The start of West Nile virus transmission began unusually early in some US states, with the most cases in early summer since 2004, the US Centers for Disease Control and Prevention (CDC) warned in July – so Delaware turned to chickens.

“They kind of are the proverbial canary in the coal mine,” said Tom Moran, head of the mosquito control section at the Delaware department of natural resources and environmental control.

Sentinel chickens, as they’re called, serve as an early warning for mosquito-borne diseases, especially Eastern equine encephalitis (EEE) and West Nile.

They can often predict when virus activity begins ramping up, allowing officials to target mosquito control efforts and warn the public about taking precautions. Doctors know to be on the lookout when patients have flu-like symptoms – especially when it’s not flu season – and that can help reach diagnosis faster. Veterinarians can make sure horses are vaccinated.

The chicken itself is a “dead-end host”, Moran said. They can test positive for the viruses, but they can’t spread the diseases, which is why they’re ideal sentinels, he said.

Delaware keeps flocks of four chickens each at 20 locations throughout the state, a program that began in the early 1980s. It’s a relatively lo-tech way to monitor and prepare for the viruses, which can be deadly.

“It is kind of the old-school gold standard,” Moran said. Some states have transitioned to testing mosquito pools, which can be a faster way to detect viruses, because it takes about a week for the chickens to develop antibodies after being bitten.

They tried both in Delaware, but decided to go with the chickens only. “The chickens sometimes pick up virus activity in an area that the mosquito pools did not,” Moran said. Unlike the insects, they don’t fly away – which means officials can be certain that infected mosquitoes are active in that exact area.

“It gives you a little more surgical information,” Moran said. “You take that information and try to use it to lessen the specter of any human transmission.”

Cases usually start appearing from north to south as migrating birds pass the viruses to mosquitoes, Moran said. Delaware’s season usually begins in early August.

The chickens could also be tested for other mosquito-borne illnesses if they emerge. That’s what happened when West Nile appeared in Delaware starting in 2000. At first, officials would test wild birds found dead to see if they had the virus, but they soon began testing the sentinel chickens they already had.

Some places test for other local mosquito-borne illnesses, suchas St Louis encephalitis. If a neighboring state were to detect a pathogen like that, “we could probably pretty quickly get up to speed and start testing for that here too”, Moran said.

The decades of data collected from sentinel chickens on virus patterns can be invaluable for creating models that predict the severity of a season. That’s what Lindsay Campbell, associate professor of entomology and nematology at the University of Florida, has focused on in her research. She digitized decades of paper records from Florida’s Department of Health and mosquito control programs, and she’s also drawn on environmental patterns – temperature, precipitation, land cover – as well as community science data like eBird to track bird migration patterns.

The data, and the models built from it, can offer insight into the conditions driving good and bad years for West Nile and EEE.

The sentinel chickens are providing important data “telling us what happened, while we would like to know what might happen in the future”, said Robert Guralnick, curator of biodiversity informatics at the Florida Museum of Natural History and a co-author with Campbell.

“It gives us the opportunity to go ahead and see the ebbs and flows, essentially, of these viruses over time, and then we can try to link them with different environmental characteristics to see if we can maybe predict when there might be elevated activity in the natural environment,” Campbell said.

There are many factors likely involved in how diseases like these spread – the hosts, like mosquitoes, birds, and other animals, as well as environmental conditions. That makes it complicated to predict how severe a year will be. But these models are examining, for instance, if weather patterns or temperatures from four or six or twelve months earlier affect how bad a viral season is.

“For some viruses, like EEE virus, we’ve seen a strong association with a certain type of landscape that is associated with the mosquito vector – that’s really important for that virus,” Campbell said. For West Nile virus, they’ve identified some precipitation and temperature patterns a few months before chickens start testing positive.

“The information that these programs have been able to provide is really invaluable,” Campbell said of sentinel chickens. Guralnick agreed. “It’s a clever way to gather really good information about what’s happening,” he said.

Over the decades, officials have tried new things and learned what works best. For the first few years of the Delaware program, they used white leghorn chickens – but they proved, in Moran’s words, to be “very flighty”. They frequently tried to escape, running into the bushes and forests.

“It’s literally like a scene from Rocky, where you’re chasing the chicken around – you can’t catch these things,” Moran said. It would sometimes take an hour to track the chickens down. Once they switched to a breed of Rhode Island red chickens, their work went much more smoothly.

Similarly, officials have experimented with other means of tracking mosquito-borne illnesses, including one research project that proposes using honey traps to catch the insects. “We’re always looking: is there a better way, a faster way, a cheaper way, a more accurate way – it checks all those boxes?” Moran said. “Folks are always pushing the ball forward.”

But for now, chickens rule the roost.