For communities across southern Africa, the summer rainy season can bring both essential water and destructive extremes. Rain supports agriculture, replenishes rivers and reservoirs, and sustains ecosystems. When intense systems stall or deliver exceptional downpours, however, water that would otherwise be welcome can overwhelm drainage networks and river catchments.
New climate research suggests that one atmospheric pattern associated with unusually wet southern African summers could become more frequent as the planet warms. Under a high-greenhouse-gas-emissions scenario, models project that extreme southward swings of the South Indian convergence zone could occur approximately once every 10 years during the twenty-first century, compared with once every 22 years in the twentieth century.
Rainfall intensity associated with those extreme swings is also projected to increase by an average of 20%. The findings, published on 20 October 2026 in Nature Climate Change, point to a possible increase in flood-related hazards in parts of the region. They do not, however, mean that every southern African town will flood twice as often, or that annual rainfall everywhere will rise by 20%.
The study is particularly relevant because its researchers examined the physical reasons these exceptional rain-band movements may become more common. Rather than simply assuming that a warmer atmosphere produces heavier rain, they traced changes in winds, the temperature contrast between Africa and the Indian Ocean, and circulation over the Southern Ocean and Atlantic.
The moving rain band that helps shape southern African summers
The South Indian convergence zone, often abbreviated SICZ, is a broad atmospheric system where moist air and circulating winds come together. It is an important part of summer rainfall variability across portions of southern Africa and the neighbouring southwest Indian Ocean.
Unlike a stationary feature on a weather map, the convergence zone changes position. During some seasons it extends or moves farther south than usual, shifting the geography of rain-producing clouds and moisture transport. In other seasons its position and strength differ, contributing to markedly different conditions across the region.
The researchers concentrated on extreme southward swings, not the everyday movement of thunderstorms and not every period of heavy rain. These swings are unusual departures from the system’s typical position and can produce conspicuously wet conditions over some southern African areas.
That distinction matters because annual rainfall averages can hide important risks. A region could experience less rain during many seasons and still face occasional exceptionally intense wet episodes. Changes in the frequency of a particular extreme circulation pattern do not necessarily move in the same direction as changes in average rainfall.
For planners, this makes the study a useful complement to broad climate statements about a warmer and potentially drier southern Africa. It asks how the distribution of rare wet conditions might change, even where mean climate trends point toward drying or shifts in the position of rain belts.
How the study reconstructed past extreme swings
Haodong Wang, Tao Geng, Wenju Cai and Xiaopei Lin combined analyses of observed climate characteristics with computer simulations. To characterize historical SICZ movement, they examined austral-summer rainfall anomalies across a region extending roughly from 30°S to 5°S and 20°E to 45°E.
The observational analysis covered 1960 to 2024 and focused on the Southern Hemisphere summer months of December, January and February. The team used a statistical approach to identify shifts in the rain band’s position from patterns in precipitation anomalies rather than treating one measuring station as representative of the whole region.
In their historical analysis, the authors identified four observed extreme southward-swing seasons: 1998/1999, 1999/2000, 2016/2017 and 2020/2021. They also distinguished seven moderate events in the observed record.
Those seasons illustrate what the researchers meant by an extreme circulation configuration. They are not a complete list of every major flood that affected southern Africa, and they should not be interpreted as a claim that all floods in those summers resulted from the same mechanism.
The team also investigated how wind and humidity anomalies contributed to shifting rainfall. Their analysis indicated that the motion of moist air through anomalous winds played a central role in the observed atmospheric patterns. That supplied a physical benchmark against which to judge the behaviour of global climate models.
Historical observations alone cannot tell researchers how often a rare event will occur in a future climate. They can, however, help identify whether a model represents the system in a reasonably credible way before its long-term projections are analysed.
Why 20 climate models were selected from 43
The researchers examined 43 models from the sixth phase of the Coupled Model Intercomparison Project, known as CMIP6. These internationally developed models simulate interactions among the atmosphere, oceans, land and other parts of the climate system.
Not every model reproduced the observed SICZ equally well. The authors therefore compared simulated summer rainfall patterns and the relationship between their SICZ movement index and the actual latitude of maximum rainfall.
On that basis, they selected 20 models that represented the rain band’s mean conditions and north-south swings more credibly. The remaining 23 models formed an important comparison group for checking whether the results depended heavily on model selection.
This screening is a methodological strength because an average across many models can conceal systematic problems. If a model poorly represents the present-day mechanism being investigated, its projections for that same mechanism deserve more caution.
The future comparison contrasted conditions simulated for the twentieth century, 1900–1999, with those for the twenty-first century, 2000–2099. The headline figures came from a high-emissions pathway called SSP5-8.5, which should be understood as a scenario for exploring possible climate outcomes rather than a forecast of humanity’s inevitable future emissions.
The researchers additionally tested sensitivity to model selection and other emissions pathways. The broader tendency toward more frequent southward swings was not confined to a single set of models, although the precise headline frequency and intensity figures are tied to the reported high-emissions comparison.
From roughly every 22 years to every 10 years
The most striking result concerns how often the SICZ moves exceptionally far south. Under the high-emissions scenario, the frequency of these extreme shifts rises from an average of one event per 22 years during the twentieth century to approximately one per 10 years during the twenty-first.
Expressed as event rates, that represents a change from about 4.5 events per century to about 10 events per century. The ratio is a little above two, which is why describing the frequency as roughly doubling is appropriate.
These figures are long-term statistical averages across climate simulations. They are not a schedule. They do not mean that the next event is due exactly 10 years after the previous one, or that every 10-year period must contain one event.
Changes in the atmospheric conditions underlying rare events could produce clusters, long gaps and substantial differences between locations and seasons. The rate simply describes how often a defined extreme pattern arises in the simulated climates on average.
The study also projects an average 20% increase in rainfall intensity during these extreme southward swings. This should not be confused with a 20% increase in normal summer rainfall, rainfall at every location or the volume of every future flood.
Where rainfall occurs, how long an event lasts and the condition of the land and rivers will remain crucial to the consequences of any particular episode. A more intense rainfall-generating circulation pattern raises concern, but cannot be translated directly into a universal percentage increase in flood damage.
What changes in a warmer atmosphere?
The first proposed mechanism involves the contrast in warming between the African continent and the Indian Ocean. Land tends to respond differently to greenhouse warming from large bodies of water. In the simulations, the African continent warms more rapidly, altering the distribution of atmospheric pressure.
That stronger land-ocean contrast favours changes in winds that move air southward. Specifically, the models indicate an intensification of northerly winds that supports unusual southward excursions of the convergence zone.
Those winds matter because rainfall patterns depend on the transport and convergence of moisture, not simply on how much water vapour a warmer atmosphere could theoretically contain. When circulation draws moist air into favourable locations, the setting for substantial rainfall becomes more likely.
A second mechanism occurs far to the south. The researchers link the poleward movement of Southern Ocean westerly winds to changing sea-surface temperature patterns in the subtropical Indian Ocean, with stronger relative warming toward the south.
That ocean temperature arrangement favours the circulation changes associated with southward SICZ swings. The moving westerlies also help establish circulation anomalies over the southeastern Atlantic that can increase moisture transport from the Atlantic toward Africa.
The mechanisms operate through different components of the climate system, but reinforce conditions favourable to an extreme southward position of the rain band. This is why the researchers describe a change in atmospheric dynamics, not merely an increase in moisture availability.
The importance of physical mechanisms is especially clear in regional climate research. A projection supported by a coherent chain of changes in temperature, winds and moisture can be examined in more detail than a simple statement that future rainfall might be more variable.
More rare wet episodes do not rule out regional drying
One potentially confusing implication is that an increase in extreme southward SICZ swings can coexist with a change in the average position of the convergence zone that favours a northward shift.
The authors explicitly examined this issue. Their additional analyses indicated that more frequent extreme southward departures can occur even when the mean-state behaviour changes differently. The probability of unusual excursions and the location of the average are two separate features of a variable climate system.
This distinction helps reconcile the study with broader research into a hotter and, in some respects, drier southern African climate. Changes in average moisture conditions and changes in the occurrence of exceptional rainfall events should be assessed together rather than treated as opposing claims.
In practice, the same region may have to prepare for greater heat or water stress over some periods and disruptive precipitation over others. The existence of one risk does not eliminate the other.
A related Research Today article on climate models and future cooling demand illustrates why model choices can have major implications for planning. The new SICZ paper adds a different consideration: whether models represent the specific atmospheric processes responsible for the most damaging rainfall episodes.
What the results mean for communities and infrastructure
Southern Africa includes cities, agricultural areas and river catchments with sharply different exposure to extreme rainfall. More frequent wet extremes could have implications for stormwater systems, transport infrastructure, settlements on floodplains, agricultural timing and emergency preparedness.
But the modelling result does not give a precise flood probability for Johannesburg, Durban, Maputo or any other specific location. Determining local impacts would require detailed information about topography, soils, river networks, drainage, settlement patterns and how the rainfall in the models translates into local storm conditions.
Flooding can become more severe where heavy rain coincides with saturated soil, high river levels, blocked drains or vulnerable housing. Equally, improvements in early warning systems, infrastructure and land-use planning can reduce harm even if a climate hazard becomes more frequent.
Previously published Research Today coverage of flood-evacuation simulations examined the operational side of such planning. Those simulations concern a different study and setting, but help illustrate why translating climate hazards into safe routes and response plans is a separate, practical task.
Research on weather forecasts and southern African pastoral communities also highlights that receiving climate information does not automatically remove social and economic vulnerability. Preparedness depends on whether warnings can be interpreted and acted upon.
The most useful implication of this new research is therefore not a single predicted flood date. It is a reason to consider the possibility of more frequent extreme rain-band conditions when updating regional risk assessments and investigating where local adaptation is needed.
How much confidence can be placed in the forecast?
The study provides a substantial modelling analysis, but all long-range climate projections involve uncertainty. The headline rates depend on the definition of an extreme SICZ swing, the simulated climate system and the emissions pathway used for the main comparison.
The observed extreme events are relatively rare. Four identified extreme seasons within the historical observational analysis provide a useful physical reference, but they are not enough on their own to establish an exact future frequency. This is one reason the researchers turned to ensembles of climate-model simulations.
Model selection introduces a further judgment. The authors’ choice to prioritize 20 better-performing CMIP6 models has a physical justification, but a different weighting or future generation of models could change the magnitude of the estimated shift.
The paper includes robustness checks involving the wider model collection, different emissions scenarios and alternative ways of accounting for changes in the mean position of the rain band. These tests strengthen the case that an increase in southward swings is a meaningful signal within the model framework, without removing all uncertainty about its size.
It is also important not to interpret the results as an attribution study of a specific recent flood. The authors investigated a regional circulation pattern and its future response to greenhouse warming. A direct claim about what caused any named disaster would require event-specific analysis.
Finally, atmospheric extremes and disaster losses are different outcomes. A 20% increase in the intensity of rainfall during a defined event does not establish a 20% increase in property losses or affected populations. Changes in exposure and vulnerability can make local consequences either greater or smaller than the change in the physical hazard.
Why the findings matter beyond a rainfall statistic
Climate-change discussions often focus on temperature averages, annual rainfall totals or the number of very hot days. These are useful indicators, but communities experience much of their risk through unusual combinations of atmospheric events and local conditions.
The new analysis draws attention to a less familiar mechanism: a rain band that may be more likely to swing unusually far south as the atmosphere and surrounding oceans warm at different rates.
For Southern Africa, the projected change is consequential enough to warrant further investigation. Under the high-emissions scenario, the simulations suggest a shift from approximately four or five extreme southward events per century to around ten, with rainfall intensity during those events averaging one-fifth higher.
That is not a prediction that every summer will be wetter or every community will flood more often. It is evidence that an important source of regional rainfall variability may behave differently in a warmer world.
The next step is to combine this large-scale atmospheric evidence with regional rainfall models, local hydrology and assessments of where people and infrastructure are most exposed. In that form, research on a moving convergence zone could help inform practical adaptation rather than remain simply another global climate projection.
Looking for studies on this topic? Use the Research Today Research Paper Finder to search scholarly research by topic, question, author or keyword.
Source Information
Study title: Increased southward swings of the South Indian convergence zone under greenhouse warming
Authors: Haodong Wang, Tao Geng, Wenju Cai and Xiaopei Lin
Journal: Nature Climate Change
Publication date: 20 October 2026
Study design: Observational analysis of austral-summer regional rainfall and atmospheric circulation, combined with historical and future CMIP6 multi-model climate simulations and robustness checks across model selections and emissions scenarios
Sample or scope: Observed summer rainfall anomalies over southern Africa during 1960–2024; 43 CMIP6 models assessed and 20 selected for the primary simulation analysis comparing 1900–1999 with 2000–2099, with the headline projection under SSP5-8.5
Key finding: Under a high-emissions scenario, projected extreme southward SICZ swings become more frequent, from approximately one per 22 years to one per 10 years, with an average 20% increase in rainfall intensity during such events
