Climate change is a major threat to human health and happiness, and to the species and ecosystems on which all economies depend, set to escalate rapidly over coming decades. Of greatest concern is that climate warming could simultaneously push many species beyond critical thermal thresholds, leading to the collapse of ecosystems and sudden and widespread biodiversity loss. The ability to predict where and when biodiversity is at risk of abrupt collapse is an urgent priority to drive emissions, conservation and food security policy, needed to avoid the most damaging impacts.

Our own research suggests that even for widespread species, most local populations exist under a relatively narrow set of temperatures, so that gradual warming can suddenly push large areas of a species geographic range into hotter conditions than they can tolerate. One analogy is to imagine standing on a beach as the tide comes in. Where the beach is flat, large areas are suddenly inundated by the sea, but where the beach is steep, the sea encroaches slowly, even though the tide moves at a constant speed. We think the same applies to climate warming. Initially species may only be exposed at the very edge of their geographic range-like a steep bank on a shore. However, once the bank is breached, small increases quickly expose large numbers of individuals to damaging conditions, causing species to disappear abruptly from communities across their range.

The big unknown is where and when fine scale variation in climate and in the ecology of organisms can reduce the risk of such abrupt collapses. One possibility is that species can disperse to nearby cooler sites (e.g. higher elevations or depths) or behave in ways that buffer extreme temperatures-either by seeking shade, suspending development, or breeding, flowering or fruiting in different months or years. Such organismal responses to environmental variation in space and time provide poorly understood resilience to climate warming, reducing the risk that populations, species and communities collapse at the same time.

This project will study the extent to which variation in the response and evolutionary history of organisms will reduce the risk of abrupt collapse. Using data on the geographic distribution of 1000s of animal species, for which exceptional data are available, we will ask: why are thermal thresholds more abrupt for some species compared to others? For example, we suspect that birds and mammals-species that maintain a relatively constant body temperature are more thermally resilient than amphibians, reptiles and insects where performance depends on environmental temperatures.

We will then test how the behavior of given species may buffer against warming, and for how long. Using data on changes in the distribution of European birds over recent decades we will ask if more dispersive species and those that can breed at different times, have more stable populations in places that our models predict are already crossing critical thresholds. Using data on the fine scale distribution and phenology of butterflies we will test if the ability to seek shelter from the heat explains why some species have persisted at low elevations where temperatures are rapidly rising, while others have become extinct.

Finally, we will include these biotic responses in our models and ask whether they are enough to avoid abrupt species collapse, whether they simply delay inevitable biodiversity collapse, or potentially even make its effect worse: like the sea wall holding back the tide that eventually crumbles. Using our models we will better understand which species and ecosystems are most at risk of abrupt collapse, and when, under different emissions scenarios, enabling us to better target conservation efforts to safeguard these places, and predict which parts of the globe will remain productive for agriculture or wild harvesting, or will persist as carbon sinks, in the decades to come.