This Saturday, May 9, is birding鈥檚 biggest day. It is called , an annual 24-hour event in which thousands of people around the world help crowdsource sightings of for the Cornell Lab of Ornithology, considered one of the leading avian research institutions in the world.
Thousands more tune in to the lab鈥檚 and use the lab鈥檚 downloadable , which can recognize bird sounds and images, both free to users.
But on Saturday people young and old will upload sightings, primarily using , a massive database, with both a website and free , with checklists birders can send in all day.
The goal of the Cornell Lab of Ornithology is to help people connect to the joys of nature. But it also turns all who participate into global citizen scientists, because the data that comes in along with decades of bird recordings from amateur and professional birders, is free to researchers who use it to study the natural world and the impact of things like climate change on birds, and subsequently all of us.
鈥淭o me, it鈥檚 all built on the power of birds,鈥� Ian Davies, international strategy lead for eBird said. 鈥淏irds bring people together. Everything is free and everything鈥檚 open source. So, researchers use this to look at climate change impacts, impacts of changing human habitation.鈥�
As we talked, we looked at a screen showing a map of the world filled with little yellow dots representing people in places like Mexico and Germany who were uploading in real time the birds they were seeing. On that same screen, there were silver discs.
鈥淵eah, those are checklists that people have submitted in the last 10 minutes or so. So today, there has been 18,000 different people who have gone out and shared what they鈥檝e seen. Ten, 20 right now as we鈥檙e talking,鈥� Davies said.
鈥淪o, if I want to go find a bird, find an American woodcock, I can go do that right now,鈥� he added. 鈥淪o, every single point on this map is a place where in the last 30 days, someone has gone out and seen the species American woodcock. You can see where to find them and also get information on what they look like and what they sound like. They have one of the most distinctive sounds of a spring evening.鈥�
The data on eBird is what teaches Cornell鈥檚 other app, Merlin, to identify bird sounds and images from citizen scientists. The lab also uses as audio tapes collected by birders over the past decades.
鈥淎 lot of our knowledge of even what a bird sounds like came from this generation of recordings,鈥� curator Glenn Seeholzer said. 鈥淥ftentimes the first recordings of a species or even finding new species.鈥�
The room we鈥檙e in is lined with shelves of old field books and binoculars from past contributors.
鈥淚f you look on Merlin in the reference recordings, a lot of those recordings are from these previous generations, because oftentimes they got to places that people can鈥檛 go now or recorded species that are now extinct or hard to find or just happened to serendipitously make the best recording that maybe anyone will ever make of one of these species,鈥� Seeholzer said
Another room is filled with audio tapes, some of which look ancient.
鈥淲ell, this is from a French recordist, he built his own private sound collection. His daughter, after he passed away, donated it to us,鈥� Seeholzer said as he looked at old bird recordings. 鈥淎nd he spent 50 years traveling throughout the world recording all this analog audio.鈥�
One canister is labeled 鈥渋vory-billed woodpecker.鈥� It鈥檚 35-millimeter film containing footage of one of the last times a pair of ivory-billed woodpeckers were documented in the wild. The images were captured by lab founders Arthur Allen and Peter Paul Kellogg.
鈥淣ow you can look online and look at these original tapes, rare and declining birds that are now extinct,鈥� said Seeholzer.
There are three million pieces of sound at the bird lab, but it all has to be annotated. Merlin doesn鈥檛 actually hear sounds. It sees a spectrogram of sound waves, markers laid down using a machine learning process developed at the lab to teach Merlin to not just distinguish species, but to hear regional accents, if you will, within a species. Cardinals, for instance, sound different depending on where they live in the country.
鈥淭he name of the game with training this machine learning model is diversity, so we don鈥檛 want just to train it on cardinal recordings from upstate New York. Right? Because then you go to the Midwest and maybe they sound a bit different there or maybe the ones in Florida sound different still,鈥� said annotator Jim McGowan.
Birds are disappearing. have shown North America has nearly 3 billion fewer birds than it did in 1970. Over 100 species have had no sightings.
But there are reasons for hope. Scientists at the lab are working with government agencies and other researchers on everything from loss of habitation due to construction and forest fires to pesticides.
鈥�Birds are sentinels of the ecosystem health,鈥� said Cornell researcher Andrew Stillman. 鈥淲hen the birds disappear, other things are disappearing as well. eBird data has been part of sounding the alarm about birds disappearing and declining. But at the same time, that same data collected by birders around the world, regular people, is the dataset that we can use to start to identify better solutions to the problem.鈥�
This article was originally published on
Copyright 2026 WBUR