In 1880, teacher Wells Cooke began writing down when migrating birds arrived; 90 years later, the project had left millions of records scientists still use today

Long before crowdsourced data collection existed on phones or websites, one of its earliest large-scale examples ran on handwritten index cards. In 1880, Wells W. Cooke was posted to the White Earth Indian Reservation in Minnesota, where he taught by trade but spent his free time observing the birds passing through each spring. A U.S. Geological Survey (USGS) history of the North American Bird Phenology Program, the agency that now holds and curates the original records, recounts that Cooke became intrigued by a simple question that was largely unanswered at the time: did migratory birds arrive on the same dates each year?

With the help of a network of friends and acquaintances that he recruited, Cooke collected first sightings for different species of birds, information which he then carefully filed onto cards sorted by species, location, and date. It was a crowdsourced endeavor. That hobby would go on to become a federal program spanning nine decades and amass one of the largest migration records of any North American bird group.

From one enthusiastic hobby to a national data machine


Cooke's little side project was noticed by the newly founded American Ornithologists' Union, which hired naturalist C. Hart Merriam to take it nationwide. A 2010 NOAA Climate.gov feature on the program's history reports that this network of birdwatchers stretched as far as the Caribbean and Canada, with an estimated 3,000 people across the U.S. submitting sightings to the program, according to the entry on the North American Bird Phenology Program page.

Administration of the program then passed to the U.S. Department of Agriculture's Division of Economic Ornithology, which later became part of the U.S. Fish and Wildlife Service before the surviving records were ultimately transferred to the USGS's Patuxent Wildlife Research Center, where they are held today. This administrative and funding structure was distinct from the volunteers themselves: the thousands of birdwatchers who mailed in their sightings were never paid, but the office that recruited them, mailed out standardized recording cards, and filed the results was a funded federal operation with paid staff. The program ran for about 90 years, from 1880 to 1970, before being discontinued, leaving behind roughly six million handwritten cards in storage that were largely left untouched for decades.

Why a box of old index cards could explain your climate anxiety

Those cards represent one of the longest continuous sets of bird arrival records in North America, used by researchers to understand how migration has changed as the world has warmed. According to a 2012 study titled “Reviving a Legacy Citizen Science Project to Illuminate Shifts in Bird Phenology,” published in the International Journal of Zoology, scientists used the reconstructed set of arrival dates to model changes in migration timing for birds like the Ruby-throated Hummingbird and the Purple Martin, to create a baseline of sorts for how the species behaved before 1970 that previously didn't exist. Were it not for Cooke's little network, there would be no reference point for how "early" or "late" these arrivals were compared to the present day.

And those changes have not necessarily been good for the birds. In a separate 2019 study titled “Phenology of nocturnal avian migration has shifted at the continental scale,” published in Nature Climate Change, a different group of researchers used 24 years' worth of weather radar data to track changes in spring and autumn migration across North America. They found that on average, birds are migrating earlier in the spring, with the most rapid shifts taking place in the areas undergoing the most rapid warming. Meanwhile, the autumn exodus seems to have seen much smaller changes in timing, hinting at a possible disconnect between migration and food availability that could affect species that rely on insect life or plant matter to survive.

An old idea, revived by the same method

Crowdsourced data collection is often associated with recent, internet-based projects, but the practice long predates them. The effort to digitize the cards and analyze them, beginning in 2009, echoes the original creation of the network in one notable way: the initial digitization effort was also carried out largely by unpaid volunteer transcribers, working through the scanned records online in their own time.

That parallel is worth keeping in mind amid ongoing debates over funding for climate and biodiversity data collection. The most enduring record of North American bird migration was not built by a single federal office with a mandate to observe and report on the country's natural resources, but by a much larger effort: a federally coordinated program that depended, from the very start, on volunteers faithfully noting every sighting they could.