Scientists are learning that the key to predicting certain epidemics - like Rift Valley fever in Africa or Hanta virus in the U.S. - lies in an unexpected place: the ocean.

by Karen Miller

It's called Rift Valley fever.

Humans can be infected as well, either through mosquitoes that carry the disease, or by handling infected tissue. Few die, but the illness can cause serious complications: meningoencephalitis, an inflammation of the brain, and lesions of the retina, which leave victims with at least some permanent loss of vision. During the most recent and devastating outbreak in 1997-98, an embargo banned exports of East African meat for one and a half years.

While no easy treatment exists for the disease, Rift Valley fever can be controlled. Animals can be vaccinated, insecticides can be spread into the soil to keep infected mosquitoes from hatching. But the disease's unpredictability has been a sticking point: without knowing when and where the disease will strike, it's hard to know how to use those controls efficiently.

more Rift Valley fever was first identified during the 1930's when it struck residents of Kenya's great Rift Valley, pictured here.

But the disease might not be so capricious after all, says Assaf Anyamba of NASA's Goddard Space Flight Centre.

Anyamba and colleagues at Goddard and at the Walter Reed Army Institute of Research have discovered that outbreaks of Rift Valley fever follow sudden floods triggered by El Niño and a similar (yet lesser-known) climate disturbance called the "Indian Ocean Dipole." Using weather satellites to track sea surface temperature patterns in the Indian and Pacific oceans, they now believe they have found a way to predict outbreaks up to five months in advance.

Sea surface temperatures can predict the likelihood of the disease because tiny variations in these temperatures cause huge shifts in air circulation patterns - shifts that alter rainfall around the globe. El Niño, for example, happens when a band of warmer-than-average water forms near the Pacific coast of South America. Meanwhile, Pacific waters near Australia and Indonesia become a bit cooler than usual.

A similar type of temperature imbalance can occur in the Indian Ocean, with the western part near Africa becoming warmer than the eastern part near Australia. Indeed, researchers liken this "Indian Ocean Dipole" to El Niño in the Pacific. Both tend to increase rainfall in East Africa.

When the two anomalies occur at the same time, buckets pour.

more This map shows the greening of East Africa during an Indian Ocean warm spell in 1983.

"The year 1997 saw the largest El Niño ever recorded simultaneous with a very large Indian Ocean Dipole (see figure)," says Christina Clark, an atmospheric scientist at the University of Colorado. "East African rainfall was then the highest on record, in many places five times the normal amount." Such floods bring Rift Valley fever because water collects in shallow depressions called "dambos" that punctuate the savannahs, providing mosquito eggs with exactly the nurturing conditions that they need to hatch.

The data about both sea surface temperature and vegetation is provided by NOAA's Advanced Very High-Resolution Radiometer (AVHRR) - a type of visible-light and infrared sensor carried on many polar orbiting weather satellites. "These satellites were not really designed to monitor land surface conditions," says Anyamba. "They were designed to monitor atmospheric conditions - basically, clouds." But Goddard scientist Compton Tucker realized, says Anyamba, that by manipulating the information provided by the AVHRR, he could produce a 'greenness index,' which measured the condition of the vegetation on the ground.

Using sea surface temperatures to predict when East Africa might be vulnerable, and using the greenness index to pinpoint exactly where, researchers can alert health officials to potential danger. "What we can do is provide public health officials with an efficient way of being able to focus their resources, rather than sending teams out all over the place," says Anyamba.

Deer mice and mosquitoes are disease vectors that can suddenly flourish in areas made wet by El Niño.

The method used to forecast Rift Valley fever could be expanded to predict other types of epidemics. It could be applied in areas that, like East Africa, are usually dry, but sometimes experience heavy rains, and that, like East Africa, have ecosystems that burgeon when precipitation comes.

Hanta virus outbreaks in the American southwest could be monitored in this way. The virus is carried by deer mice and can kill people who have been exposed to it. Like Rift Valley fever, says Anyamba, Hanta virus is correlated with rainfall. "The US southwest is really a very dry environment," he says, "and you are likely to see there the same kind of bioclimatic rhythms that you see in East Africa."

Right now, says Anyamba, "we're in operational mode [for East Africa]." Every month, he and his colleagues post their findings on the web, so that people in the field can "check the animals, check the people, see whether there's any activity."

Things are quiet - for now.

But sea temperatures will shift again. And when they do, the work of Anyamba and his colleagues will save lives