Scientists say they have developed a computer model that can tell them whether a malaria season will be mild or severe five months before it occurs. Experts say the information provided by the model gives public health officials in countries with limited resources time to prepare for a severe malaria outbreak.
Controlling malaria is a top priority for international health officials. The mosquito-borne illness infects 500 million people worldwide each year, killing an estimated one million, most of them children. The problem is greatest in sub-Saharan Africa.
With advance warning of an epidemic, a country could stock up on medications and redirect more resources to combating a malaria outbreak in any given region.
Weather is one of the key factors in predicting outbreaks, and as any expert knows, predicting the weather can be a tricky business.
Simon Mason is a climatologist with the Earth Institute at Columbia University in New York.
"Most people joke about us not getting the weather forecast right over the next few days, so it's even harder for us to forecast the climates if we're going to forecast the climates three, four months in advance," he said.
Columbia's Earth Institute collaborated with European scientists to develop the DEMETER project.
DEMETER is a computer ensemble of all the available climate models around Europe. The models are then used to make weather predictions up to four months ahead of time in tropical countries, where malaria is endemic.
Each model within DEMETER predicts the effects of ocean warming, as occurs every three to four years with a phenomenon known as el Nino, sea surface temperatures, winds, rainfall and dry conditions on climate.
Tim Palmer led a team of scientists at the European Center for Medium Range Weather Forecasts in England, which devised DEMETER.
"When things are predictable, all of the models give the same answer," he noted. "They are consistent within each other. When we are entering a very unpredictable or chaotic period, then the models diverge in what they say. They give different answers and then we can't be sure of what's going to happen."
Scientists say the model was more successful at predicting very low epidemic years, which it did 85 percent of the time.
Because malaria epidemics usually follow wet rainy seasons, DEMETER is used to forecast the rainy season that precedes the malaria season, which peaks in March and April in Botswana, according to Mr. Palmer.
"So, if we can make a forecast of the rainfall amount ahead of the rainy season then we have a good estimate of the likelihood of a malaria epidemic or, indeed, a below average year months ahead of the outbreak of the disease," he added.
Researchers are already starting to use DEMETER to help countries in southern Africa predict whether they should brace for a malaria outbreak.
Earth Institute researcher Simian Mason says the 2006 malaria season should not be too bad.
"There was a risk of a little higher rainfall than normal," he explained. "However, we didn't feel this would translate into a big increase in the risk of malaria epidemics this year."
Researchers describe their method for forecasting malaria epidemics in the latest issue of the scientific journal Nature.