We are experiencing an increase in extreme natural events. Or at least, that is the general perception. Just reading the pages of this journal, for example, brings news of the effects on biodiversity of extremes of high1 and low2 temperature, drought3, precipitation4, wind5 and fire6.

However, showing definitively that such events are increasing has often proved elusive. In this issue of Nature Ecology & Evolution, Cunningham et al. present two decades of global satellite data that show that the incidence of extreme wildfires has indeed more than doubled in this time. To demonstrate this, they needed to be precise about exactly what was being measured. Most fires are low intensity and are often ignited by humans; they are not especially damaging and they can have a crucial role in ecosystem health. These sorts of fires have decreased over the past century, as has average fire intensity.

By contrast, energetically extreme fires have increased. Cunningham et al. show this by measuring fire radiative power summed across multiple nearby locations, to capture the most extreme events. Not only have these increased in frequency, but they have also doubled in magnitude over the time period studied. Moreover, the six most extreme years for such events have all been since 2017.

As detailed for the 2020 fires in Australia6, an increase in extreme fire events will have major effects on biodiversity. Other types of extreme event can also be detrimental to biodiversity, so any rise in their frequency is a matter of concern. For example, hurricanes can affect coral dispersal and spread disease7 and also affect evolution and extinction in insects8; extreme cooling events are linked to mortality in marine species9; marine heatwaves can cause regime shifts in intertidal ecosystems1; and both hot and cold extreme events affect reproductive fitness in birds10. These studies also illustrate the diversity of approaches that are needed to assess the effect of extreme events, including long-term monitoring of wild populations, experimental manipulations and modelling.

Not all species will be equally affected, and it is important to understand any systematic differences. A recent meta-analysis in this journal showed differences in how native and non-native species respond to extreme weather events11. That analysis found most marine species were overall insensitive to extreme weather events, except for heatwaves — but individual studies, including some cited above, indicate that this cannot be taken as general rule. The meta-analysis also found that effects are more widespread in terrestrial and freshwater systems, but with a much higher proportion of native than non-native species involved. This also illustrates the potential for interactions between extreme events and other global change phenomena, such as invasive species.

The rise in extreme events that could catastrophically affect biodiversity will bring an increase in rapid decision making about prioritizing what to protect. In many cases, the need to protect biodiversity will compete with the needs of human life, infrastructure and cultural assets. A recent study that surveyed the priorities of the Australian public showed, unsurprisingly, that human life is valued above all else, but — perhaps more surprisingly — that biodiversity is often valued above property12. However, the desire to protect biodiversity was much higher for an iconic species of marsupial (the koala) than for critically endangered invertebrate or plant species.

The only way to decelerate the increase in extreme events is the rapid reduction of emissions. However, even if this were to occur, a certain degree of increase is inevitable. It is therefore important to direct research efforts at understanding how to live with these increases. Further work is also needed to document how different types of extreme events have increased and will continue to increase. The nuanced ways that they affect different types of species and ecosystems — via demographic, physiological and evolutionary mechanisms — need to continue to be documented both in real and experimental systems. And conservation practitioners and managers will need to further incorporate extreme events into both short-term and long-term planning decisions.