A woman blowing her nose.

Humidity and temperature both play a role in flu outbreaks.


‘Messy math’ from sardine studies could help fight flu outbreaks

It’s about that time, in the Northern Hemisphere at least: flu season. Each winter in the temperate parts of the world, the influenza virus seemingly awakens from its slumber, leaving fevers, sniffles, and aches in its wake. Now, using mathematical techniques created for studying fluctuating fish populations, a team of researchers has shown that both temperature and humidity influence the timing of flu outbreaks all over the world—a finding that could lead to new ways to fight the virus.

The relationship between climate and the flu looks simple in places like North America and Europe: Winter’s colder and drier weather means more outbreaks. But in the tropics, those cycles disappear, and things get decidedly harder to predict. Some researchers thought that meant climate just isn’t that important a factor in tropical flu outbreaks. But George Sugihara, a mathematical biologist at Scripps Institution of Oceanography in San Diego, California, suspected there might be a deeper rule connecting how the virus acts in the two regions.

The problem, according to Sugihara, is that past studies examining flu prevalence have failed to look at it as a whole natural system, instead trying to analyze the impact of individual pieces like temperature and humidity. Sugihara is a pioneer of mathematical methods that throw out the equations and embrace the messiness of natural systems. His lab’s “empirical dynamic modeling” techniques use time-series data to look at the invisible ways these complicated systems are connected: like plucking one string out of a jumbled network and seeing which other strings echo back. In 2012, his group published a widely cited paper describing some of these techniques and their usefulness for studying sardine and anchovy populations. These methods have since been used for all sorts of different problems—from probing how greenhouse gases affect the climate to diagnosing early onset Alzheimer’s disease.

A group of researchers led by Sugihara’s postdoctoral researcher Ethan Deyle now has applied their fishy methods to flu, looking at 18 years of data on worldwide influenza outbreaks. They found that humidity is the strongest factor in driving influenza, but that temperature also plays a role, combining in a complex way that past studies were unable to pick up on. At temperatures up to about 24°C, drier climates encourage flu outbreaks, as seen in North America. Hotter than that, and the relationship flips: Suddenly, wetter environments are better for the flu, they write today in the Proceedings of the National Academy of Sciences.

The way this relationship flips around 24°C also lines up with what researchers already know about the flu virus. It has a “viral envelope”—similar to a cell membrane—that has two distinct weaknesses: It can swell and burst if the climate is cold and humid, and it dries up when the environment is hot and dry. The authors say that these twin forces could explain why the relationship between temperature and humidity flips around 24°C. If this idea turns out to have merit, then it might help fight winter flu outbreaks by, say, simply putting humidifiers in schools, homes for the elderly, and hospitals.

The study also has an answer to why flu patterns are so different between tropical and temperate areas. The winter cycles of flu in temperate regions are due to the big, consistent changes in humidity and temperature. In the tropics, the same underlying rules link flu to the climate, but seasonal swings in temperature and humidity are so weak that a “flu season” doesn’t emerge from all of the noise.

The work is “a great analysis,” says Jeffrey Shaman, an infectious disease climate scientist at Columbia University. But Shaman cautions that the team’s proposed mechanism is just one of a number of ways climate could act on the flu and needs to be tested further. He agrees that modifying indoor environments has untapped potential as a way of controlling disease transmission—but even if it’s possible to make the environment harder for a flu virus to spread, that could just create the ideal habitat for moisture-loving bacteria and molds. “There is no such thing as a free lunch,” he says.

Others are less impressed with the analysis. Julian Tang, a clinical virologist at the United Kingdom’s University Hospitals of Leicester NHS Trust, notes that the humidity and temperature data used in the study come from outdoor weather monitoring stations, whereas it’s believed most flu transmission occurs indoors. “If you don’t use the right data,” Tang says, “whatever model you develop and apply, no matter how innovative, is likely to give you the wrong conclusions.”