are all negatively impacted by variable or late rainfall onset, with the latter highly correlated with shorter growing season length (Reason et al. 2005). Accordingly, understanding the controls on rainfall onset and improving its predictability has long been a key goal for forecasters (Phakula et al. 2018; Ratnam et al. 2018).
The August–October pre-onset circulation over tropical southern Africa is characterized by a unique and distinguishing large-scale feature, the Congo air boundary (CAB). The CAB occurs at the interface of tropical deep convection in the southern Congo basin and dry stable subtropical air to the south (Torrance 1979). With the CAB in place, tropical convection is locked into the Congo basin to the north while the subtropics to the south are dominated by deep subsidence, low surface dewpoint temperatures and a dry regime (Howard and Washington 2019). The breakdown of the CAB occurs contemporaneously with the southward advance of convection over the subcontinent as subsidence inhibition and moisture limitation is lost, facilitating broad rainfall over southern Africa.
The presence or absence of the CAB is a critical control of transition-season rainfall over the region. Figure 1 presents the odds ratios of a rainy day in October or November under CAB presence or absence. Rainy days are 8 times more likely over the core of the southern African subtropics (centered on southern Zambia, northern Zimbabwe, northern Botswana, and southeast Namibia) in the absence of the CAB, with this increase locally as large as 40 times in the Zambezi river valley. Understanding the dynamics of the CAB, and particularly the causes of CAB breakdown, is therefore integral to understanding onset at the southern African tropical edge.
Odds ratio of a rainy day (>0.1 mm day−1) in October or November given the presence or absence of the CAB. The color map is scaled exponentially. Rainfall data are from the CHIRPS V2 precipitation dataset, while daily CAB detection follows the canny-edge method of Howard and Washington (2019) (full details in methods). The odds ratio is calculated from a 1981–2018 climatology, with 816 CAB present days and 1502 CAB absent days. Stippling indicates the regions where October–November CAB frequency at the day scale is positively associated with ERA5 rainfall onset date [following the method of Liebmann et al. (2012)] at the 95% significance level in a 1979–2018 climatology.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
It is well established in the literature that midlatitude disturbances can be associated with precipitation extremes in subtropical and tropical regions (Knippertz 2007; de Vries 2021). As illustrative examples over Africa, Ward et al. (2021) show a remote influence of upper-level midlatitude troughs to Saharan warming and northern Congo rainfall, while Hart et al. (2010) link three summer rainfall extremes over south Africa to cloud bands and Rossby waves in the Southern Hemisphere westerly jet. Given these well-documented tropical–extratropical linkages, it is plausible that midlatitude disturbances play a role in the breakdown of the CAB and onset of the rains over tropical southern Africa. While literature addresses subseasonal modulation of rainfall over southern Africa by the midlatitudes, the contribution of Rossby wave dynamics to onset processes is unexplored. To address this research gap, this paper will first identify and characterize CAB breakdown events. Subsequently, it will investigate the causes of these breakdown events and the potential role of midlatitudes disturbances in the breakdown process.
2. Background
a. Onset predictability
Predictive skill in the forecasting of southern African summer rains onset has remained elusive. While Ratnam et al. (2018) were able to produce onset forecasts with realistic spatial distributions over the subcontinent using a seasonal prediction model ensemble, root-mean-square error values were greater than 30 days over most of South Africa and southern Botswana, Zimbabwe and Mozambique. Similarly, Phakula et al. (2018) showed that while global coupled models had skill predicting 3-month seasonal rainfall totals over a similar domain, they struggled with forecasting rainfall totals for the onset months of October and November. Clearly, onset is a key period for which seasonal forecasting is lacking in skill. Underpinning this dearth in predictive skill is a poor understanding of the controlling dynamics of rainfall onset. This means that it is difficult to point to the particular deficiencies in numerical model simulations that require improvement to enhance forecasting skill.
b. The Congo air boundary
Forming during the pre-onset period, the CAB is a distinct feature of the tropical edge. The CAB is defined as a convergence zone and dryline where near saturated low-level westerlies originating from the Atlantic and Congo meet easterly Indian Ocean trade winds, which are dried by strong heating of the continental surface and diluted by deep tropospheric subsidence south of the CAB (Howard and Washington 2019). It represents the southern limit of the area of tropical deep moist convection, with a stable subtropical troposphere dominating to the south.
Figure 2 demonstrates the climatological importance of CAB dynamics to onset over southern Africa. From August to October, the CAB edges southward from the Congo basin, maintaining integrity but displacing dry conditions with moist convection. In October and November, spells of weather occur where the CAB breaks down and the strong meridional gradient in humidity is lost, allowing deep moist convection and significant rainfall over the central subtropics (Howard and Washington 2019). In the last week of September (shown as “A” in Fig. 2), where the climatological daily CAB frequency is 80%, a strong dryline is clear in near-surface specific humidity and the high values of convective available potential energy (CAPE) are contained in the Congo basin north of 10°S. Conversely, by the first week of November (“B” in Fig. 2), where CAB frequency is 32%, the CAB dryline is absent from the climatological specific humidity field and CAPE values > 250 J kg−1 are prevalent over large parts of southern Africa. The CAB typically does not reform after the latter half of November, marking the full transition to a post-onset rainfall regime where widespread convection and rainfall occurs through a variety of mechanisms such as cloud bands and tropical lows (Hart et al. 2013; Howard et al. 2019).
(a) A 40-yr (1979–2018) climatology of daily CAB frequency from 1 Sep to 1 Dec. CAB detection follows the canny-edge method of Howard and Washington (2019). (b),(c) Composite plots of ERA5 climatological (1979–2018) 850-hPa specific humidity (g kg−1) and 1200 UTC convective available potential energy (CAPE) (J kg−1) for the last 7 calendar days of September (“A”) when CAB frequency is 80% and (d),(e) the first 7 calendar days of November (“B”) when CAB frequency is 32%.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
The CAB exercises an important control on rainfall over the core of the southern African subtropics (Fig. 1). High odds ratio values centered on southern Zambia, northern Zimbabwe, northern Botswana, and southeast Namibia indicate that the CAB strongly controls precipitation over the region, with rainy days 5–40 times more likely when the CAB is absent. In the figure, the climatological location of the CAB is clear as the ∼1000 km southwest–northeast diagonal line around 10°–12°S that demarcates the northern limit of the region of high odds ratios (Figs. 1 and 2b). Odds ratios decline south of 20°S as midlatitude rainfall mechanisms such as tropical temperature troughs (TTTs), cutoff lows (COLs), and ridging highs become increasingly important. These contribute to rainfall over the far south of the subcontinent independent of CAB presence or absence. Midlatitude precipitation mechanisms may also contribute to rainfall over the central domain through tropical–extratropical interactions, but importantly this is only (i) in the absence of the CAB or (ii) through causing its breakdown, hence the importance of the feature as a regional precipitation control.
Figure 3 presents a schematic of pre- and post-CAB breakdown regimes over southern Africa. Under the pre-breakdown regime, moist air is locked into the Congo basin to the north of the CAB, with dry air capped by subsidence and a strong Botswana high circulation to the south (Moses et al. 2023). The influence of midlatitude precipitation mechanisms is confined to the south of the subcontinent. Under the post-breakdown regime, the CAB is absent. Moist air covers most of southern Africa with a new dryline forming over the Kalahari (Howard and Washington 2019). The midlevel Botswana high is replaced by tropical convection through the depth of the troposphere and midlatitude influences may modulate precipitation over the entirety of the subcontinent under tropical–extratropical interactions. The transition from a pre- to post-breakdown regime occurs over time scales of 2–5 days and is what we term in this work “CAB breakdown.” Given the importance of the CAB to early-summer rainfall over tropical southern Africa, considering the causes of this breakdown is a crucial step toward understanding the controlling dynamics of rainfall onset.
A schematic of key southern African regional circulation features with the CAB (a) present and (b) absent. Features are as given in the legend and are not to scale. Locations are approximate. Gold stars mark the September–November hotspots of anticyclonic Rossby wave breaking identified by Ndarana and Waugh (2011). CAB breakdown represents the transition of the regional circulation from (a) a CAB-present state to (b) a CAB-absent state over time scales of 2–5 days. Tropical temperature troughs here are indicative of a variety of midlatitude precipitation mechanisms.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
c. Midlatitude influence on the (sub)tropics
The midlatitudes influence the subtropics and tropics through baroclinic Rossby waves. These are long horizontal wavelength (>2000 km) oscillations in the midlatitude westerlies that predominantly occur in the upper troposphere and lower stratosphere. The propagation of these Rossby waves creates perturbations in the climatologically meridional stratification of potential vorticity (PV); upper-level ridges create poleward intrusions of high-PV air, while upper-level troughs create equatorward intrusions of low-PV air (in the Southern Hemisphere PV is generally negative and decreases toward the poles). Troughs may be involved in Rossby wave breaking (RWB) where Rossby waves amplify and overturn such that the typical meridional gradient of PV becomes reversed (McIntyre and Palmer 1983). Southern Africa is a particular Rossby wave breaking hotspot; wave breaking typically initiates in the South Atlantic, consolidates over the continent and dissipates in the Indian Ocean (Postel and Hitchman 1999). We include the upstream and downstream wave breaking hotspots for September–November identified in Ndarana and Waugh (2011) Southern Hemisphere wave breaking climatology in our Fig. 3 schematic.
Numerous studies document the link between Rossby waves, Rossby wave breaking and subtropical extreme precipitation events (Knippertz 2007). For example, Hart et al. (2010) link PV troughs with tropical–extratropical cloud bands (regionally known as TTTs) while Ndarana and Waugh (2010) show that Rossby wave breaking is associated with 89% of Southern Hemisphere cutoff lows (COLs), both of which are linked to extreme spring and summer precipitation events over southern Africa (Hart et al. 2010; Favre et al. 2013). Ridging highs, important in advecting moisture from the Indian Ocean to the continental interior, also have a well-established relationship with RWB (Ivanciu et al. 2022). More broadly, de Vries (2021) link >80% of extreme subtropical precipitation events to Rossby wave breaking in a global review. These Rossby wave–related precipitation extremes are typically concurrent with intense poleward moisture transport from (moist) tropical to (dry) subtropical regions (de Vries 2021). Qualitatively, this anomalous subtropical precipitation and poleward moisture transport is analogous to CAB breakdown. However, studies that address the extent to which baroclinic Rossby waves modify atmospheric stability and moisture fluxes over southern Africa have largely focused on the austral summer (Hart et al. 2010; Macron et al. 2014) and the southern regions of the continent which are known to receive rainfall of midlatitude origin (Reason and Rouault 2005; Favre et al. 2013; Ndarana et al. 2018, 2021, 2022; Ivanciu et al. 2022). Less is known about the potential role of such midlatitude disturbances in onset dynamics over central southern Africa and the tropical edge.
While midlatitude disturbances are the key to transition-season rainfall over the far south of the subcontinent, any role in rainfall onset at the tropical edge is mediated by the CAB. Midlatitude precipitation mechanisms cannot contribute to rainfall at the tropical edge when the CAB is present (Fig. 1). Only when the CAB is absent, or through causing CAB breakdown, may they contribute to precipitation at the tropical edge and thus rainfall onset. In other words, CAB breakdown “unlocks” the possibility of tropical–extratropical interactions over southern Africa and allows midlatitude influences to modulate precipitation from the extratropics to the Congo basin as is well established in the literature. Given this, it is plausible that midlatitude disturbances play a role in the timing of onset at the southern African tropical edge if midlatitude baroclinic disturbances are a cause of CAB breakdown. To investigate this, in the subsequent section we identify and characterize CAB breakdown events before investigating breakdown causes.
3. Data and methods
a. Data
This work uses hourly or daily mean surface and pressure-level reanalysis fields from the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 product (Hersbach et al. 2020). The ERA5 dataset is a detailed record of past climate from 1950 to present, with a 31-km grid spacing, 137 vertical levels, and hourly output for a range of atmosphere, land, and ocean variables (Hersbach et al. 2020). Only data after 1979 are used as routine satellite observations over the Southern Hemisphere began after this date (Tennant 2004). ERA5 is favored over other reanalyses due to its spatial resolution, record length, and established verification. To ensure that results presented here are not sensitive to reanalysis product, we also use the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2), which is known to perform well in the Congo basin (Hua et al. 2019). MERRA2 fields are omitted for brevity as MERRA2 and ERA5 were remarkably consistent in the evolution of dynamic and thermodynamic fields across breakdowns, likely due to the large spatial scales across which the processes of interest occur.
Outgoing longwave radiation (OLR) data from the OLR version 1.2 dataset provided by the National Oceanic and Atmospheric Administration (NOAA) Climate Data Record (CDR) at a 1° × 1° resolution are used to identify anomalous convection over southern Africa (Lee 2014).
Daily precipitation data are used to characterize rainfall location and intensity. Similar results were produced when using ERA5 precipitation and satellite-derived precipitation data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS v2) dataset. ERA5 precipitation is presented owing to data availability for full period of OLR data (1979–2018) and consistency with the dynamical fields.
b. Congo air boundary breakdown detection
We identify the CAB on a daily time scale in ERA5 following Howard and Washington (2019) whereby a canny edge detection algorithm is applied to smoothed fields of surface specific humidity to identify strong gradients in the continental sector 5°–18°S. Subsequent shape/orientation/length filtering removes nonsignificant segments and those not associated with the CAB. The canny edge detection algorithm of Howard and Washington (2019) is available publicly on GitHub.1
To identify CAB breakdown events, a daily time series of total detected CAB grid cells was produced for August–December 1979–2018. This was smoothed with a 5-day running mean. CAB breakdown events were then defined as instances where the 5-day mean CAB gridcell count fell below 40 for 4 consecutive days (>40 is day −1, <40 is days 0 to 3). A total of 108 breakdown events were identified for the 1979–2018 period. Breakdown frequency increases with the gridcell threshold chosen, with a threshold of 35 (45) leading to the detection of 96 (115) events.
To confirm CAB presence at D − 2 and CAB absence at D + 2, visual inspection of day-mean surface specific humidity was carried out. 23 cases were rejected on the grounds that the CAB was either absent at D − 2 (n = 12) or present at D + 2 (n = 11). We take this stringent approach to CAB breakdown identification as the objective of this work is to investigate only the large-scale breakdown events associated with significant increases in specific humidity over subtropical southern Africa, which are by extension the most important to early season rainfall. The total number of CAB breakdowns identified from 1979 to 2018 was 85, an average of 2.13 breakdown events per year. Breakdown was most frequent in October (n = 40), followed by November (n = 24), September (n = 17), August (n = 3) and December (n = 1).
c. Compositing
We present in the following section composite means of the 85 breakdowns during 1979–2018. All 85 breakdown events were bought into phase by assigning D = 0 to the day on which breakdown was identified. Because of differences in the location, amplitude, and phase of various features of interest there will necessarily be some smearing and smoothing in both time and space within the composites. The logic of calculating a composite mean is that the signal to noise ratio will be enhanced.
4. Results
a. Congo air boundary breakdown
First, we characterize CAB breakdown events through composite analysis of the 85 breakdown events identified in section 3b. Figure 4 shows that a key feature of CAB breakdown is the loss of a strong meridional specific humidity gradient and a southward advance of the area of high specific humidity. Initially (D − 3) the CAB is present as a strong near-surface gradient at 12°S. From D = 0, moist (>10 g kg−1) air advances southward and the CAB dryline is lost. The 7 g kg−1 specific humidity contour advances around 10° latitude (1100 km) southward over 3 days.
Lead (D − 3, D − 2, D − 1) and lagged (D = 0, D + 1, D + 2) daily composite vertical profiles of specific humidity (g kg−1) by latitude for 1979–2018 CAB breakdown events (n = 85) averaged across 20°–30°E. The day of CAB breakdown is designated D = 0 for each of the 85 included breakdown events.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
Figure 5 plots the 4-day evolution of near-surface specific humidity anomaly, surface temperature anomaly and upper-level streamfunction anomaly around CAB breakdowns. We use the 250-hPa pressure level for “upper levels” through this work as this pressure surface is in close proximity to both the subtropical and midlatitude dynamical tropopause positions (Ndarana and Waugh 2010). The specific humidity field shows a significant increase in moisture over much of southern Africa during breakdown, and particularly over southeast Angola, Zambia, Zimbabwe, and Botswana. The region of increased moisture is strongly collocated with the region of high rainfall sensitivity to the CAB identified in Fig. 1, highlighting the importance of moisture limitation to the CAB–rainfall relationship. Concurrently, there is a transition through the sequence from broadly positive to broadly negative temperature anomalies over central southern Africa. Streamfunction at 250 hPa shows positive (cyclonic) anomalies over the subcontinent, which shift northward and increase in magnitude through the sequence. The region of maximum specific humidity change and the leading edge of streamfunction anomalies are strongly collocated.
Lead (D − 1) and lagged (D = 0, D + 1, D + 2) composite day-mean (a)–(d) 850-hPa specific humidity (g kg−1), (e)–(h) 2-m temperature anomaly (°C), and (i)–(l) 250-hPa streamfunction anomaly (against an October climatology) (m2 s−1) for all 1979–2018 CAB breakdown events (n = 85) from D − 1 to D + 2. In (i)–(l) a −9 s−1 absolute vorticity contour at 250 hPa is plotted over streamfunction anomalies. Positive streamfunction anomalies are cyclonic. Stippling indicates where anomalies are significantly different from the October climatology or zero at the 0.05 level.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
Figure 5 suggests a coherent midlatitude signature during CAB breakdowns. Positive (cyclonic) streamfunction anomalies diagnose the presence of a stationary upper-level trough during the sequence. Troughs are cool events over southern Africa due to the influence of cold air advection and cloud shading (Kuete et al. 2020), visible as the transient cool anomaly at the tip of the subcontinent from D − 1 to D + 1. This signal is significant in a composite of all 1979–2018 breakdown events, suggesting that troughs may play an important role in forcing the cooling and moistening of southern Africa during CAB breakdown.
It is notable that cyclonic streamfunction anomalies persist and even intensify over southern Africa through the 4-day breakdown sequence. Typically, troughs propagate eastward with the zonal flow at a phase speed set by their wavenumber (Holton 2012). Stationarity is possible in two ways; low-wavenumber Rossby waves with zero phase speed, or Rossby wave breaking (RWB). Streamfunction anomalies associated with a low-wavenumber (high-wavelength) Rossby wave would exhibit greater zonal separation between positive and negative anomalies, meaning RWB is the likely cause of the anomaly patterns of Fig. 5. There is also evidence of a transient midlatitude disturbance in the west–east migration of the most southerly streamfunction and temperature anomalies through the sequence. The reviews of Knippertz (2007) and de Vries (2021) highlight the large body of work that links midlatitude troughs and RWB to enhanced tropical convection, providing a mechanism by which such disturbances may be causally related to CAB breakdown. The strength of signals emerging from Fig. 5 are weak, however, as troughs may not occur with every breakdown.
In light of these results, we pose the following research questions:
-
RQ1: What proportion of CAB breakdowns are related to the troughs in the midlatitudes?
-
RQ2: What are the characteristics of Rossby waves that are related to breakdown?
-
RQ3: How do Rossby waves contribute to breakdown?
b. What proportion of breakdowns are associated with the midlatitudes?
To determine the proportion of CAB breakdown events associated with the midlatitudes (RQ1), we begin by testing for the proportion of breakdown events that are immediately preceded by (i) troughs or (ii) RWB. The methods to identify each are outlined below.
1) Strong trough identification
A strong trough is considered present on a given day if TPday_max is greater than 1 standard deviation above the August–December mean (TPday_max > 1134). This threshold resulted in 15% of days from August to December 1979–2018 being classed as strong trough days. A threshold of 1200 (1000) flagged 14% (18%) of days as strong trough days. We consider one standard deviation above the mean an appropriate middle value that flags significant troughs but is not highly sensitive to assumptions of the index method about wavelength, location and orientation.
2) Rossby wave breaking identification
Daily relative vorticity (ζ) at 0.25° × 0.25° resolution was obtained from ERA5 archives, summed with planetary vorticity (Coriolis parameter f), and smoothed. Smoothing is achieved through conversion of globally gridded data to spectral data and then a truncation of spherical harmonics at T15. This ensured that the method only identifies large-scale overturning as a wave breaking event, rather than small-scale perturbations in the vorticity field. A RWB detection algorithm then searches for locations along select absolute vorticity contours (“C”) where a single meridian intersects C at least three times, and these intersection points are flagged as an overturning point. Where two or more overturning points are in longitudinal proximity (i.e., overturning spans two meridian lines) they are grouped together as a RWB event. Only RWB events for which the mean latitude and longitude of all overturning points falls within the sector of interest (20°–70°S, 0°–50°E) are retained, and small-scale events with <20° along the meridian between the most northern and most southern intersection points are removed. RWB in the sector of interest was identified on 34% of days from August to December 1979–2018, including both anticyclonic and cyclonic wave breaking. Barnes and Hartmann (2012) show that RWB identification in pressure-level absolute vorticity is a robust method that produces a similar distribution to RWB identification using potential vorticity (PV) on isothermal surfaces. We prefer a pressure-level absolute vorticity metric for its simplicity and for consistency through this paper; the properties of PV (conservation, invertibility) that have made it such a key tool for unlocking midlatitude dynamics are not relevant to this work (Hoskins 2015). Unless otherwise stated, for data visualization purposes absolute vorticity is presented with a spherical truncation at T21 in plots to enhance signal to noise.
3) Association with breakdown
To determine Rossby wave association to breakdown, we test if breakdown is accompanied by RWB or a strong trough event in the time interval from two days prior to the day of detected breakdown. An “RWB” event set is constructed as a subset of the 85 identified breakdown events where RWB was detected within 20°–70°S, 0°–50°E from 2 days prior to the day of detected breakdown, while a trough (“TR”) event set was constructed as a subset of the 85 identified breakdown events where a strong trough was detected in the same period. Breakdown events flagged by both the trough and RWB detection schemes were assigned only to the RWB category. To enhance the strength of signals in subsequent composite analyses and account for the differing dynamical evolution of each breakdown, the date of a RWB breakdown event (D = 0) is defined as the first day of detected overturning of absolute vorticity. Dates of TR breakdown events (D = 0) are defined as the day of maximum TP index values. Relative to CAB-gridcell-detected breakdown dates, RWB or TR dates were either (i) unchanged, (ii) 1 day, or (ii) 2 days previous.
Overall, 53% (45/85) of 1979–2018 CAB breakdown events are associated with Rossby wave breaking over southern Africa. A further 18% (15 of 85) are associated with troughs in the midlatitude westerly jet. In total, 60 (71%) of the 85 CAB breakdown events from 1979 to 2018 co-occur with troughs or RWB. The proportion of associated breakdowns decreases through the onset period from 100% in August (n = 3) to 46% in November (n = 11), as shown in Table 1, suggesting a decreasing association of midlatitude Rossby waves and breakdown through the transition season.
Monthly proportion of 1979–2018 CAB breakdown events related to midlatitude Rossby waves. Breakdowns are associated if occurring within 2 days of RWB or strong trough events.
Figure 6 presents a composite of 250-hPa absolute vorticity, near-surface specific humidity anomaly and CAPE for the 60 CAB breakdown events we relate to midlatitude Rossby waves. Upper-level absolute vorticity shows a westward-leaning trough at D − 1, with a strong CAB in place, negative specific humidity anomalies over southern Africa and CAPE locked into the Congo basin. At D = 0, anticyclonic RWB is clear from the overturning of the upper-level absolute vorticity contours. Through D = 0 and D + 1, the strong meridional gradient of specific humidity and CAPE is eroded. By D + 2, the RWB event has begun to dissipate but significant increases in specific humidity south of the CAB are clear, matched by higher 1200 UTC CAPE over a broad region of southern Africa. It is clear that RWB is contemporaneous with a large and spatially significant increase in specific humidity and CAPE over subtropical southern Africa.
Lead (D − 1) and lagged (D = 0, D + 1, D + 2) composite day mean (a)–(d) 250-hPa absolute vorticity (s−1), (e)–(h) 850-hPa specific humidity anomaly (g kg−1), and (i)–(l) 1200 UTC convective available potential energy (CAPE) (J kg−1) for all CAB breakdown events related to midlatitude Rossby waves (both RWB and TR event sets, n = 60). For RWB breakdowns D = 0 is defined as the first day of overturning, while for TR breakdowns D = 0 is defined as the day of maximum “TP” index values. Stippling indicates where values are significantly different from the October climatology or zero at the 0.05 level.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
c. What Rossby waves cause breakdown?
Having demonstrated that RWB is contemporaneous with breakdown and large increases in specific humidity and CAPE over subtropical southern Africa, we next investigate the characteristics of the Rossby waves that are concurrent with breakdowns (RQ2). To aid this analysis, a counterfactual event set (CF) is constructed featuring 32 September–November strong trough events (TPday_max >1 std dev above mean) from 1979 to 2018 that did not lead to CAB breakdown. The distribution of these events in the annual cycle was broadly similar to that of the TR and RWB event sets (Table 2).
Mean position and standard deviation of the Julian day of CAB breakdown events in the annual cycle for ALL, RWB, TR, and CF. The “ALL” event set is centered on the day of detected CAB breakdown, while the “RWB” event set is centered on the first day of detected overturing and the “TR” and “CF” sets are centered on the day of maximum “TP” index values.
Figure 7 shows the composite 5-day change in 850-hPa specific humidity and pentad precipitation change for each event set. CAB breakdown, shown by significant increases in specific humidity south of the CAB’s preferred location, is clear in the ALL, RWB and TR event sets, with no significant change in CF. Rainfall is strongly collocated with increases in specific humidity; the three breakdown cases contribute significantly to precipitation over the subcontinent while the CF case remains dry.
Composite 5-day change in 850-hPa specific humidity (g kg−1) and pentad precipitation change (difference in total rainfall between the previous and following 5 days) (mm) for all CAB breakdowns (ALL) and each of the RWB, TR, and CF event sets. The RWB event set is a subset of ALL where breakdown is associated with Rossby wave breaking, while TR is a subset of ALL where breakdown is associated with troughs. RWB and TR trigger large-scale responses over southern Africa. The CF event set is a counterfactual of trough events that do not trigger a large-scale response over the domain.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
1) Wavenumbers
Here we analyze wavenumber to determine the characteristics of the waves bound up with the RWB, TR and CF cases. Higher-wavenumber waves propagate more rapidly eastward with the zonal flow while low-wavenumber planetary waves are stationary or retrogress (Holton 2012). Table 3 details the amplitude and explanatory variance of the first 10 harmonics of 250-hPa geopotential height anomaly at 40°S for each of the three event sets. For the RWB case, the greatest explanatory variance is for wavenumbers 4 and 5, accounting for 18.4% and 20.4% of variance, respectively. For the TR case, the greatest explanatory variance is for wavenumber 3, accounting for 26.7% of variance and with an amplitude of 45.3 gpm. And for the CF case, the greatest explanatory variance is for wavenumber 6, accounting for 25.0% of variance and with an amplitude of 38.1 gpm. The Rossby waves that are important to breakdown therefore appear to be characterized by much lower wavenumbers than those unrelated to breakdowns. Extending the analysis over further latitudinal bands, the low-wavenumber signals of the RWB and TR cases are dominant consistently through to 55°S, while the wavenumber 6 peak of the CF case holds dominant explanatory power only from 35° to 40°S. We interpret that the Rossby waves related to breakdown are latitudinally coherent low-wavenumber disturbances across both the polar and subtropical jet, while those unrelated to breakdown are weaker higher wavenumber disturbances confined to only the subtropical jet.
Amplitude and variance of the first 10 harmonics of composite 250-hPa geopotential height anomaly between −180° and 180° at 40°S for each of the RWB, TR, and CF cases. Harmonics with the largest explanatory variance are highlighted in boldface.
2) Wave trains
Next, we consider the large-scale wave trains within which the vorticity anomalies of interest are embedded. Figure 8 plots Southern Hemisphere upper-level geopotential height anomalies for the “RWB,” “TR,” and “CF” event sets. In all three cases, a negative geopotential height anomaly indicative of an upper-level trough overlies southern Africa. In the RWB case this negative anomaly is broad and somewhat zonally orientated, covering a large range of latitudes and longitudes, while for the TR and CF cases it is confined to the southern tip of the continent.
Composite day mean 250-hPa geopotential height anomaly (gpm) for D = 0 of the (a) RWB, (b) TR, and (c) CF event sets. A solid black contour plots −9-s−1 absolute vorticity at 250 hPa. Dashed black closed contours plot OLR anomaly (W m−2) for the 5 days previous to the event shown at an interval of −4 W m−2. Stippling indicates where geopotential height anomalies are significantly different from zero at the 0.05 level.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
At the larger scale, the RWB case is characterized by a deeply extratropical trough–ridge structure in the South Atlantic, with associated geopotential height anomalies extending from 30° to 70°S (Table 3, Fig. 8a). Anticyclonic Rossby wave breaking occurs on the downstream equatorward flank of this wave train.
The TR composite is characterized by a wavenumber-3 trough centered over southern Africa with ridges fore and aft (Table 3, Fig. 8b). Compared to RWB, anomalies are weakly defined in the extratropical South Atlantic, but stronger over southern Africa. The extent to which RWB and TR differ is noteworthy; the two cases are products of very different wave trains and RWB does not appear to be a dynamical extension of TR troughs where overturning occurs.
While structurally similar to TR, the geopotential height anomalies of the CF case are weaker and less meridionally extensive, with stippling generally confined to lower latitudes (<50°S). We interpret this difference between TR and CF as further support that CF cases are weaker transient wavenumber-6 troughs confined to the subtropical jet, while TR cases are deeper low-wavenumber troughs involving superposition of the polar and subtropical jets (Table 3, Fig. 8).
3) Wave sources
Finally, we consider potential sources of the Rossby waves related to CAB breakdown. Sardeshmukh and Hoskins (1988) show that regions with divergent winds, such as those occurring in areas with strong meridional gradients of zonal winds, can advect vorticity and act as a source of Rossby wave energy. As well as stationary influences such as orography, vortex stretching as a result of convection can enhance divergence and act as a Rossby wave source.
The dashed contours of Fig. 8 plot mean OLR anomaly for the 5 days preceding each event set to identify upstream regions of anomalous convection that may be a Rossby wave source. In the TR composite, anomalous convection in the South Atlantic convergence zone (SACZ) is clear. SACZ convection can excite Rossby waves that propagate downstream to cause rainfall in southern Africa. Indeed, Grimm and Reason (2015) document a 5-day lagged summer teleconnection between rainfall in South America and Southern Africa as a result of this wave excitation. Dynamically, upper-level anticyclonic vorticity tendency induced by enhanced convection stalls the upper-level flow, contributing to downstream divergence and exciting a Rossby wave that grows and propagates toward southern Africa. Subsequent wave superposition with an extratropical disturbance produces a deep and energetic trough that causes CAB breakdown. The RWB and CF cases lack such a clear upstream convective signal across a 5-day lead. A more detailed inspection of the evolution of daily geopotential and OLR anomalies at a 0–6-day lead indicates that RWB and CF are led by deeply extratropical disturbances in the South Atlantic off the eastern coast of South America. These OLR anomalies are transient and decrease in magnitude with an increasing lead time, suggesting that a coherent upstream convective signal is not a feature of these two cases. This is likely a result of the averaging of a diverse number of upstream evolutions. Only the TR composite is characterized by clear upstream convection in the SACZ across a 5-day lead.
d. How do Rossby waves contribute to breakdown?
In this section we consider the evolution of each of the “RWB,” “TR,” and “CF” cases over southern Africa to investigate how Rossby waves dynamically contribute to CAB breakdown.
1) Vertical velocity
Figures 9 and 10 show lagged composite vertical velocity by latitude and at 500 hPa, respectively, for each of the three composite cases. In Fig. 9, the 7 g kg−1 specific humidity contour identifies the tropical region of high humidity. At D − 1 for all three event sets the CAB is clearly in place, with moist air locked into the Congo basin and a strong demarcation between deep tropospheric ascent in the Congo (0°–10°S) and the midlevel subsidence of the subtropics to the south of the CAB. Initial (D − 1) south-of-CAB subsidence is stronger in the CF case than others, possibly due to the earlier seasonal distribution of the CF events against an October climatology (Table 2). This south-of-CAB subsidence is key to the maintenance of a CAB-in-place, pre-onset state, diluting the easterly winds and forcing a highly stable atmospheric profile. D − 1 streamlines indicate this subsidence is related to both the midlatitudes and local overturning. By the end of the breakdown sequence (D + 2), there is a clear southward extension of the region of high humidity in the two CAB breakdown cases (RWB and TR), while the counterfactual (CF) remains unchanged. Midlevel subsidence is replaced by ascent in the two CAB breakdown cases but persists in CF.
Lead (D − 1) and lagged (D = 0, D + 1, D + 2) composite latitudinal profiles (20°–30°E) of vertical velocity (Pa s−1) for each of the (a)–(d) RWB, (e)–(h) TR, and (i)–(l) CF event sets. Topography is shaded black. A solid black contour plots 7 g kg−1 specific humidity. Stippling indicates where vertical velocity is significantly different from the October climatology at the 0.05 level. Streamlines show vertical velocity and the northward wind component, with vertical velocity scaled by 1000.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
Lead (D − 1) and lagged (D = 0, D + 1, D + 2) composite 500-hPa vertical velocity (Pa s−1) and OLR anomaly (W m−2) (against an October climatology) each of the (a)–(d) RWB, (e)–(h) TR, and (i)–(l) CF event sets. A solid black contour plots −9-s−1 absolute vorticity at 250 hPa. Stippling indicates where vertical velocity is significantly different from the October climatology at the 0.05 level.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
Figure 10 shows that the latitudinal changes in vertical velocity identified in Fig. 9 can be attributed to Rossby waves in the midlatitude westerly jet. A solid black contour plots −9-s−1 absolute vorticity, showing wave breaking in the RWB case and transient troughs migrating eastward in the TR and CF cases. In all cases, patterns of ascent and subsidence are tightly coupled to upper-level vorticity anomalies with ascent leading and subsidence trailing.
In the RWB case, wave breaking leads to the establishment of quasi-stationary ascent over the subcontinent and decent off the southwest coast. Through the sequence we see south-of-CAB subsidence eroded by this quasi-stationary ascent, so that by the end of the sequence ascent dominates through the depth of the troposphere and large negative OLR anomalies establish over the majority of subtropical southern Africa (Figs. 9 and 10). The dynamical impact of wave breaking here tightly reflects that demonstrated by Funatsu and Waugh (2008) in the tropical Pacific, whereby upper-level PV anomalies promote convection through destabilizing the lower troposphere and causing upward motion ahead of the vorticity anomaly. We view the location of wave breaking here as key; wave breaking must be upstream of southern Africa and vorticity anomalies located over the western seaboard for this ascent to consolidate over the continental interior and promote CAB breakdown. Wave breaking at this location is identified as a hotspot in Ndarana and Waugh’s (2011) Southern Hemisphere wave breaking climatology, but is less common than wave breaking downstream of southern Africa.
In the TR case, we see vertical velocity patterns associated with the transient trough migrate eastward. Strong ascent is associated with the trough’s divergent leading edge and subsidence associated with the convergent trailing edge. A cloud-band structure of negative OLR anomalies, tightly coupled to the region of maximum vertical ascent, displaces south-of-CAB subsidence and leads to regionwide negative OLR anomalies by D + 2.
The CF case is structurally similar to TR. However, the northward extension of perturbed vertical velocity is less and no permanent change to the dominant south-of-CAB subsidence occurs through the sequence. A weak cloud-band OLR signature is seen over the Indian Ocean, but it does not have any continental extension. A key distinction between TR and CF is that we see the TR case instigate a tropical–extratropical interaction over southern Africa that displaces south-of-CAB subsidence with regionwide ascent. In contrast, the CF case is insufficient to force this and south-of-CAB subsidence remains in place.
In all three cases, negative OLR anomalies lead the upper-level vorticity anomaly. Latent heating associated with this convection introduces an upper-level anticyclonic vorticity tendency which counteracts the cyclonic circulation around the vorticity anomaly, slowing wave propagation by stalling the leading edge of the trough. This contributes to wave amplitude growth and increases the probability of overturning (wave breaking). We observe this effect in all three cases as the disturbances transit southern Africa; wave propagation totally stalls in RWB while we see minor overturning of absolute vorticity in TR and an increasing wave amplitude and westward tilt in CF. This is not a surprising result; Ndarana and Waugh (2011) show a clear wave breaking hotspot downstream of southern Africa. In the RWB case, this feedback facilitates the persistence of synoptic ascent and further promotes breakdown. This is to a certain extent also true of the TR case, as although the trough remains transient, we see wave amplitude growth and eastward propagation is slowed. In the CF case, convection in the Mozambique channel slows wave propagation and contributes to RWB, a strong ridging high and enhanced south-of-CAB subsidence (Ivanciu et al. 2022). This adds dry upper-level air to the southern African domain and acts to reinforce a CAB-present state (Figs. 9k,l).
2) Wind field modification
Figure 11 plots moisture fluxes to investigate near-surface wind field modification during breakdowns. All three cases exhibit a D = 0 reduction in easterly moisture fluxes where they would be convergent with the low-level Congo westerlies at the CAB. This is an apparent common feature of midlatitude troughs over southern Africa; transient lows force an eastward displacement of the Mascarene anticyclone, weakening geopotential gradients around the Mozambique channel and slackening the easterlies across Mozambique and Zambia. At the same time, and likely related to the loss of convergence at the CAB, we see enhanced northwesterly moisture fluxes in the Congo. This leads to a southward advance of the region of moist air from the Congo basin in all cases.
Lead (D − 1) and lagged (D = 0, D + 1, D + 2) composite 850-hPa moisture fluxes (vectors) and anomaly of the moisture fluxes (shaded, against an October climatology) (g kg−1 m s−1) for each of the (a)–(d) RWB, (e)–(h) TR, and (i)–(l) CF event sets. Black contours plot 850-hPa geopotential height contours at 10-m intervals from 1520 to 1580 m. Moisture flux vectors are plotted where the u or υ component of the moisture flux anomaly is significantly different from the October climatology at the 0.05 level.
Citation: Journal of Climate 37, 8; 10.1175/JCLI-D-23-0446.1
In the latter stages of the sequence, ridging of the South Atlantic high is also clear in all three cases. The importance of ridging highs in advecting moisture from the Mozambique channel to the continental interior is well established in the literature, and we see evidence of enhanced moisture fluxes along this axis in all three cases (Ndarana et al. 2021, 2022; Ivanciu et al. 2022). In the RWB and TR cases, anomalous positive moisture fluxes are seen in the Limpopo and Zambezi low-level jet regions, particularly in the TR case at D + 2. This is largely driven geostrophically by a ridging high pressure system in RWB, whereas for TR the moisture flux appears dominantly ageostrophic. In the CF case, these moisture fluxes are generally weaker. Low-level jets act as “atmospheric rivers” for the inflow of Indian Ocean moisture into the subcontinent, which is then available for precipitation, providing fuel for the breakdown event (Munday et al. 2021; Spavins-Hicks et al. 2021; Barimalala et al. 2021). Where this is combined with a mechanism of synoptic ascent, such as RWB or cloud bands, breakdown may lead to significant rainfall over southern Africa.
It is interesting to note that in general, Fig. 11 is characterized by negative moisture flux anomalies over central southern Africa associated with weaker moisture fluxes embedded in the tropical easterlies. The resulting loss of CAB convergence appears key to allowing moisture-laden Congo air to advance southward and fuel convection over subtropical southern Africa. Moisture fluxes of the CF case look broadly similar to TR and RWB (but with reduced intensity and shorter-lived anomalies), underscoring the controlling importance of vorticity-forced changes in static stability to breakdown.
3) Summary
This section has identified two primary mechanisms that summarize the contribution of midlatitude troughs to CAB breakdown. These are as follows:
-
The replacement of south-of-CAB tropospheric subsidence with Rossby wave–associated synoptic ascent.
-
Modification of the climatological wind field that leads to (i) southward moisture transport from the Congo basin and (ii) anomalous moisture fluxes from the Indian Ocean into the continental interior by ridging highs.
To ensure results presented here are not sensitive to reanalysis product, we reproduced key figures using Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2), known to perform well in the Congo basin (Hua et al. 2019). An analysis of specific humidity, vertical velocity, upper level absolute vorticity and u and υ wind fields yields remarkably consistent dynamic and thermodynamic evolutions as those we have identified in ERA5, giving confidence the mechanisms identified here are not sensitive to reanalysis product.
Both mechanisms 1 and 2 are complimentary in their contributions to CAB breakdown and the switch from a dry-subtropical to moist-tropical regime over southern Africa. The starkest differences against the counterfactual are seen in the vertical velocity fields, where breakdown leads to a total removal of south-of-CAB subsidence. The loss of south-of-CAB subsidence facilitates both (i) convection through reduced atmospheric stability and (ii) enhanced moisture fluxes from the Mozambique channel. If subsidence remains, modified moisture fluxes are insufficient to force breakdown.
The importance of Rossby wave breaking to CAB breakdown is mediated by the location of wave breaking (i.e., the location of upper-level vorticity anomalies). A total of 53% of 1979–2018 CAB breakdowns were associated with wave breaking upstream of southern Africa on the equatorward flank of an extratropical trough–ridge in the South Atlantic. Upstream wave breaking promotes quasi-stationary ascent over the continental interior, eroding south-of-CAB subsidence and promoting breakdown. While we only searched for wave breaking as far as 50°E, we found no evidence of association with downstream wave breaking despite this being more common in September–November (Ndarana and Waugh 2011).
Similarly, it appears that the importance of transient troughs to CAB breakdown is mediated by trough energetics (i.e., the magnitude of upper-level vorticity anomalies). We view the TR cases, accounting for 15% of 1979–2018 CAB breakdowns, as a subset of the most energetic transition season troughs over southern Africa, and show that this is related to upstream Rossby wave source activity in the SACZ. This convection excites a Rossby wave that propagates downstream and superposes with an extratropical wave to create a highly energetic trough that causes CAB breakdown.
5. Discussion
a. Low-latitude rainfall and upper-level troughs
The CAB is a distinct feature of the tropical edge, marking the southern limit of the tropical rain belt at around 10°–15°S. The demonstration of a midlatitude control on CAB breakdown highlights the significant equatorward extent of midlatitude influences and shows midlatitude control of transition-season rainfall across a significant geographical area well north of the traditionally defined latitudes of midlatitude precipitation (Fig. 1).
The review of Knippertz (2007) outlines how penetration of upper-level troughs into the tropics is frequently associated with enhanced convection, with the climatology of de Vries (2021) building on this further by providing a global perspective on the importance of Rossby wave breaking to subtropical precipitation extremes. Here we demonstrate the importance of PV troughs and Rossby wave breaking to not only anomalous subtropical precipitation but also to the seasonal cycle of rainfall and onset timing over central southern Africa. Extratropical disturbances are associated with patterns of weather where CAB breakdown occurs over a time scale of a few days, facilitating a transition from a dry subtropical to wet tropical regime over the region and unlocking the possibility of further tropical–extratropical interactions.
An outstanding question raised by Knippertz (2007) is the extent to which such anomalous low-latitude precipitation is a direct causal result of extratropical upper-level troughs. Put differently, in the instances of PV–trough related CAB breakdowns does wave amplification and breaking directly cause convection and rainfall, or does convection and latent heating cause wave amplification and breaking? This is an area that requires further investigation. Idealized dry simulations similar to that of Matthews and Kiladis (2000) in the tropical Pacific and Davidson et al. (2007) over the Australian monsoon region may be fruitful in teasing apart the relative roles of dry wave dynamics and diabatic heating associated with convection in the synoptic-scale processes important to CAB breakdown.
b. Implications for onset prediction
The results of this work have implications for the subseasonal prediction of onset over southern Africa and our understanding of the limits to that predictability. The midlatitudes are often considered a stochastic influence on the extratropics; although dynamics are represented well under quasigeostrophic theory they are generally predictable only for short lead times. The chaotic nature of the atmosphere means inevitable initial condition errors go on to influence all scales of motion in a finite time, so that explicit determination at the synoptic scale is not possible beyond a threshold of about 2 weeks (Lorenz 1963, 1984). Given the implication of midlatitude Rossby waves in a significant proportion of breakdown events, the prediction of onset at long lead times may be challenging.
Seasonal and subseasonal prediction typically relies on the fact that certain circulation types have intrinsically longer predictability times than individual synoptic weather patterns, and that predictable longer-term variability influences the probability of certain modes of atmospheric circulation and the probabilistic distribution of possible outcomes (Palmer and Anderson 1994). This may be true of the midlatitude circulations important to breakdown. The Southern Annular Mode is a control on mean jet latitude, while La Niña has been shown to increase the occurrence of a distinct subtropical jet over the region compared to El Niño (Gillett et al. 2006; Hart et al. 2018). The breakdown of the stratospheric polar vortex is predictable at longer lead times and is associated with a well-documented shift toward a negative Southern Annular Mode (Shen et al. 2020). If a relationship between these modes of variability and the probability of occurrence of the Rossby waves identified in section 4c can be shown, there is potential for the derivation of probabilistic predictability statements related to breakdown and onset.
c. Future change
Coupled Model Intercomparison Project (CMIP) Phase 3, 5, and 6 models simulate a statistically robust rainfall decline over southern Africa in the twenty-first century as a response to rising global temperatures, with models dynamically forcing this drying trend through a later average CAB breakdown date (Shongwe et al. 2009; He and Soden 2017; Howard and Washington 2020). The implication of the midlatitudes in CAB breakdown raises novel questions about this future drying.
A number of studies have outlined an expected poleward shift in the midlatitude westerly jets under climate change, with Barnes and Hartmann (2012) showing that this poleward shift of the Southern Hemisphere midlatitude jet is accompanied by a poleward shift in anticyclonic wave breaking in all GCMs across multiple climate forcing scenarios. Such a trend has been identified in the Northern Hemisphere historical record by Jing and Banerjee’s (2018) study of changes in RWB from 1981 to 2015 in MERRA2. It is plausible that a poleward shift in the westerly jets and consequently RWB may lead to a reduced role of midlatitude breakdown processes in the future, contributing to a later mean CAB breakdown and rainy season onset over southern Africa. This is particularly true given the importance of the location of wave breaking in determining impact on CAB breakdown. A minor change in jet dynamics could impact the frequency of wave breaking upstream of southern Africa and consequently have large impacts on RWB-related CAB breakdowns. Some of the regional rainfall impacts of a poleward shift in RWB have been considered by Ivanciu et al. (2022), who modeled the impact of poleward shifting RWB on southern African ridging highs and found both a decreasing frequency and southward shift in the area of associated rainfall. The contribution of midlatitude future change to CAB breakdown dynamics therefore represents an exciting area for further study.
6. Conclusions
This work has diagnosed a midlatitude influence on CAB breakdown and rainfall onset over the core of the southern African subtropics. Specifically, it has shown that propagating and breaking low-wavenumber Rossby waves in the Southern Hemisphere westerly jet can force CAB breakdown and a transition from a dry pre-onset regime to one characterized by broad precipitation over southern Africa. This demonstration of a remote midlatitude influence on CAB breakdown is novel and has important implications for how rainfall onset is considered at the tropical edge.
Analysis showed that that >70% of large-scale CAB breakdown events from 1979 to 2018 were related to midlatitude Rossby waves, cooccurring with strong troughs (18%) or RWB (53%) (RQ1). The proportion of midlatitude associated breakdowns decreases through the onset period from 100% in August (n = 3) to 46% in November (n = 11), suggesting a decreasing seasonal importance of the midlatitudes to breakdown as tropical influences become more important.
The types of Rossby waves important to breakdown include energetic large amplitude low-wavenumber troughs, and troughs exhibiting anticyclonic Rossby wave breaking (RQ2). The former appear to be related to anomalous SACZ convection, while the latter is associated with a deeply extratropical ridge–trough structure in the South Atlantic.
Midlatitude dynamics contribute to breakdown through (i) the replacement of broad south-of-CAB tropospheric subsidence with synoptic ascent from either RWB or cloud bands and (ii) modification of the climatological wind field leading to a reduction in convergence at the CAB and its southward advance, along with greater moisture flux from the Indian Ocean from ridging highs and convergence in the continental interior (RQ3). Our results suggest that importance of Rossby wave breaking to CAB breakdown is mediated by the location of wave breaking, while the importance of transient troughs to CAB breakdown is mediated by trough energetics.
The demonstration of a midlatitude influence on the CAB has implications for global understandings of the general circulation, highlighting the significant equatorward extent of dynamical midlatitude influences and precipitation control. The novel demonstration of a remote midlatitude influence on onset timing over southern Africa adds to the body of evidence for midlatitude precipitation control at the tropical edge.
Multiple subtropical regions, such as southern Africa, are highly vulnerable to current and future rainfall variability. It is clear that an integrated consideration of both tropical and midlatitude influence and future change is essential for rigorous understandings of present-day and future precipitation variability.
Acknowledgments.
The first author is funded by the Keble College Bigg Scholarship in African Climate Science. Professor Richard Washington is partly funded by the NERC Decreasing Rainfall to Year 2100—Role of the Congo Air Boundary (DRY-CAB) Grant (NE/V011928/1). The authors thank Kitty Attwood, Dr. Callum Munday, Dr. Emma Howard, Dr. Neil Hart, and Dr. Sebastian Engelstaeder for helpful discussions. They would also like to thank the three reviewers of this manuscript for a number of thought-provoking comments and their suggestions for improving manuscript presentation and clarity.
Data availability statement.
Results were generated using Copernicus Climate Change Service Information 2020 ECMWF Reanalysis 5 (ERA5), available at https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5. The outgoing longwave radiation–daily CDR used in this study can be acquired from NOAA’s National Center for Environmental Information (http://www.ncei.noaa.gov) and was originally developed by Hai-Tien Lee and colleagues for NOAA’s CDR Program. The CHIRPS v2 precipitation product used for Fig. 1 is available at https://data.chc.ucsb.edu/products/CHIRPS-2.0/. The canny edge CAB detection method of Howard and Washington (2019) is available from a GitHub repository at https://github.com/EmmaHoward/drylines.
REFERENCES
Barimalala, R., R. C. Blamey, F. Desbiolles, and C. J. C. Reason, 2021: The influence of southeastern African river valley jets on regional rainfall. Climate Dyn., 57, 2905–2920, https://doi.org/10.1007/s00382-021-05846-1.
Barnes, E. A., and D. L. Hartmann, 2012: Detection of Rossby wave breaking and its response to shifts of the midlatitude jet with climate change. J. Geophys. Res., 117, D09117, https://doi.org/10.1029/2012JD017469.
Conway, D., and Coauthors, 2015: Climate and southern Africa’s water–energy–food nexus. Nat. Climate Change, 5, 837–846, https://doi.org/10.1038/nclimate2735.
Davidson, N. E., K. J. Tory, M. J. Reeder, and W. L. Drosdowsky, 2007: Extratropical–tropical interaction during onset of the Australian monsoon: Reanalysis diagnostics and idealized dry simulations. J. Atmos. Sci., 64, 3475–3498, https://doi.org/10.1175/JAS4034.1.
de Vries, A. J., 2021: A global climatological perspective on the importance of Rossby wave breaking and intense moisture transport for extreme precipitation events. Wea. Climate Dyn., 2, 129–161, https://doi.org/10.5194/wcd-2-129-2021.
Favre, A., B. Hewitson, C. Lennard, R. Cerezo-Mota, and M. Tadross, 2013: Cut-off lows in the South Africa region and their contribution to precipitation. Climate Dyn., 41, 2331–2351, https://doi.org/10.1007/s00382-012-1579-6.
Funatsu, B. M., and D. W. Waugh, 2008: Connections between potential vorticity intrusions and convection in the eastern tropical Pacific. J. Atmos. Sci., 65, 987–1002, https://doi.org/10.1175/2007JAS2248.1.
Gillett, N. P., T. D. Kell, and P. D. Jones, 2006: Regional climate impacts of the Southern Annular Mode. Geophys. Res. Lett., 33, L23704, https://doi.org/10.1029/2006GL027721.
Grimm, A. M., and C. J. C. Reason, 2015: Intraseasonal teleconnections between South America and South Africa. J. Climate, 28, 9489–9497, https://doi.org/10.1175/JCLI-D-15-0116.1.
Hart, N. C. G., C. J. C. Reason, and N. Fauchereau, 2010: Tropical–extratropical interactions over southern Africa: Three cases of heavy summer season rainfall. Mon. Wea. Rev., 138, 2608–2623, https://doi.org/10.1175/2010MWR3070.1.
Hart, N. C. G., C. J. C. Reason, and N. Fauchereau, 2013: Cloud bands over southern Africa: Seasonality, contribution to rainfall variability and modulation by the MJO. Climate Dyn., 41, 1199–1212, https://doi.org/10.1007/s00382-012-1589-4.
Hart, N. C. G., R. Washington, and C. J. C. Reason, 2018: On the likelihood of tropical–extratropical cloud bands in the South Indian convergence zone during ENSO events. J. Climate, 31, 2797–2817, https://doi.org/10.1175/JCLI-D-17-0221.1.
He, J., and B. J. Soden, 2017: A re-examination of the projected subtropical precipitation decline. Nat. Climate Change, 7, 53–57, https://doi.org/10.1038/nclimate3157.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Holton, J., 2012: An Introduction to Dynamic Meteorology. 5th ed. International Geophysics Series, Vol. 88, Academic Press, 552 pp.
Hoskins, B. J., 2015: Potential vorticity and the PV perspective. Adv. Atmos. Sci., 32, 2–9, https://doi.org/10.1007/s00376-014-0007-8.
Howard, E., and R. Washington, 2019: Drylines in southern Africa: Rediscovering the Congo air boundary. J. Climate, 32, 8223–8242, https://doi.org/10.1175/JCLI-D-19-0437.1.
Howard, E., and R. Washington, 2020: Tracing future spring and summer drying in southern Africa to tropical lows and the Congo air boundary. J. Climate, 33, 6205–6228, https://doi.org/10.1175/JCLI-D-19-0755.1.
Howard, E., R. Washington, and K. I. Hodges, 2019: Tropical lows in southern Africa: Tracks, rainfall contributions, and the role of ENSO. J. Geophys. Res. Atmos., 124, 11 009–11 032, https://doi.org/10.1029/2019JD030803.
Hua, W., L. Zhou, S. E. Nicholson, H. Chen, and M. Qin, 2019: Assessing reanalysis data for understanding rainfall climatology and variability over central equatorial Africa. Climate Dyn., 53, 651–669, https://doi.org/10.1007/s00382-018-04604-0.
Ivanciu, I., T. Ndarana, K. Matthes, and S. Wahl, 2022: On the ridging of the South Atlantic anticyclone over South Africa: The impact of Rossby wave breaking and of climate change. Geophys. Res. Lett., 49, e2022GL099607, https://doi.org/10.1029/2022GL099607.
Jacobeit, J., 1987: Variations of trough positions and precipitation patterns in the Mediterranean area. J. Climatol., 7, 453–476, https://doi.org/10.1002/joc.3370070503.
Jing, P., and S. Banerjee, 2018: Rossby wave breaking and isentropic stratosphere-troposphere exchange during 1981–2015 in the Northern Hemisphere. J. Geophys. Res. Atmos., 123, 9011–9025, https://doi.org/10.1029/2018JD028997.
Kanemasu, E. T., J. I. Stewart, S. J. van Donk, and S. M. Virmani, 1990: Agroclimatic approaches for improving agricultural productivity in semiarid tropics. Advances in Soil Science: Dryland Agriculture: Strategies for Sustainability, R. P. Singh, J. F. Parr, and B. A. Stewart, Eds., Advances in Soil Science, Vol. 13, Springer, 273–309, https://doi.org/10.1007/978-1-4613-8982-8_14.
Knippertz, P., 2007: Tropical–extratropical interactions related to upper-level troughs at low latitudes. Dyn. Atmos. Oceans, 43, 36–62, https://doi.org/10.1016/j.dynatmoce.2006.06.003.
Kniveton, D. R., R. Layberry, C. J. R. Williams, and M. Peck, 2009: Trends in the start of the wet season over Africa. Int. J. Climatol., 29, 1216–1225, https://doi.org/10.1002/joc.1792.
Kuete, G., W. P. Mba, and R. Washington, 2020: African Easterly Jet South: Control, maintenance mechanisms and link with southern subtropical waves. Climate Dyn., 54, 1539–1552, https://doi.org/10.1007/s00382-019-05072-w.
Lee, H.-T., 2014: NOAA Climate Data Record (CDR) of Daily Outgoing Longwave Radiation (OLR), version 1.2. NOAA, accessed 28 June 2023, https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ncdc:C00875.
Liebmann, B., I. Bladé, G. N. Kiladis, L. M. V. Carvalho, G. B. Senay, D. Allured, S. Leroux, and C. Funk, 2012: Seasonality of African precipitation from 1996 to 2009. J. Climate, 25, 4304–4322, https://doi.org/10.1175/JCLI-D-11-00157.1.
Lorenz, E. N., 1963: Deterministic nonperiodic flow. J. Atmos. Sci., 20, 130–141, https://doi.org/10.1175/1520-0469(1963)020<0130:DNF>2.0.CO;2.
Lorenz, E. N., 1984: Irregularity: A fundamental property of the atmosphere. Tellus, 36A, 98–110, https://doi.org/10.1111/j.1600-0870.1984.tb00230.x.
Macron, C., B. Pohl, Y. Richard, and M. Bessafi, 2014: How do tropical temperate troughs form and develop over southern Africa? J. Climate, 27, 1633–1647, https://doi.org/10.1175/JCLI-D-13-00175.1.
Matthews, A. J., and G. N. Kiladis, 2000: A model of Rossby waves linked to submonthly convection over the eastern tropical Pacific. J. Atmos. Sci., 57, 3785–3798, https://doi.org/10.1175/1520-0469(2000)057<3785:AMORWL>2.0.CO;2.
McIntyre, M. E., and T. N. Palmer, 1983: Breaking planetary waves in the stratosphere. Nature, 305, 593–600, https://doi.org/10.1038/305593a0.
Moses, O., R. C. Blamey, and C. J. C. Reason, 2023: Drought metrics and temperature extremes over the Okavango River basin, southern Africa, and links with the Botswana high. Int. J. Climatol., 43, 6463–6483, https://doi.org/10.1002/joc.8215.
Munday, C., R. Washington, and N. Hart, 2021: African low-level jets and their importance for water vapor transport and rainfall. Geophys. Res. Lett., 48, e2020GL090999, https://doi.org/10.1029/2020GL090999.
Ndarana, T., and D. W. Waugh, 2010: The link between cut-off lows and Rossby wave breaking in the Southern Hemisphere. Quart. J. Roy. Meteor. Soc., 136, 869–885, https://doi.org/10.1002/qj.627.
Ndarana, T., and D. W. Waugh, 2011: A climatology of Rossby wave breaking on the Southern Hemisphere tropopause. J. Atmos. Sci., 68, 798–811, https://doi.org/10.1175/2010JAS3460.1.
Ndarana, T., M.-J. Bopape, D. Waugh, and L. Dyson, 2018: The influence of the lower stratosphere on ridging Atlantic Ocean anticyclones over South Africa. J. Climate, 31, 6175–6187, https://doi.org/10.1175/JCLI-D-17-0832.1.
Ndarana, T., S. Mpati, M.-J. Bopape, F. Engelbrecht, and H. Chikoore, 2021: The flow and moisture fluxes associated with ridging South Atlantic Ocean anticyclones during the subtropical southern African summer. Int. J. Climatol., 41, E1000–E1017, https://doi.org/10.1002/joc.6745.
Ndarana, T., T. S. Rammopo, C. J. C. Reason, M.-J. Bopape, F. Engelbrecht, and H. Chikoore, 2022: Two types of ridging South Atlantic Ocean anticyclones over South Africa and the associated dynamical processes. Atmos. Res., 265, 105897, https://doi.org/10.1016/j.atmosres.2021.105897.
Palmer, T. N., and D. L. T. Anderson, 1994: The prospects for seasonal forecasting—A review paper. Quart. J. Roy. Meteor. Soc., 120, 755–793, https://doi.org/10.1002/qj.49712051802.
Phakula, S., W. A. Landman, and A. F. Beraki, 2018: Forecasting seasonal rainfall characteristics and onset months over South Africa. Int. J. Climatol., 38, e889–e900, https://doi.org/10.1002/joc.5417.
Postel, G. A., and M. H. Hitchman, 1999: A climatology of Rossby wave breaking along the subtropical tropopause. J. Atmos. Sci., 56, 359–373, https://doi.org/10.1175/1520-0469(1999)056<0359:ACORWB>2.0.CO;2.
Ratnam, J. V., T. Doi, W. A. Landman, and S. K. Behera, 2018: Seasonal forecasting of onset of summer rains over South Africa. J. Appl. Meteor. Climatol., 57, 2697–2711, https://doi.org/10.1175/JAMC-D-18-0067.1.
Reason, C. J. C., and M. Rouault, 2005: Links between the Antarctic Oscillation and winter rainfall over western South Africa. Geophys. Res. Lett., 32, L07705, https://doi.org/10.1029/2005GL022419.
Reason, C. J. C., S. Hachigonta, and R. F. Phaladi, 2005: Interannual variability in rainy season characteristics over the Limpopo region of southern Africa. Int. J. Climatol., 25, 1835–1853, https://doi.org/10.1002/joc.1228.
Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45, 1228–1251, https://doi.org/10.1175/1520-0469(1988)045<1228:TGOGRF>2.0.CO;2.
Shen, X., L. Wang, and S. Osprey, 2020: Tropospheric forcing of the 2019 Antarctic sudden stratospheric warming. Geophys. Res. Lett., 47, e2020GL089343, https://doi.org/10.1029/2020GL089343.
Shongwe, M. E., G. J. van Oldenborgh, B. J. J. M. van den Hurk, B. de Boer, C. A. S. Coelho, and M. K. van Aalst, 2009: Projected changes in mean and extreme precipitation in Africa under global warming. Part I: Southern Africa. J. Climate, 22, 3819–3837, https://doi.org/10.1175/2009JCLI2317.1.
Spavins-Hicks, Z. D., R. Washington, and C. Munday, 2021: The Limpopo low-level jet: Mean climatology and role in water vapor transport. J. Geophys. Res. Atmos., 126, e2020JD034364, https://doi.org/10.1029/2020JD034364.
Tadross, M. A., B. C. Hewitson, and M. T. Usman, 2005: The interannual variability of the onset of the maize growing season over South Africa and Zimbabwe. J. Climate, 18, 3356–3372, https://doi.org/10.1175/JCLI3423.1.
Tennant, W., 2004: Considerations when using pre-1979 NCEP/NCAR reanalyses in the Southern Hemisphere. Geophys. Res. Lett., 31, L11112, https://doi.org/10.1029/2004GL019751.
Thurlow, J., T. Zhu, and X. Diao, 2012: Current climate variability and future climate change: Estimated growth and poverty impacts for Zambia. Rev. Dev. Econ., 16, 394–411, https://doi.org/10.1111/j.1467-9361.2012.00670.x.
Torrance, J. D., 1979: Upper windflow patterns in relation to rainfall in south-east central Africa. Weather, 34, 106–115, https://doi.org/10.1002/j.1477-8696.1979.tb03418.x.
Ward, N., A. H. Fink, R. J. Keane, F. Guichard, J. H. Marsham, D. J. Parker, and C. M. Taylor, 2021: Synoptic timescale linkage between midlatitude winter troughs Sahara temperature patterns and northern Congo rainfall: A building block of regional climate variability. Int. J. Climatol., 41, 3153–3173, https://doi.org/10.1002/joc.7011.
