Abstract

Teleconnection of climate anomalies between various parts of the tropics and extratropics is a well-established feature of the climate system. Building on previous work showing that a teleconnection exists between the South American monsoon system and interannual summer rainfall variability over southern Africa, this study considers intraseasonal variability over these landmasses. It is shown that strong teleconnections exist between South African daily rainfall and that over various areas of South America, with the latter leading by 4–5 days, for both winter and summer, involving regions with strong rainfall in these seasons. During the summer, the mechanisms involve both a modulation of the local Walker cell as well as extratropical Rossby wave trains. For winter, the latter mechanism is more important. While in summer tropical convective anomalies over South America play an important role, in winter the subtropics become more important. In both cases, these modulations lead to regional changes in circulation over southern Africa that are favorable for the dominant synoptic rainfall-producing weather systems such as cutoff lows and tropical extratropical cloud bands.

1. Introduction

Situated in the subtropics, South Africa mainly receives rainfall during austral summer (Fig. 1) except in the southwest (winter rainfall) and the south coast (all-season rainfall). Characteristic of southern African climate is marked variability on intraseasonal (e.g., Pohl et al. 2007; Mapande and Reason 2005), interannual (e.g., Lindesay 1988; Mason and Jury 1997; Reason et al. 2006), interdecadal (e.g., Tyson et al. 1975; Reason and Rouault 2002; Allan et al. 2003), and longer time scales.

Fig. 1.

(top) Selected regions and annual cycles of precipitation in South Africa. (middle)–(bottom) The 1° boxes in South America with precipitation significantly correlated to the lagged precipitation (5 days) in each selected region in South Africa. Dark squares (triangles) indicate confidence levels higher than 90% for positive (negative) correlation; open squares (triangles) are for confidence levels between 85% and 90%. Ellipses indicate the regions with maximum correlation. White areas are void of data. The highest correlation coefficient for box 1 is 0.48, and for boxes 2 and 3, it is 0.42.

Fig. 1.

(top) Selected regions and annual cycles of precipitation in South Africa. (middle)–(bottom) The 1° boxes in South America with precipitation significantly correlated to the lagged precipitation (5 days) in each selected region in South Africa. Dark squares (triangles) indicate confidence levels higher than 90% for positive (negative) correlation; open squares (triangles) are for confidence levels between 85% and 90%. Ellipses indicate the regions with maximum correlation. White areas are void of data. The highest correlation coefficient for box 1 is 0.48, and for boxes 2 and 3, it is 0.42.

Grimm and Reason (2011) suggested an interannual teleconnection between Brazilian and southern African rainfall during summers when there are Benguela Niño (Niña) events, anomalous warm (cool) sea surface temperature (SST) in the South Atlantic near Angola and Namibia. This teleconnection involves anomalous rainfall during particular South American monsoon seasons, which generate Rossby wave trains across the Atlantic and, hence, affect southern African circulation and rainfall.

Here, we focus on intraseasonal time scales, less well studied over South Africa than interannual variability, and on teleconnections between South American rainfall and that in South Africa, although it is recognized that there are other important influences such as ENSO. There is significant intraseasonal variability over South America and the Pacific (e.g., Paegle et al. 2000) that can modulate synoptic-scale variability. The enhanced synoptic anomalies can then produce teleconnections that perturb weather and circulation over Africa. Evidence is given of such teleconnections during summer (December–February) and winter (June–August) for those parts of South Africa that receive significant rainfall during these seasons.

2. Data and methods

Observed daily gauge precipitation data for 1970–99 are gridded to 1° resolution for South America and 2.5° for South Africa. At each grid point, anomalies of daily precipitation are calculated and submitted to a bandpass Lanczos filter to isolate intraseasonal oscillations in the 20–90-day band. For each season, the filtered precipitation anomalies for the South African grid boxes are correlated with filtered precipitation anomalies in the grid boxes over South America. Lags from 0 up to 12 days are applied to the South African data, in order to investigate convection anomalies over South America that could produce atmospheric perturbations associated with South African precipitation anomalies. The significance of correlation between the filtered data takes autocorrelation into account and uses an effective sample size calculated according to Dawdy and Matalas (1964).

Although significant correlations are revealed for several regions, the results shown represent the best correlations for different precipitation regimes: the Limpopo (summer), southwestern Cape (winter), and South Coast (all-season) (Fig. 1). The correlation maps show the South American grid boxes with significant correlation, to confidence levels of 85% and 90%. The isolines of correlation coefficients are not shown, to keep the figures cleaner. The correlation coefficients may be slightly affected by different quantities of data at each grid point, since there may be different missing daily data at each grid point and each season. On the other hand, the significance levels take into account the actual number of data in each grid box and therefore are comparable. Nevertheless, it is convenient to give an idea of the magnitude of the correlation coefficients corresponding to given confidence levels, and their maximum value for each case. It may vary a little for each grid box (because of different autocorrelations). When using around 2500 data points, the correlation coefficients corresponding to confidence levels of 85% and 90% are 0.19 and 0.23, respectively. The maximum correlation coefficient for box 1 (summer) is 0.48 and for boxes 2 and 3 (winter) it is 0.42.

NCEP–NCAR reanalyses (Kalnay et al. 1996) are used to composite intraseasonal anomalies in OLR, 200-hPa streamfunction, and vertically integrated moisture flux associated with South African positive filtered precipitation anomalies above one standard deviation (positive phases). The OLR composite anomalies are also calculated for 5 days before the days of the positive phase, in order to capture the previous OLR anomalies over and near South America. The statistical significance of the composite analysis is calculated with a t test. Since the daily anomalies are filtered, they have some serial dependence, and therefore, the effective sample size must take into consideration the autocorrelation of the series (Wilks 1995).

The possible origin of the atmospheric circulation anomalies associated with those positive phases is determined using influence functions (IFs) of a vorticity equation model with a divergence source (Grimm and Silva Dias 1995a, b). The model is linearized about a realistic basic state and includes the divergence of this state and vorticity advection by divergent wind:

 
formula

where

 
formula

Here F′ depends only on the anomalous divergent flow. In these equations, ζ is absolute vorticity; D is divergence; Vχ and Vψ are the divergent and rotational components of the wind, respectively; and A′ is the damping term, including linear damping and biharmonic diffusion. The model is applied at 200 hPa, near the level of maximum divergence associated with convective outflow in the tropics and an equivalent barotropic level in the extratropics. Its stationary version may be written as

 
formula

where M is a linear operator and ψ′ is the anomalous streamfunction. Then the IF based on divergence forcing is defined by

 
formula

where is the delta function. Thus, the IF for a target point with longitude and latitude (λ, φ) is, at each point (λ′, φ′), equal to the model response at (λ, φ) to an upper-level divergence located at (λ′, φ′). Maps with contours of IF for a given target point indicate the regions in which the anomalous upper-level divergence is most efficient in producing streamfunction anomalies around that target point. Upper-level anomalous divergence (convergence) in regions with positive values of the IF produces positive (negative) streamfunction anomalies around the target point, and the opposite is true for negative values of the IF. More information about the model and the IFs, their usefulness and drawbacks can be found in Grimm and Silva Dias (1995a, b). This model is also used for simple simulations of the observed streamfunction anomalies at 200 hPa in response to anomalous convection over South America.

Here, the IFs are shown only for the action centers of circulation anomalies directly associated with anomalous convection in southern Africa, usually cyclonic anomalies southwest of the analyzed regions. Before the IF analysis, selection of regions whose upper-level anomalous divergence might contribute to generate streamfunction anomalies leading to anomalous South African precipitation is primarily based on the correlation patterns in Fig. 1 (regions within the ellipses). The OLR composite anomalies calculated for 5 days before the days of positive phase in South Africa (Figs. 2b, 3b, and 4b) are used only to confirm the results of the correlations, and to add information about regions within or nearby South America (e.g., the Atlantic) where precipitation data are not available and anomalous convection might be important for the teleconnections. It is convenient to remark that these OLR anomaly composites for 5 days before also show enhanced convection in South Africa because the anomalies are filtered (smoothed) by a 20–90-day bandpass filter.

3. Summer teleconnections

There are significant correlations for a lag of few days between filtered daily precipitation anomalies in several summer rainfall regions of South Africa and South America [for precipitation regimes in South America, see Grimm (2011)]. As an example, we use box 1, in the north of the country (Fig. 1).

For box 1, the strongest correlation is at a 4–5-day lag with rainfall over parts of northeastern and central Brazil. The OLR anomalies for positive phases in box 1 show a northwest–southeast (NW–SE)-oriented connection of negative values between tropical southern Africa and the southwestern Indian Ocean (Fig. 2a), consistent with the formation of tropical-temperate troughs (TTT), the main summer synoptic rainfall system (Harrison 1984; Hart et al. 2010, 2013). Over South America 5 days before (Fig. 2b), they are predominantly negative (positive) within the ellipse with positive (negative) correlations in Fig. 1, coherent with enhanced (suppressed) precipitation associated with positive phases in box 1.

Fig. 2.

Anomaly composites for the days of positive phases in box 1, in summer: (a) OLR and (b) OLR 5 days before (contour interval 1.5 W m−2), (c) 200-hPa streamfunction (contour interval 0.3 × 106 m2 s−1), and (d) vertically integrated moisture flux (arrows, m g s−1 kg−1) and its divergence (contour interval is 0.75 × 10−2 g s−1 kg−1). Dark (light) gray shading indicates confidence levels higher than 90% for negative (positive) anomalies, and only significant moisture fluxes are shown. Zero isolines are not shown. (e) Influence function for summer basic state for the easternmost action center numbered in (c). The values shown in each location are proportional to the streamfunction response at the target point to a unitary upper-level divergence anomaly in this location. (f) The anomalous 200-hPa prescribed divergence and (g) corresponding steady anomalous streamfunction, with the zonal components removed, for the box 1 experiment.

Fig. 2.

Anomaly composites for the days of positive phases in box 1, in summer: (a) OLR and (b) OLR 5 days before (contour interval 1.5 W m−2), (c) 200-hPa streamfunction (contour interval 0.3 × 106 m2 s−1), and (d) vertically integrated moisture flux (arrows, m g s−1 kg−1) and its divergence (contour interval is 0.75 × 10−2 g s−1 kg−1). Dark (light) gray shading indicates confidence levels higher than 90% for negative (positive) anomalies, and only significant moisture fluxes are shown. Zero isolines are not shown. (e) Influence function for summer basic state for the easternmost action center numbered in (c). The values shown in each location are proportional to the streamfunction response at the target point to a unitary upper-level divergence anomaly in this location. (f) The anomalous 200-hPa prescribed divergence and (g) corresponding steady anomalous streamfunction, with the zonal components removed, for the box 1 experiment.

The 200-hPa streamfunction cyclonic (anticyclonic) anomalies behind (ahead) box 1 are coherent with a TTT (Fig. 2c), being part of a wave train that originated from the strongest South American convective anomalies. In the tropics, there is also a Walker cell–like pattern across the South American–African sector (pairs of anticyclones/cyclones either side of the equator). Although the filtered data do not contain isolated synoptic rainfall systems, their positive phases occur when these systems are more intense and frequent, and therefore, the anomalies during these phases are representative of them. The moisture flux anomaly (Fig. 2d) shows increased inflow of moist marine air from the southwestern Indian Ocean over Mozambique toward South Africa. Anomalous moisture flux also occurs from the Atlantic, through Angola/Namibia, consistent with Cook et al. (2004). Relative convergence occurs over South Africa, Zimbabwe, and Botswana, including the TTT region. The cyclonic moisture flux anomaly over western southern Africa and the southeastern Atlantic favors both increased uplift over subtropical southern Africa and inflow of warm tropical air from the tropical South Atlantic toward the TTT source region, the Angola low (Reason et al. 2006). An enhanced cloud band is evident in the negative OLR anomaly band (Fig. 2a) that stretches southeastward from southeastern Angola, across the Limpopo box to the Agulhas Current region where it joins with another band of relative uplift. TTT formation is triggered by the arrival of an upper-level trough over southern Africa and associated planetary waves (Hart et al. 2010), which may originate from South America.

A first indication of South American influence is the wave train visible in Fig. 2c, originating south of the Brazilian precipitation anomalies (Fig. 1). As a second indication, the IF for the action center 4 of this wave train (Fig. 2e) shows clearly that the anomalous upper-level divergence and convergence in the ellipses in Fig. 1 are in the right position to produce cyclonic anomalies around point 4. The IF is positive (negative) in the region with anomalous upper-level divergence (convergence) over central-eastern (southeastern) South America, indicating that both positive and negative divergence anomalies are able to produce cyclonic anomalies (positive streamfunction anomalies) around point 4. Although there are also OLR anomalies in the Pacific and the Atlantic, either they are in regions with weaker IF or the associated upper-level divergence does not have the correct sign, according to the IF, to produce a cyclonic anomaly around point 4. No other regions with strong OLR anomalies have such strong IF values with such coherent signs. A third indication of the influence of South America anomalous convection is the upper-level streamfunction response (Fig. 2g) to upper-level divergence anomalies in the marked ellipses (Figs. 1 and 2f). The wave trains in Figs. 2c and 2g are similar, although complete agreement is not possible with such an idealized experiment.

Our results are also consistent with Macron et al. (2014), who showed that not all extratropical waves produce a TTT; Rossby waves leading to TTT systems are already stronger and show associated enhanced convection over the subtropical southern Atlantic and eastern South America prior to the TTT development over southern Africa, indicating the contribution of a wave train that originated in the tropics.

4. Winter teleconnections

Western South Africa receives mainly winter rainfall (Fig. 1). The only other region in southern Africa that has significant winter rainfall is the south coast. Both these regions have a significant relationship on intraseasonal scales with rainfall upstream over certain South American regions.

For the winter rainfall southwestern region (box 2), the strongest correlation with South American winter rainfall regions is southern Brazil/Uruguay (positive), at a 4–5-day lag (Fig. 1). The OLR anomaly composite confirms the enhanced convection in southwestern South Africa (Fig. 3a) and 5 days before in a region off of southeastern South America, an important winter cyclogenesis region (Fig. 3b). Most of the rainfall in southwestern South Africa is brought by cold fronts, with cutoff lows that are sometimes important. Figure 3c shows a wave pattern in the midlatitudes, similar to that identified by Weldon and Reason (2014) as important for southern South African rainfall, with large cyclonic features over western South Africa and southern Argentina and Chile. The latter extends over the western South Atlantic important area for cyclogenesis, and thus favors more frontal systems approaching southwestern South Africa. On the other hand, an anticyclonic anomaly is present near southeastern South America. These patterns are consistent with the positive correlations between box 2 rainfall and southern Brazil/Uruguay, and negative OLR anomaly composites off of southeastern South America (Fig. 3b). The cyclonic anomaly over and south of western South Africa is favorable for strengthening the approaching westerly systems. A similar pattern is evident in the moisture flux (Fig. 3d), indicating its barotropic characteristics, with enhanced westerly/northwesterly inflow from the South Atlantic toward western South Africa and relative moisture convergence there, favorable for increased rainfall (Reason et al. 2002; Reason and Jagadheesha 2005). Negative OLR anomalies (uplift) predominate upstream over the west coast of South Africa/Namibia and the adjacent ocean area (Fig. 3a).

Fig. 3.

As in Fig. 2, but for box 2 in winter.

Fig. 3.

As in Fig. 2, but for box 2 in winter.

The streamfunction anomalies associated with the positive phases of the winter precipitation in box 2 (Fig. 3c) could be attributed to a wave train circling the globe and its migratory extratropical cyclones. However, one can show that the anomalous convection over southeastern South America and in the cyclogenetic region off the southern Brazil/Uruguay coast is able to influence the circulation anomalies that lead to stronger South African precipitation. There are three factors that suggest that the relationship between South American and South African rainfall is not simply due to an extratropical wave train across the Southern Hemisphere that equally influences South American and South African rainfall, without influence of the South American anomalous convection. First, the streamfunction anomalies are much stronger in the Atlantic than in the Pacific and Indian Oceans. Second, the IF of the action center 2 of the cyclonic anomaly near southern South Africa in Fig. 3c indicates very clearly that anomalous convection (and upper-level divergence) over southeastern South America and the cyclogenetic region off the coast is very efficient in producing this cyclonic anomaly (Fig. 3e). Third, the simple model simulation with a forcing over the maximum correlation for box 2 (Fig. 1) and maximum OLR negative anomalies (Fig. 3b), as indicated in Fig. 3f, reproduces the main circulation features observed over the Atlantic (around centers 1 and 2 in Fig. 3c), as shown in Fig. 3g. These features are important to produce rainfall over box 2. Even the anticyclonic anomalies over tropical southern Africa extending toward Madagascar and the cyclonic ones over Indian Ocean are reproduced. These results indicate that if a wave train passes over southern South America without causing strong convective anomalies over the southeastern continent and off the coast, there is likely to be no strong anomaly over southern South Africa. These connections may potentially be useful for prognostic purposes.

For the south coast (box 3), there is strong positive correlation at a 4–5-day lag with southeastern and central Brazil; a negative correlation appears over northern Argentina and southern Brazil, where winter rainfall is strong (Fig. 1). Negative (positive) OLR anomalies also appear 5 days before in these regions with positive (negative) correlations (Fig. 4b).

Fig. 4.

As in Fig. 2, but for box 3 in winter.

Fig. 4.

As in Fig. 2, but for box 3 in winter.

The streamfunction anomalies at 200 hPa (Fig. 4c) show a wave train between South America and southern Africa. Over South Africa, a large area of negative OLR anomalies (relative uplift) is apparent over the south coast and inland, except in the far southwest (Fig. 4a). Both the 850- (not shown) and 200-hPa streamfunction anomalies (Fig. 4c) are characteristic of cutoff low conditions since they have a large anticyclonic anomaly (center 2) to the south of South Africa and a cyclonic anomaly over the landmass (center 3) (Singleton and Reason 2006, 2007a). The fact that these patterns are similar at the lower and upper levels is consistent with the development of deep convective storms such as cutoff lows. Almost all of the flooding events in southern South Africa result from these weather systems (Singleton and Reason 2007b). Relative moisture flux convergence is evident over eastern South Africa, including the eastern part of the south coast where there are easterly onshore anomalies (Fig. 4d).

The influence of South American anomalous convection in forcing the wave train between the two continents is confirmed by the IF of center 3 (Fig. 4e), which displays strong positive values over southeastern and central Brazil, around 20°S, indicating that upper-level divergence in this region is able to produce a positive streamfunction around center 3. The IFs of the other centers (not shown) indicate that also the subtropical upper-level convergence is important in generating that wave train. Furthermore, a model simulation of the streamfunction anomalies at 200 hPa forced by anomalous divergence (convergence) within the continuous (dashed) ellipse on the correlation map for box 3 (Fig. 1), as represented in Fig. 4f, shows a wave train over the southern Atlantic (Fig. 4g) similar to the observed one (Fig. 4c).

5. Summary and conclusions

Grimm and Reason (2011) showed that a teleconnection exists between the South American monsoon and summer rainfall over tropical southern Africa on interannual scales. Here, evidence is shown that teleconnections exist between South American rainfall variability and that of various South African regions for both summer and winter on intraseasonal scales. Although not as strong as summer/winter, there is evidence of rainfall teleconnections for the transition seasons too. It, therefore, appears that the rainfall relationships between these two Southern Hemisphere landmasses can exist at any time of the year. Since there are teleconnections at both intraseasonal and interannual scales, they might appear on interdecadal scales, too, since there is strong interdecadal precipitation variability over South America (Grimm and Saboia 2015) as well as southern Africa (Reason and Rouault 2002).

The mechanisms by which these teleconnections occur involve the generation of wave trains across the South Atlantic that then impact on regional circulation and moisture flux convergence over South Africa. In summer, there is also an anomalous Walker-type circulation in the tropical Atlantic region, associated with anomalous tropical convection. While in summer tropical convective anomalies play an important role, in winter the subtropics become more important. Even when there is a wave pattern circling the globe in the subtropics and midlatitudes that could be associated with anomalous convection in South America and southern Africa some days after, it is possible to show that the anomalous convection over South America and the neighboring Atlantic region is able to influence South African precipitation.

Obviously there are other influences over southern Africa rainfall on interannual, interdecadal, and intraseasonal time scales. The focus of this manuscript is on the possible influence of the South American anomalous convection on the precipitation variability in southern Africa on intraseasonal time scales. The strongest relationships between the intraseasonal variability of South American and South African rainfall exist at lags of 4–5 days of the South African rainfall behind that over South America. This aspect then suggests that there may be some possibilities for improving forecasting skill of wet and dry spells over South Africa based on near-real-time monitoring of rainfall upstream over South America. Understanding and predicting dry and wet spell frequencies during the rainy season is very important for users in agriculture and other sectors of the economy (Usman and Reason 2004).

Acknowledgments

This work was supported by Program PROAFRICA (CNPq-Brazil), and IAI-CRN3035, which is supported by the U.S. NSF (Grant GEO-1128040). The authors thank Eduardo Machado and Gisele Martins for help with the figures.

REFERENCES

REFERENCES
Allan
,
R. J.
,
C. J. C.
Reason
,
J. A.
Lindesay
, and
T. J.
Ansell
,
2003
:
Protracted ENSO episodes and their impacts in the Indian Ocean region
.
Deep-Sea Res. II
,
50
,
2331
2347
, doi:.
Cook
,
C.
,
C. J. C.
Reason
, and
B. C.
Hewitson
,
2004
:
Wet and dry spells within particular wet and dry summers in the South African summer rainfall region
.
Climate Res.
,
26
,
17
31
, doi:.
Dawdy
,
D. R.
, and
N. C.
Matalas
,
1964
: Statistical and probability analysis of hydrologic data. Part III: Analysis of variance, covariance and time series. Handbook of Applied Hydrology, a Compendium of Water-Resources Technology, Ven Te Chow, Ed., McGraw-Hill Book Company, 8.68–8.90.
Grimm
,
A. M.
,
2011
:
Interannual climate variability in South America: Impacts on seasonal precipitation, extreme events and possible effects of climate change
.
Stochastic Environ. Res. Risk Assess.
,
25
,
537
554
, doi:.
Grimm
,
A. M.
, and
P. L.
Silva Dias
,
1995a
:
Use of barotropic models in the study of the extratropical response to tropical heat sources
.
J. Meteor. Soc. Japan
,
73
,
765
779
.
Grimm
,
A. M.
, and
P. L.
Silva Dias
,
1995b
:
Analysis of tropical–extratropical interactions with influence functions of a barotropic model
.
J. Atmos. Sci.
,
52
,
3538
3555
, doi:.
Grimm
,
A. M.
, and
C. J. C.
Reason
,
2011
:
Does the South American monsoon affect African rainfall?
J. Climate
,
24
,
1226
1238
, doi:.
Grimm
,
A. M.
, and
J. P. J.
Saboia
,
2015
:
Interdecadal variability of the South American precipitation in the monsoon season
.
J. Climate
,
28
,
755
775
, doi:.
Harrison
,
M. S. J.
,
1984
:
A generalized classification of South African summer rain-bearing synoptic systems
.
J. Climatol.
,
4
,
547
560
, doi:.
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
, doi:.
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
, doi:.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
Lindesay
,
J. A.
,
1988
:
South African rainfall, the Southern Oscillation and a Southern Hemisphere semi-annual cycle
.
J. Climatol.
,
8
,
17
30
, doi:.
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
, doi:.
Mapande
,
A.
, and
C. J. C.
Reason
,
2005
:
Links between rainfall variability on intraseasonal and interannual scales over western Tanzania and regional circulation and SST patterns
. Meteor. Atmos. Phys.,
89
,
215
234
, doi:.
Mason
,
S. J.
, and
M. R.
Jury
,
1997
:
Climate variability and change over southern Africa: A reflection on underlying processes
.
Prog. Phys. Geogr.
,
21
,
23
50
, doi:.
Paegle
,
J. N.
,
L. A.
Byerle
, and
K. C.
Mo
,
2000
:
Intraseasonal modulation of South American summer precipitation
.
Mon. Wea. Rev.
,
128
,
837
850
, doi:.
Pohl
,
B.
,
Y.
Richard
, and
N.
Fauchereau
,
2007
:
Influence of the Madden–Julian oscillation on southern African summer rainfall
.
J. Climate
,
20
,
4227
4242
, doi:.
Reason
,
C. J. C.
, and
M.
Rouault
,
2002
:
ENSO-like decadal variability and South African rainfall
.
Geophys. Res. Lett.
,
29
, doi:.
Reason
,
C. J. C.
, and
D.
Jagadheesha
,
2005
:
Relationships between South Atlantic SST variability and atmospheric circulation over the South African region during austral winter
.
J. Climate
,
18
,
3339
3355
, doi:.
Reason
,
C. J. C.
,
M.
Rouault
,
J.-L.
Melice
, and
D.
Jagadeesha
,
2002
: Interannual winter rainfall variability in SW South Africa and large scale ocean–atmosphere interactions. Meteor. Atmos. Phys., 80, 19–29, doi:.
Reason
,
C. J. C.
,
W.
Landman
, and
W.
Tennant
,
2006
:
Seasonal to decadal prediction of southern African climate and its links with variability of the Atlantic Ocean
.
Bull. Amer. Meteor. Soc.
,
87
,
941
955
, doi:.
Singleton
,
A. T.
, and
C. J. C.
Reason
,
2006
:
Numerical simulations of a severe rainfall event over the eastern Cape coast of South Africa: Sensitivity to sea surface temperature and topography
.
Tellus
,
58A
,
355
367
, doi:.
Singleton
,
A. T.
, and
C. J. C.
Reason
,
2007a
:
A numerical model study of an intense cutoff low pressure system over South Africa
.
Mon. Wea. Rev.
,
135
,
1128
1150
, doi:.
Singleton
,
A. T.
, and
C. J. C.
Reason
,
2007b
:
Variability in the characteristics of cut-off low pressure systems over subtropical southern Africa
.
Int. J. Climatol.
,
27
,
295
310
, doi:.
Tyson
,
P. D.
,
T. G. J.
Dyer
, and
M. N.
Mametse
,
1975
:
Secular changes in South African rainfall: 1880 to 1972
.
Quart. J. Roy. Meteor. Soc.
,
101
,
817
833
, doi:.
Usman
,
M. T.
, and
C. J. C.
Reason
,
2004
:
Dry spell frequencies and their variability over southern Africa
.
Climate Res.
,
26
,
199
211
, doi:.
Weldon
,
D.
, and
C. J. C.
Reason
,
2014
:
Variability of rainfall characteristics over the south coast region of South Africa
.
Theor. Appl. Climatol.
,
115
,
177
185
, doi:.
Wilks
,
D. S.
,
1995
: Statistical Methods in the Atmospheric Sciences. Academic Press, 467 pp.