Cloudy issues, burning answers: Jeremy Cherfas finds the best way to study the effect of fires on climate is to start one

ON 16 September, an ecologist will walk into the bush in the middle of South Africa's Kruger National Park and drop a match. The flames crackling across the veld will signal the start of an unusual project by the Southern Africa Fire-Atmosphere Research Initiative (Safari), a massive international effort to gather data on fires and how they affect the earth's atmosphere.

The man who will drop the match, Johann Georg Goldammer, is head of the Fire Ecology and Biomass Burning Research Group at the University of Freiburg in Germany, and co-chair of the Safari steering committee. He is also a driving force in what he calls 'global fire science', and at a recent meeting on fire and the environment organised by the Dahlem Konferenzen in Berlin, insisted that 'we must know the impact of biomass burning on world atmosphere'.

Why? Because vegetation fires fill the atmosphere with staggering amounts of chemicals: a quarter of the global carbon monoxide and nitrogen oxides, a third of the non-methane hydrocarbons, and a third of the ozone in the troposphere, along with many other reactive compounds. Are these pollutants, or integral parts of natural cycles? And if biomass burning - most of which takes place in the developing South - does prove to be a factor in global climate change, will that take the pressure off the developed North to clean up its industrial emissions?

Although there are good models for predicting the behaviour of fires, most have been developed to deal with the kinds of conflagrations that beset the North American timber industry. They are not much good for treating fires being used to clear rainforests, and almost useless for savannah fires.

For example, these models do not clearly distinguish between smouldering and blazing fires. As a fire spreads, it generally advances along a blazing front. Gases produced by the breakdown of the vegetation burst into flame, producing a preponderance of oxidised compounds and smoke, which absorbs sunlight.

Once the flame has passed by, the rest of the fuel smoulders, still burning, but without a flame. The products of a smouldering fire are different, producing more reduced compounds, such as methane, carbon monoxide, and smoke that is less light-absorbing.

Those differences almost certainly imply great variation in the impact of smouldering and blazing fires on atmospheric chemistry and global climate, but nobody knows how much vegetation flames and how much smoulders.

Safari will try to find out the answer, at least for savannah fires, which are among the most common types of deliberate fire. Dr Goldammer says 1.5 billion hectares (3.6 billion acres) of savannah are 'waiting to be burnt' each year, because setting fire to range-lands turns the unpalatable dead thatch into fertiliser that encourages a flush of palatable and nutritious young growth for cattle and other grazers.

The interactions between these fires and atmospheric chemistry are intricate and far-reaching, many of them mediated by the hydroxyl radical. Hydroxyls oxidise important atmospheric pollutants, mostly in equatorial regions.

Paul Crutzen, an atmospheric chemist at Stockholm University, says: 'The hydroxyl radicals are nature's 'detergents' and the tropics her 'washing machine'.' Emissions from fires are cleaned up by hydroxyl radicals, but in the process they destroy those radicals. That, Dr Crutzen says, 'lowers the oxidative capacity of large regions of the atmosphere'. In other words, biomass burning uses up the detergent, leaving the washing machine unable to do its job properly.

Then there is ozone, created by light in the presence of oxides of nitrogen. Nasa scientists have detected excess ozone streaming out over the southern Atlantic during the burning season in southern Africa. Safari should help establish how much ozone these fires create, which will enable modellers to assess its impact on climate.

And of course there is no fire without smoke, which may actually be keeping the planet cooler by reflecting sunlight. Smoke also provides nuclei on which raindrops can condense, thereby altering local rainfall patterns.

Fires have many effects, says Dr Goldammer. They shift nutrients around, change ecosystems, and probably affect climate.

The big issue is whether any of this is of wider significance. Biomass burning has been going on for at least 300 million years, ever since there were plants on land to ignite. And people have probably been using fire for at least half a million years. 'What's new,' says Steve Pyne, a historian and anthropologist at Arizona State University West in Phoenix, Arizona, 'is that global climate is making regional fires a world problem.'

Dr Crutzen explains: 'There is much more biomass burning than there was 100 years ago, of that I am convinced absolutely. Fossil fuels are definitely a driving force on global climate, but biomass burning, that's a very young idea.'

That is why Dr Goldammer is so keen to take everyone with an interest, from fire ecologists to global modellers, on safari. Almost 60 scientists will take part, and although the disastrous droughts in southern Africa have required some scaling back - there just isn't enough vegetation to burn in some places - Safari remains ambitious.

On the ground, crews will study the amount and type of fuel available. They will measure the behaviour of the fires, while overhead, scientists in low-flying aircraft will sample emissions in the smoke plume. High-flying planes and balloon-borne instruments will extend the measurements and tracking, and the whole will be watched over by satellite-borne sensors. Later efforts will look at the impact of fires on plant nutrients and ecosystem changes.

The goal is an improved model that integrates fires, weather and climate. 'We would like a technologically and geographically seamless system,' says Dr Goldammer. Such a model would integrate satellite information about the amount of fuel with recent weather information, to predict the risk of fire. It would also calculate the effects on the atmosphere of fires spotted from space.

(Photograph omitted)