What Does A Migration Look Like?

Every afternoon at about three, my officemate puts on her coat, picks up her binoculars, and walks down to Sakuma Viewpoint, a little patch of park across the street from our building on Portage Bay, to do a five-minute survey of all the birds in sight. She is what you might call a bird nerd—her research is on seabirds, and when she needs a break from her research, she goes outside and counts them for fun. Every so often, I join her, to get a breath of fresh air and remind myself how little I actually know about birds.

These short surveys are part of a larger project, run by the Audubon Society and the Cornell Lab of Ornithology, called eBird. Birdwatchers from all over the country do their own short surveys and send in their sightings electronically. What good are a motley assortment of observations, neither fully random nor systematic in design, made by a bunch of amateurs? Turns out, when done on a continental scale and stitched together with some geostatistical wizardry, a whole lot. Sightings made by eBird’s network of citizen scientists, coupled with geographical and climatological knowledge, allow the scientists at Cornell to produce truly spectacular maps of bird migrations across the United States.

The map above shows the peregrinations of the wood thrush, Hylocichla mustelina, during 2008. The brightness of the colors indicates the probability a wood thrush occurs in a given county. Thrushes move up the Mississippi Valley in the spring, spread out to cover the entire Eastern U.S, and then quickly file south again along the Appalachians in late summer, headed to their wintering grounds in Central America.

This is just one of a long and growing list of migration maps the Lab of Ornithology has produced; all can be viewed here. These maps are a truly spectacular use of the eBird data, but the models behind them are hindered by a lack of data in some regions—in particular, the great plains and the inland Northwest. If you’re in Montana and reading this, consider heading outside and adding a few data points.