Image: Satellite view of Africa
By closely monitoring the vegetation in the region affected by increased rainfall using weather satellites, scientists can identify the actual areas affected by outbreaks of Rift Valley Fever in East Africa. Scientists use satellite images to show regions of Africa that are greener (and wetter) than normal or more brown (and drier) than normal.
updated 3/23/2004 8:34:20 PM ET 2004-03-24T01:34:20

Despite the advances of modern medicine, diseases like malaria, dengue fever and even plague still afflict millions of people each year, crippling some while proving fatal for others. Many of the diseases are spread through infected, bloodsucking mosquitoes, which can cause widespread epidemics by feeding on people or animals, then flying to another target.

But scientists are hoping to stem future outbreaks — or at least reduce their severity — using not just medicine, but also Earth-watching satellites capable of identifying potential disease hotspots before an outbreak has time to spread.

The method combines computer modeling of satellite data with good old-fashioned fieldwork. It could give local health officials more lead time to distribute preventive medical care before a disease strikes.

"We actually have all the pieces to do this," Ronald Welch, of NASA's Global Hydrology and Climate Center, said in a telephone interview. "The technology is now available and people have done enough research. It's just time to integrate all this."

Welch is leading an effort to develop an early warning system he hopes will accurately predict the progression of malaria across the Mewat region of India, a predominantly rural area south of New Delhi. If his model is successful, it could be used for any number of diseases that can be spread through mosquitoes.

According to World Health Organization statistics, malaria alone infects up to 500 million people each year, killing at least a million.

The spread of other mosquito-borne sicknesses, such as dengue fever, West Nile virus and hantavirus can also lead to outbreaks across small villages, towns and whole regions.

Welch's research team has already conducted ground and space studies in parts of Guatemala, using high-resolution satellite images to find tiny pools of standing water where mosquitoes and malaria can grow. The task now is to build a model that not only accurately shows where the disease could develop, but effectively predicts an outbreak.

"But it's one thing to build a model, it's something else entirely to validate it," Welch said.

Peering down for impending disease
Welch's team uses data from satellites like Landsat 7 and NASA's Terra orbiter to build a climate profile of a region's rainfall, temperature variations, local vegetation and soil moisture. That profile is then combined with high-resolution imagery from commercial satellites Ikonos and QuikBird, which have enough resolution to spot an object just 31 inches (80 centimeters) in size, from orbit, to determine where mosquito spawning areas are likely to appear.

To spread malaria, mosquitoes need pools of stagnant water where they can lay their eggs. In Mewat, these appear as puddles after heavy rains, as well as in swamplands and rain-filled buckets typically left outside. The temperature must be warmer than 64 degrees Fahrenheit (18 degrees Celsius), with humidity levels hovering between 55 percent and 75 percent, for the pests to survive. All of those factors can be gleaned from weather satellite data.

But gathering the satellite data is the easy part.

"What you're tracking [with satellites] is not really a disease," said Assaf Anyamba, a research scientist with NASA's Goddard Space Flight Center. "You're tracking the conditions that lead to the disease, and it's putting that whole chain of events together that is challenging."

In a separate study, independent of Welch's, Anyamba and his colleagues used weather satellites watching Earth's oceans to associate outbreaks of Rift Valley fever, which like malaria is spread by mosquitoes, with changes in sea surface temperatures. The study helped scientists develop what they believe is a five-month warning system for the disease.

But at the heart of any outbreak prediction model is the biology of a disease and its relation to the ecosystem in which it lives, Anyamba told To get that, he added, requires a surveillance team on the ground.

Fastidious fieldwork
For Welch's study, ground researchers need to know such basic facts as how the local people sleep at night, whether they protect themselves with insect nets and how often they leave buckets of water out where mosquitoes can breed and spread disease.

That called for 10 scientists, in teams of two each, to canvass Mewat, a region about two-thirds the size of Rhode Island with 700,000 residents across nearly 500 settlements. It will be up to those ground researchers to determine what local species of mosquito is present, how it behaves and where it tends to bite people (indoors or outdoors). The locations of homes and cattle pens must be inserted into the computer model because mosquitoes tend to seek out livestock before dining on human blood.

"Getting this ground data, this demographic data, is the hard part," Welch said, adding that it is also important to learn which forms of malaria (there are two, though only one is fatal) or other pathogens are present.

India's health officials have spent a long time studying their regional malaria problem, and Welch's team has been working with the country's Malaria Research Center to install ground instruments to measure local climate and soil over the last year of so.

"Right now we have people in the area to collect data from those instruments," Welch said. "We're hoping to have a preliminary prediction model up and ready by the end of the year."

Follow-up studies
With the preliminary model in hand, Welch and his colleagues hope to check its predictions with follow-up studies to see if malaria appears as predicted. Epidemics of the disease tend to run in cycles, with the next one likely in the next two years, he said.

In addition to curbing the human cost of disease outbreaks, preventing the progression of sickness can also affect a region's economy as well.

For example, an epidemic of Rift Valley Fever in East Africa during the late 1990s infected cows throughout the region, prompting a ban on the region's cow trade for beef and milk markets. In addition to mosquito bites, the disease can be passed to humans through milk and meat consumption.

"It had very devastating impacts on the local economy in East Africa, especially since the only thing people there do is grow these animals," Anyamba said. Knowing ahead of time that the disease could appear soon could have alerted local East African governments or cattle farmers to slaughter or export their livestock early in preparation, he added.

Disease alerts
The idea of an early warning system to predict disease outbreaks is not a new one. Since the 1960s, researchers have discussed how to use what's called "landscape epidemiology" to predict the spread of a disease based on its surrounding geography and climate. Landscape epidemiology also takes into account to local cultural practices of people living in a given geographical area prone to disease outbreaks.

But it was only recently that the daily Earth examinations conducted by satellites in geosynchronous orbits — hovering above a fixed surface location by orbiting at the same pace as the planet rotates — provided epidemiologists with enough long-term data to begin making associations between climate and disease.

"For a long time there was no good systematic collection of information that would let us establish that ecosystem-disease relationship," Anyamba said. "But now we have a coherent picture from which you can now begin to see how that relationship forms."

© 2013 All rights reserved. More from


Discussion comments


Most active discussions

  1. votes comments
  2. votes comments
  3. votes comments
  4. votes comments