James Hurrell of the National Center for Atmospheric Research (NCAR) will present the findings on May 24 in New Orleans at the spring meeting of the American Geophysical Union. The study, conducted with Martin Hoerling (National Oceanic and Atmospheric Administration), was supported by NOAA and the National Science Foundation, NCAR's primary sponsor.
The analysis, which draws on 60 simulations of global climate from five computer models, provides new evidence linking drought in southern Africa to the warming of the Indian Ocean. However, it contradicts earlier studies that also connected the Sahelian drought of northern Africa to the Indian Ocean. Instead, the new results point to a late 20th-century cooling of the North Atlantic Ocean as having been key to Sahelian drought. A subsequent switch to North Atlantic warming, partly consistent with the impact of greenhouse gas increases, is the main factor behind the Sahel's recent swing from drought to moist conditions, the researchers believe.
"Changes in the Indian and Atlantic oceans are causing large regional effects in Africa, and these have substantial impacts on people. Now we can explain these climatic effects," says Hurrell.
Recurrent drought since the 1970s has plagued southern Africa, including Angola, Zambia, and Zimbabwe. Meanwhile, the nearby Indian Ocean has warmed more than 1 degree Celsius (0.6 degree Fahrenheit) since 1950. As showers and thunderstorms develop in the rising air above the warming ocean, says Hurrell, they help lead to sinking air and drought in a surrounding ring that includes southern Africa.
"In our models, the Indian Ocean shows very clear and dramatic warming into the future, which means more and more drought for southern Africa," says Hurrell. "It is consistent with what we would expect from an increase in greenhouse gases."
Hurrell and Hoerling compared model results from 1950-99 to several control runs that omitted the Indian Ocean warming. None of those runs showed the magnitude of drying that actually occurred in southern Africa. When the models did include the Indian Ocean warming, southern Africa consistently dried out. The models also project that by 2049, monsoons across southern Africa could be 10% to 20% drier than the 1950-99 average.
A different process appears to shape rainfall in the Sahel. When sea-surface temperatures are warmer in the South Atlantic than in the North, it pulls the Sahelian monsoon cycle south as well, depriving the region of its usual rains.
"This was the situation during much of the latter half of the 20th century." says Hurrell. "We believe the North Atlantic Ocean cooling was natural and masked an expected greenhouse-gas warming effect."
Since 1990, the sea-surface temperature pattern has reversed, warming more rapidly in the North Atlantic than in the South. The models examined by Hurrell and Hoerling show this trend intensifying in future decades. They project that the Sahel monsoon will be some 20% to 30% wetter by 2049 compared to the 1950-99 average.
The warming of Indian Ocean waters is well beyond the range expected from natural processes. This strengthens the case that greenhouse gases are involved, says Hurrell. In the Atlantic, natural variability affects ocean temperatures more strongly, making it more difficult to attribute changes there to greenhouse-gas effects.
Paleoclimate records show that even greater climate swings have occurred in Africa's monsoons, most likely related to past v ariations in solar output and in Earth's orbit. "From a paleoclimate perspective, the recent African dryings appear to be neither unusual nor extreme," says Hurrell.
Monsoon rains, critical to life in much of Africa, shift north and south with the seasons. They normally reach the Sahel from July to September and the southern part of the continent from February into April. Low-pressure centers moving west from the Sahel during the monsoon often serve as seed for tropical storms and hurricanes in the North Atlantic. Hurrell's work does not address the possible impact of increased rains in the Sahel on future Atlantic hurricane activity.
For their study, Hurrell and Hoerling examined output from computer models at NCAR, NASA, NOAA, the European Centre for Medium-Range Weather Forecasts, and France's National Center for Meteorological Research (CNRM).