Figure 1
a) rainfall totals from January to March
in mm (NCEP Reanalyses)
b) contribution of the JFM season to annual totals, in %
Rainfall is a major climatic element in tropical regions, and consequently, a significant socio-economic factor in the developing countries, whose economy is mainly based on a rain-dependent agriculture. In the regions where dry and wet seasons alternate, either with a bimodal rhythm with two dry seasons and two wet seasons, or with a monomodal rhythm with a dry season generally much longer than the wet season, the amount and distribution of rainfall during the wet season determine crops, and therefore the economy, and in some cases, survival of the people. Several droughts in the 1980s (i.e. Ethiopia in 1984) and 1990s, and moreover, their persistence over multi-annual or even decadal periods (see Sahel case in S. Janicot's article) drew the attention of the general public to the fragility and climatic dependence of the farming systems, and, more widely, to the socio-economic systems of the countries concerned. The issue of African rainfall variability south of the Equator drew much less attention, at least here in France. No significant trend towards a decrease in rainfall has been observed. However, inter-annual climate variability is actually marked and there were major droughts in the 1980s and 1990s (Harsch 1992) which induced a huge decrease in farming production (Vogel 1994). The Southern African population depends basically on water resources as regards food production, economic development and protection of the natural systems. In countries such as South Africa, the implementation of many waterworks enables a relative balance to be reached between the increasing urban, industrial and agricultural needs. Nevertheless, this balance is fragile and disturbed by the recurrent droughts affecting the country (1982-83 and 1992-93 for instance).
Tropical rainfall usually occurs within the Inter-Tropical Convergence Area (see S. Janicot's article). The latitude of this area shifts and follows, with some weeks of delay, the zenithal position of the sun. Schematically, it is in the northern hemisphere during northern summer, and in the southern hemisphere during southern summer. Its average position during the year is north of the Equator because of the unequal distribution of continents and oceans between the two hemispheres. Southern African rainfall is linked to this seasonal shift and mainly occurs between October and March. Only the Cape Province, in South Africa, and some coastal regions have a different seasonal rhythm. Figure 1a shows the distribution of the average total rainfall during the January-March term, which is when rainfall is the heaviest (figure 1b).
In comparison with West Africa, a different organisation of rainfall is noticeable. There is no structure in zonal strips. The Kalahari is only a semi-desert. Its small area implies it cannot be compared with the Sahara. Minimum rainfall occurs along the western coast of the sub-continent. It corresponds to the Namib Desert, due to the presence of coastal upwellings. Mainly located along the Namibian coasts, the rising of cold and deep waters influences the temperature of the lower layers of the atmosphere, their water vapour content and consequently the stability and possibility of convection and rainfall. Maximum rainfall is located at 15°S in Angola, Zambia and Northern Mozambique. It corresponds to the average position of the Inter-Tropical Convergence Area during this season. A secondary area of flux confluence (inter-oceanic confluence) develops parallel to the western coast of the sub-continent. This area is on an average located around 20°E. A secondary maximum rainfall can be observed south-east of the sub-continent, on the eastern side of the Drakensberg. This distribution of rainfall and main flux convergence/confluence areas is linked to the geometry of the southern part of the African continent. In particular, the meridian coastal orientation entails that the high meridian gradient of energy in the lower layers, which is the origin of monsoon circulation, does not operate in Southern Africa. The circulation of the south-east trade wind enables the advection of humidity from the Indian Ocean, and destabilises the atmosphere over the continent. However, rainfall is influenced by the horizontal temperature gradients between the continent, with its low thermic capacity, and the ocean, able to absorb more radiation. This is why anomalies in ocean surface temperatures play a major part in rainfall variability.
Many works have dealt with the relationship between rainfall in the various regions of Southern Africa and ocean surface temperatures. These temperatures are clues as regards forecasting because of their high thermic inertia. The most direct relationships have been found in regional ocean surface temperatures, near the sub-continent. They influence thermic gradients and the humidity of the lower layers of the atmosphere. This is why warm anomalies developing in the equatorial central Indian Ocean are generally associated with droughts (Mason 1995, Jury 1995). At a regional scale, cold anomalies observed in the south-west of the Indian Ocean (Mozambique Canal and area of re-circulation of the Aiguilles Stream) are also associated with rainfall deficits in a major part of South Africa (Reason 1999). Lastly, negative anomalies in the ocean surface temperatures in the Benguela system are linked to deficits in Angola (Hirst and Hastenrath 1983, Nicholson and Entekhabi 1987).
The relationship between Southern African rainfall and the El Niño phenomenon has been noticed for a long time. Ropelewski and Halpert (1987, 1989, 1996) have shown a positive correlation between the Southern Oscillation Index (SOI) and Southern African rainfall. This link is more easily explained for Southern Africa than for the Sahel because the wet season corresponds to the mature phase of the ENSO, when the range of anomalies in ocean surface temperatures and in atmospheric parameters is the largest. Moreover, anomalies in ocean surface temperatures in the Indian Ocean, linked with events in the Pacific, may relay with each other.
El Niño impacts are at a maximum in the south-east of the continent (Matarira 1990, Shinoda and Kawamura 1996, Rocha and Simmonds 1997), in the height of the southern summer, in January-February-March (Lindesay 1988, Lindesay and Vogel 1990). El Niño occurrences (warm anomalies in the Eastern Pacific) are usually associated with droughts in a large part of the sub-continent. However, El Niño events are not always accompanied by rainfall deficits (Mason and Nimmack 1992). There are usually few correlations between the rainfall index in Southern Africa and the several indexes reflecting the state of the Pacific. The relationship between Southern African rainfall and El Niño/La Niña events seems to be unstable in time. Regarding South Africa, authors mention a modulation of the relationship through the stratospheric QBO1 (Mason and Tyson 1992, Mason and Lindesay 1993, Jury et al. 1994) or a "mask" due to the low frequency variability (quasi-periodic oscillation of 18-20 years) of rainfall (Kruger 1999).
Most of these studies are concerned with administrative areas (most of them with the Republic of South Africa). The lack of obvious trend, as in the Sahel, and a more complex rainfall distribution entail that a "natural" coherent area can be defined from the criterion of inter-annual rainfall variability. Rainfall being heavier at the end of the summer (JFM) according to the annual totals, and the impact from the Pacific events being more significant within this period (Lindesay 1988), the rest of the present work focuses on this season, which is the most influenced by tropical circulation (D'Abreton and Lindesay 1993).
A way of determining a regional rainfall index in order to study inter-annual variability is to analyse the main components2 of monthly or seasonal totals of rainfall after subtracting the seasonal cycle. The first specific mode of rainfall variability in JFM obtained with this method is presented in figure 2a. The area outlined in black is defined as having a common inter-annual variability. This variability represents 31.9% of the total variance of the field over the calculation period (1946-88). This area may be considered as homogeneous as regards inter-annual rainfall variability. The spatial mean of the associated rainfall constitutes the Southern African regional rainfall index.
Figure 2b shows the temporal evolution of this index (dashed area) under the form of reduced anomalies (related to the standard deviation of the series) and the red line, the smoothed evolution of this index (overlapping means over 20 years). The very small low-frequency variability of the series and the lack of downward trend as in the Sahel are worth noting. The index essentially shows an inter-annual variability, with a variable range however. This range is high during the years 1910-1920, the variability is then lower in the years 1930-1950 and the range is highest during the years 1960-1980. The pink area shows the The variation in the range of variability of ocean surface temperatures in the eastern Pacific, or of the Southern Oscillation Index, has been noticed in many works, likewise the frequency and unfolding of the events themselves (for instance, refer to Graham 1994, Rasmusson et al. 1994, Clarke and Li 1995, Gu and Philander 1995, Wang 1995, Wang and Wang 1996, Trenberth and Hoar 1996, Zhang et al.1997). Part of this variation is shown in the low frequency mode presented in S. Janicot's figure 4b, and expresses the recent modifications of ocean surface temperatures in the tropical Pacific and in the extra-tropical latitudes and the Indian Ocean as well. Wang and Ropelewski 1995 have also demonstrated the association between the variations in the range of El Niño/La Niña events and a slow evolution, screened at 30 years, of global ocean surface temperatures interpreted as a basic state to which are superimposed specific El Niño/La Niña events. A warmer basic state is associated with a higher amplitude of the Pacific events and vice-versa.
A deeper inspection of the relationship between the Southern African rainfall index and the Southern Oscillation Index over the more recent period 1950-1960, which is better documented (fig. 2c), shows that, as is the case for the Sahel, El Niño events are associated with droughts. This becomes obvious from the end of the 1960s. Figure 2d, which presents correlations on overlapping periods, confirms this impression: correlations which were insignificant increase and become significants as soon as the events at the beginning of the 1970s are taken into account. This result is similar to the evolution of the relationships between Sahelian rainfall and El Niño shown by S. Janicot.
The composite analysis of the anomalies in ocean surface temperatures conditioned by dry years before and after 1970 show important differences. Before the end of the 1960s, droughts are associated with negative anomalies in ocean surface temperatures of a regional size (figure 3a). Negative anomalies are observed in the subtropical South Atlantic and Indian Ocean, according to the works already quoted. On the other hand, the droughts which have occurred since the end of the 1960s are associated with warmer temperatures in the eastern and central tropical Pacific, in the tropical Indian Ocean and in the equatorial Atlantic (fig. 3b). The spatial structure of these anomalies is very much like the first high frequency mode of the variability of global ocean surface temperatures (fig. 4a of the previous article) and may be associated with El Niño/La Niña events. This likeness underlines the fact that anomalies in Southern African rainfall have an inter-annual feature (high frequency) without trend.
The part played by the global context of ocean surface temperatures has been studied in numerical simulations (Richard et al. 2000). During El Niño events, in the global context of the years 1950-1960, numerical simulations have correctly reproduced the observed rainfall patterns: there is a lack of generalised deficits on the area studied. Similarly, during El Niño events, in the global context of the years 1970-1990, and according to the observations, experiments on sensitivity have simulated more important deficits, which affect the whole continent. Warmer waters in the Indian Ocean seem to create conditions favouring the impact of El Niño events on Southern Africa.
The results of studies on observation data show the importance of ocean surface temperatures on the amount of rainfall in Africa south of the Equator. A modification in the structure of anomalies in ocean surface temperatures associated with droughts has been observed since the end of the 1960s. The anomalies, which were usually of a regional size and located in the Indian and Atlantic oceans, are substituted by anomalies in the ocean surface temperatures in the Pacific, associated with El Niño/La Niña events. A change in the general context of ocean surface temperatures seems to be involved in this modification of relationship. However, no decadal variation in the average amount of rainfall has occurred, unlike in the Sahel.
Bigot S., Camberlin P., Moron V., Richard Y. 1997, "Structures spatiales de la variabilité des précipitations en Afrique : une transition climatique à la fin des années 1960?",
C.R. Acad; Sci. Paris, 324, série Iia, 181-188
Clarke A.J. Li B. 1995, " On the Timing of Warm and Cold El Niño-Southern Oscillation Events ", J. Climate, 8, 2571-2775
D'Abreton P.C., Lindesay J.A. 1993 "Water vapour transport over Southern Africa during wet and dry early and late summer months", Int. J. of Climatol., 13, 151-170
Graham N.E. 1994, "Decadal-scale climate variability in the tropical North Pacific during 1970s and 1980s : observations and model results ", Clim. Dyn., 10, 135-162
Gu D., Philander S.G.H. 1995, "Secular Changes of Annual and Interannual Variability in the Tropics during the Past Century",
J. Climate, 8, 864-876
Harsch E. 1992, " Drought devastates Southern Africa", Drought Network News, 2, 17-19
Hirst A.C., Hastenrath S. 1983, "Atmosphere-ocean Mechanisms of climate Anomalies in the Angola-Tropical Atlantic Sector",
J. Phys. Ocean., 13, 1146-1157
Jury M.R. 1995 "A review of research on ocean-atmosphere interactions and South African climate variability. ", South African J. of Sciences, 91, 289-294
Jury M.R., Mac Queen C., Levey K. 1994, "SOI and QBO Signals in the African Region", Theor. Appl. Climatol., 50, 103-115
Kruger A.C. 1999, " The influence of the decadal-scale variability of summer rainfall on the impact of El-Niño and La Niña events in South Africa", Int. J. Climatol., 19, 59-68
Lindesay J.A. 1988, "South African rainfall, the Southern Oscillation and a southern hemisphere semi-annual cycle", J. Climatol., 8, 17-30
Lindesay J.A., Vogel C.H. 1990, "Historical Evidence for Southern Oscillation - Southern African Rainfall Relationships.", Inter. J. Climatol., 10, 679-689
Mason S.J. 1995, "Sea-Surface Temperature - South African Rainfall Associations, 1910-1989", Inter J Climatol, 15, 119-135
Mason S.J., Lindesay J.A. 1993, "A Note on the Modulation of Southern Oscillation - Southern African Rainfall Associations with the Quasi-Biennial Oscillation", J. Geophys. Res., 98, 8847-8850
Mason S.J., Nimmack G.M. 1992, "The use of bootstrap correlation coefficients for the correlation coefficient in climatology", Theor. Appl. Climatol., 45, 229-233.
Mason S.J., Tyson P.D. 1992, " The modulation of sea surface temperature and rainfall associations over southern Africa with solar activity and the Quasi-Biennial Oscillation",
J. Geophys. Res., 97: 5487-5856
Matarira C.H. 1990, "Drought over Zimbabwe in a Regional and Global Context", Int. J. Climatol., 10, 609-625
Nicholson S., Entekhabi D. 1987, " Rainfall Variability in Equatorial and Southern Africa : Relationships with Sea Surface Temperatures along the Southwestern Coast of Africa", J. Clim. Appl. Meteorology, 26, 561-578
Rasmusson E.M., Wang X.L., Ropelewski C.F. 1994, "Secular Variability of the ENSO Cycle", in Decade to Century Time Scales of Natural Climate Variability, Academic Press
Reason C.J.C. 1999 "Warm and cold events in the southeast Atlantic/southwest Indian Ocean region and potential impacts on circulation and rainfall over southern Africa", Meteorology and Atmospheric Physics, 69, 49-65
Richard Y., Trzaska S., Roucou P., Rouault M. 2000, "Modification of the Southern African Rainfall variability/ ENSO Relationship since the late 1960's", Clim. Dyn., in press
Rocha A., Simmonds I. 1997, "Interannualvariability of southestern African summer rainfall. Part1: relationships with air-sea interaction processes. ", Inter. J. Climatol., 17, 235-265
Ropelewski C.F., Halpert M.S. 1987, "Global and regionAL SCALe precipITATIon and temperature patterns associated with El Niño/Southern Oscillation", Mon. Wea. Rev., 115, 1606-1626
Ropelewski C.F., Halpert M.S. 1989, "PrecipiTation patTERNS AssociateD WITH the high indices phase of the Southern Oscillation", J. Climate, 2, 268-284
Ropelewski C.F., Halpert M.S. 1996, "QuantifYing SouthERN OSCillation -PREcipitation relationships", J. Climate, 9,1043-1059
Shinoda M., Kawamura R. 1996, "Relationships between rainfall over semi-arid Southern Africa and geopotential heights and sea surface temperatures", J. Meteorol. Soc. Japan, 74, 21-36
Trenberth K.E., Hoar T.J. 1997, "El Niño and climate change", Geophys. Res. Lett., 24, 3057-3060
Vogel C.H. 1994 "(Mis)management of Droughts in South Africa : past, present and future", S. Afr. J. Sci., 90, 4-5
Wang B. 1995, "Interdecadal Changes in El Niño Onset in the Last Four Decades", J. Climate, 8, 267-285
Wang X.L., Ropelewski C.F. 1995, "An assessment of ENSO-Scale secular variability.", J. Climate, 8, 1594-1599
Wang B., Wang Y. 1996, "Temporal Structure of the Southern Oscillation as Revealed by a Waveform and a Wavelet Transform ", J. Climate, 9, 1586-1598
Zhang Y., Wallace J.M., Battisti D.S. 1997, "ENSO-Like Interdecadal Variability: 1900-93 ", J. Climate, 10, 1004-1020