Every dot is land, tinted by its temperature that month. Every line is an animal that was really there. Watch the world breathe through the seasons while the animal moves, and notice how the ocean-wanderers keep the same colour the whole way around.
A sooty tern covers 160° of longitude inside a band just 3.4°C wide. It isn’t wandering the map, it is following the water. Land-breeding migrants do the opposite: the bar-tailed godwit travels just as far, but punches straight through 19.6°C of thermal range, from Arctic tundra to tropical coast.
cold → warm · width = 5th to 95th percentile of temperature experienced
Colour is how many species have been tracked through each 5° cell. Grey is land where no tracked animal has ever been recorded. This is not a map of animals, it is a map of where science has pointed its antennas.
Coverage flatters us. A 5° cell is 550 km across, so one stork crossing the Sahara once makes it look studied. Measured by intensity, a typical tropical cell holds 381 fixes against the subtropics’ 1,287, and polar land holds a median of just one species and eleven fixes.
Every limit below was learned by being refuted, a confident result that fell apart under checking. They are the reason to trust the findings above them.
Asked where migration converges, the corpus answers “Belgium,” the home of its largest data source. Three methods each rediscovered the funders, not the biology.
Stork arrival dates vary ~13 days between birds, so a 5.5-day shift is detectable. Wintering latitude varies by 70°, so a 1–3°/decade shift never will be.
Raw minima and maxima produced a killifish at the North Pole, a 430,000 km pintail, and five phantom circumnavigators that were Bering Strait animals straddling the dateline.