Climate models have a simple story: as the world warms, animals move toward the poles or up mountains, following the cooler air. It's intuitive, mechanistic, and almost entirely wrong. According to recent analysis of nearly 30,000 documented range shifts driven by temperature changes, less than half of observed instances moved to higher latitudes or elevations as expected. Many didn't move at all. Some moved south or downslope—into warmer territory.
The standard assumption is that temperature is the dominant force shaping where animals live. When it gets hotter, creatures should flee toward cooler refuges. This logic drives conservation planning, migration corridors, and protected area networks across the globe. Climate scientists and ecologists have spent decades modeling exactly which species will end up where as the planet warms, assuming temperature gradients act like an invisible escalator pushing life toward the poles and peaks.
But the empirical reality is messier. A striking example comes from the red squirrel. Despite warming winters that should theoretically allow it to venture higher into mountain ranges, the red squirrel has instead moved downslope into regenerating forests at lower elevations, according to research highlighted by Mongabay. This isn't a rare exception. The analysis of nearly 30,000 range shift observations found that fewer than half aligned with the predicted pole-ward or upslope movement. The rest either stayed put or shifted in unexpected directions, including toward warmer areas entirely.
The mechanism behind these counterintuitive movements reveals how incomplete our models were. Temperature isn't the only driver of animal distribution—or even always the primary one. Food availability, habitat structure, and interspecies competition often outweigh thermal comfort. A warming climate that triggers earlier plant growth might pull a species downslope toward new food sources, even as higher elevations become thermally tolerable. The red squirrel's downslope shift, for instance, tracks the availability of suitable forest regeneration rather than simply the thermometer. Evolutionary history matters too: species may be physiologically adapted to exploit opportunities in their current range before investing in expanding into new territory, even if that territory becomes technically habitable.
There's also a lag effect that models rarely capture. Species don't respond to climate change in real time. A population might still be establishing itself in a recent range when the climate has already shifted again. Dispersal itself is slow and constrained by fragmented landscapes, predation, and competition. What looks like "wrong direction" movement from a model's perspective might simply reflect that organisms are responding to multiple, overlapping signals—not just temperature—and that the biotic world operates on timescales that don't align neatly with physical climate projections.
The practical implication is humbling: we may be building conservation infrastructure based on incomplete predictions of where species will actually need to be. Protected areas designed to intercept pole-ward migrations might miss populations shifting laterally or downslope. The unpredictability also suggests that climate adaptation for wildlife isn't simply about connecting cooler refuges; it's about preserving diverse habitats and ecological networks flexible enough to accommodate range shifts we can't yet predict. In a warming world, it turns out, the animals know something our models don't.