Euan Nisbet joins the band of dedicated experts who monitor the Earth's lungs, while Andrew Goudie considers life's bleak future in a warming world

The parties to the Kyoto Protocol met in Montreal, Canada, for the United Nations Climate Change Conference. They agreed to meet again. Here is a story of the people who keep watch on the greenhouse gases. They agreed to meet again too.

Denver airport, Colorado. After a squashed flight (but our lucky poster describing European methane monitoring upgrades to business class), we find the bus queue. A man bounces down the bus steps. The bus fare is $10, the driver has no change and the man only has a $20. If I give him $10, he'll pay for us both: we can solve the problem together. I recognise him - he's an American atmospheric genius and a good violinist to boot.

While the bus rumbles along to Boulder we put the world to rights, then spill cheerfully out of the bus onto the tarmac by a gas station. We hump our bags in light rain another half-mile along a busy road to our comfy and cheap motel. The reunion has begun.

É«ºÐÖ±²¥

Next morning over muffins and coffee, the friendly Finns appear. So does a team from the Max Planck Institute - they are superstars able to measure oxygen with extreme precision (the flip side of the carbon dioxide coin). It's a 20-minute walk to the gathering. The meeting room is flanked by coffee urns and piles of doughnuts and fruit. We pin our posters up by the strawberries.

There are tables to spread out on: the chap in front connects up via an extraordinary three-decker adaptor plug. Welcome to the 13th biannual meeting of the "keepers of the greenhouse": the people who measure carbon dioxide. In officialese, we're the UN World Meteorological Organisation/International Atomic Energy Agency Global Atmosphere Watch Experts Panel on CO2 and Tracer Measurement.

É«ºÐÖ±²¥

The Global Atmosphere Watch is just that. Everyone watches. Mostly we collect air and watch lab instruments, but some sniff air from watchtowers - high TV towers.

All the greenhouse measurement community is represented. Next to me the South African researcher chats to his Indian counterpart. Seated around the tables are Brazilian, Hungarian, Kyrgyz, Spanish, Russian, Swiss, Dutch and Italian scientists, as well as greenhouse measurement's Premier League: the Canadian, French, German, Japanese, Chinese and Kiwi teams. CO2's equivalent of Manchester United, Arsenal and Chelsea are the US groups.

Some participants are from government institutes, others are university based. Many are women.

But there's a gap - the superb Aussie team, keepers of the South, is decimated. World-leading scientists have been put out to grass by brutal restructuring. It's a terrible loss, not just to Australia but to the whole planet. The Kiwis are trying to fill the gap.

Most groups are growing though. Our group is reporting on a just-ended European Union project, thanks to travel support from the Royal Society.

There's another gap, as usual. The UK is the major absentee.

The Natural Environment Research Council's remit excludes greenhouse monitoring, which is not "blue-skies" research (actually CO2 is about the infra-red skies). Fortunately, the Department for the Environment, Food and Rural Affairs helps support work for Nasa and French networks to measure the Atlantic background in Ireland while our small university and EU-supported long-term monitoring keeps going. Australia used to monitor Scotland (now Germany and Edinburgh University do, through EU support) and the US monitors remote British islands such as Bermuda.

We start the meeting on a sad note: Dave Keeling has died. These meetings started in the 1970s, when Keeling and friends got together round a small table - a Kiwi and a Canadian from the original meeting still attend. In the 1950s, he started to measure carbon dioxide in the best place he could find - on Mauna Loa, Hawaii. What he found through his painstaking work forms the basis of the Kyoto Treaty. The Keeling curve shows the climb in CO2. It also records the seasonal "breathing" of the Earth - the North exhales in autumn and inhales when leaves grow in spring. The CO2 and methane "rugs", or flying carpet, which the US meticulously measures, is the life breath of the planet.

É«ºÐÖ±²¥

É«ºÐÖ±²¥

We measure the Earth's lungs as it inhales and exhales. From New Zealand, we can backtrack carbon to African fires. From Galway we isotopically smell methane from Canadian swamps and forest fires. Our Siberian airtanks sample the giant gas fields that supply Europe. In London, we look at local emissions; methane from landfill is isotopically different from gas-leak methane. By continuously measuring London air and radon emissions as a calibrator, we can identify greenhouse gas sources because these vary from morning rush hour to bathtime.

On this score, there is good news and bad. Since former Environment Secretary John Gummer's reforms, London's air has been much cleaner, containing less methane and carbon monoxide, but its CO2 level is rising fast.

Ours is a technical meeting. The press wouldn't like it because the heroes here are US teams. Global CO2 monitoring depends on US government scientists, who are deeply worried about climate change.

Talk focuses on co-operation - intercomparison and standards - the keys to global partnership. If we measure 1,820 parts per billion methane in Atlantic air blowing towards London, which subsequently reaches Holland with 1,850ppb, then London has added 30ppb. But that is true only if we are using the same measuring scale as the Dutch. How can we check they are the same?

For atmospheric work, meticulous calibration (against the global US standard) and mutual intercomparison (measuring the same sample tanks in London and Holland) are essential. It is hard to fund this (though Brussels helps), so we mostly have to bootleg it. The expert is a much respected German. She leads the final discussion on "Melons" and "Sausages" - different types of containers of air - which are shipped around the world for testing. By comparing their analyses, separate national monitoring programmes weld into a global dataset that the modellers can use. We have become a worldwide network of friends, sharing similar problems and swapping gas cylinders.

Four of us from the EU lunch daily during the gathering around a wobbly courtyard table, planning proposals. Dinner one night is with the Kiwis, who are among the very best. One leads a key part of the next Intergovernmental Panel on Climate Change, the others run measuring programmes from the high Arctic to Antarctica. One evening, we dine with the South Africans. On another night, a welcoming US team asks everyone to a lab supper: strawberries and pasta among the mass spectrometers. People trade precious not-yet-published catalytic methods, chomping pizza as complex automated valves click open and shut. The US Government funds a posh dinner in a football stadium. Months later, when my EU proposal, shaped over lunch in Boulder, goes to Brussels, the US offers warm support.

Sitting next to me most of the week is an EU friend. One day, amid the Melon-and-Sausage discussion, his laptop goes into meltdown after a barrage of e-mails. His paper on the 2003 European heatwave has just appeared in the journal Nature , and everyone from the BBC to Radio-Canada is trying to get hold of him. He shows me his lab's website. They have started "chemical weather" forecasts. Maybe they'll save a few lives.

At present, greenhouse emissions are self-declared. There's no independent atmospheric check. It's like income tax with no external inspector. In a few years, we should be able to verify Kyoto compliance by sniffing the air and assessing how much CO2 and methane each nation or region emits.

Moreover, as we begin to comprehend the Earth's breathing, we understand the biosphere's metabolism, trading carbon between plants and animals, forests and peat bogs, land and ocean. All this needs to be done from the ground - satellites help but painstaking near-surface measurement is essential. True understanding can come only when the un-loved intercomparison work succeeds in creating a global dataset.

We meet again in Helsinki in 2007. Yes, we try to keep our air travel to a minimum but these meetings are the UN at its best - maybe even the White House will listen.

É«ºÐÖ±²¥

Euan Nisbet is professor of geology at Royal Holloway, University of London.