NEW YORK (AP) — A 10 percent chance of showers today, a 70 percent chance of flu next month.
That's the kind of forecasting that scientists are trying to invent, as they incorporate more and more weather data — rainfall, temperatures, other conditions — into their computer models for predicting disease outbreaks.
Health officials are excited by this sort of work, hoping it can be used to fine-tune vaccination campaigns and other disease-prevention efforts.
In one recent study, two scientists reported they were able to predict, more than seven weeks in advance, when flu season would peak in New York City.
On the other hand, skeptics say human behavior and other factors influence illness as much, or more, than weather does. They say weather-based disease predictions still have a long way to go.
The concept is not new. Scientists have been working on mathematical models to predict disease outbreaks for decades and have long included weather in their calculations. They have known, for example, that temperature and rainfall affect the population of mosquitoes that carry malaria, West Nile virus and other dangerous diseases.
What are new are improvements in weather tracking, including satellite technology, and more sophisticated computer processing. As a result, “in the last five years or so there's been quite an improvement and acceleration” in weather-focused disease modeling, said Ira Longini, a University of Florida biostatistician who has worked on such projects.
Some claim success.
In the United States, researchers at Johns Hopkins University and the University of New Mexico tried to predict outbreaks of hantavirus in the late 1990s. They used rain and snow data and other information to study patterns of plant growth that attract rodents. People catch the disease from the droppings of infected rodents.
“We predicted what would happen later that year,” said Gregory Glass, one of the Johns Hopkins researchers.
More recently, in east Africa, satellites have been used to predict rainfall by measuring sea temperatures and cloud density. That in turn has been used to generate “risk maps” for Rift Valley fever, a virus that can cause blindness or death. Researchers say they've been able to give two to six weeks' warning.
Last year, other researchers using satellite data in east Africa said they found that a small change in average temperature was a good warning sign that cholera cases would double within four months.
“We are getting very close to developing a viable forecasting system” against cholera, said Rita Reyburn of the International Vaccine Institute, based in South Korea. That, in turn, can help health officials in those countries ramp up emergency vaccination.
Some diseases are especially hard to forecast, such as West Nile virus. Last year was one of the worst for the U.S. since the mosquito-borne virus arrived in 1999 — about 2,600 serious cases and nearly 240 deaths.
Officials said the mild winter, early spring and very hot summer were partly to blame. But the danger wasn't spread uniformly. In Texas, the Dallas area was hard-hit, while other places with similar weather and mosquitoes got off easier.
“Why Dallas, and not areas with similar ecological conditions? We don't really know,” said Roger Nasci of the Centers for Disease Control and Prevention.
Some think flu should be easier to forecast. There's already some predictability to the annual winter flu season. But it's been tricky, too.
Researchers have tried using flu reports from doctors' offices, Google searches and Twitter tweets to beef up their predictions. But their work hasn't yielded much more warning than traditional methods.
In the New York study, the scientists claimed seven weeks of warning by using weather data, focusing on the fact that the flu virus spreads better when the air is dry and turns colder. They looked at humidity readings and blended that data with Google Flu Trends, a website that tracks how many people are searching each day for flu-related information, often because they're starting to feel sick.
Copyright 2013 Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.