If we are lucky, an international climate agreement will be forged in Copenhagen later this year and emissions targets set in a bid to limit global warming to 2 °c above preindustrial levels. Agreement in Denmark or not, I find the breezy way so many politicians and commentators talk as if such an increase were no big deal truly amazing. The truth is we have no idea what will happen at those places where you and I live if we raise the global thermostat 2 °C or more. Worse, there is another factor that could crank that thermostat still higher, making life quite as intolerable as the well understood threat from rising sea levels: fire. The fossil record shows that fires started to occur soon after vegetation was established on Earth during the Silurian, about 420 million years ago. Ancient terrestrial life’s exposure to fire gives us good reason to think fire is an important evolutionary factor and, more controversially, that life co-evolved with fire. Fire can be seen as a physicochemical process, a “fire triangle” of oxygen, fuel and heat for ignition. Combustion can occur if the concentration of oxygen is higher than 13 per cent, and variation in oxygen levels correlates with fire activity in Earth’s history. Fluctuations in atmospheric oxygen through geological time significantly affected fire risk. For example, in the Permian, oxygen levels were substantially higher than now, and even moist giant moss forests sometimes burned. Those Permian coal fossils flag up another key detail: the burial of decay-resistant charcoal and organic matter may have led to long-term reductions in the proportion of CO2 in the atmosphere and increased relative oxygen levels, because the carbon is geologically sequestered as coal and the oxygen left in the atmosphere (until chemical weathering draws it down). So fire in the biosphere should be considered both as a physicochemical process and as a fundamental biogeochemical process, feeding back between biosphere, geosphere and hydrosphere.

Seeing fire as biology seems odd, but landscape fires only occur because life creates fuel: so some ecologists now say fires should be seen as “biologically constructed”, drawing parallels with decomposition and herbivory. For me, then, the Greeks had it right with their classification of fire, air, earth and water. But we have much to learn about how life and fire affect each other. Take tropical savannahs, the most flammable vegetation on Earth. Researchers think that falling atmospheric CO2 concentrations about 8 million years ago stimulated the global development of tropical savannah dominated by grasses which use the C4 photosynthetic pathway. These tropical grasses are highly productive in hot, wet climates, and under low CO2 concentrations they have a physiological advantage over woody vegetation, which uses the C3 pathway. Savannah was stimulated because those grasses produce large quantities of fine and well-aerated fuels, greatly increasing the frequency of fire, further disadvantaging woody plants because frequent fires create a population bottleneck by killing “fire-tender” juveniles. Savannah trees had to develop rapid growth to escape the fire trap. Expanding savannahs may also have caused a climate feedback that created hotter, drier conditions, favoring yet more savannah : one of the great examples of fire-vegetation-climate feedback proposed by some ecologists.