The UN Framework Convention on Climate Change, signed in Rio de Janeiro in 1992, poses a bold challenge to the world. It sets an ultimate objective to stabilise greenhouse gas concentrations "at a level that would prevent dangerous anthropogenic interference with the climate system."
Moreover, the Convention says, this objective "should be achieved within a time-frame sufficient to allow ecosystems to adapt naturally to climate change, to ensure that food production is not threatened and to enable economic development to proceed in a sustainable manner."
Twelve years on, agreement on what these words mean in practice is as elusive as ever. Defining a "dangerous" level is an inherently political decision - and one for which world leaders have so far shown little appetite.
The conference in February, arranged by the Environment Department (DEFRA) and held at the Met Office's headquarters in Exeter, was intended to serve two main functions.
The first was to bolster Prime Minister Tony Blair's plans to use the UK's G8 presidency to push for global agreement on controlling greenhouse gas emissions (see box ). Indeed, 2005 is shaping up to be a critical year in the drive to agree an international policy framework to succeed the existing Kyoto Protocol targets, which only run to 2012.
The second aim of the conference was to begin to address the thorny question of what might constitute "dangerous" climate change.
The conference, attended by over 200 experts from 30 countries, provided a timely update on the state of scientific knowledge. This was particularly important because of an unfortunate gap in the work of the Intergovernmental Panel on Climate Change (IPCC), the main forum for developing scientific consensus. The IPCC's last assessment report was published in 2001, and the next is not due until 2007.
In 2001, the IPCC concluded that there was strong evidence that human emissions of greenhouse gases are already changing the climate. Average global temperatures now stand at 0.6ºC above pre-industrial levels - and atmospheric concentrations of carbon dioxide have reached 379ppm, up from pre-industrial levels of 280ppm.
Moreover, the IPCC said, average global temperatures are likely to rise by 1.4-5.8ºC during this century. The wide range reflects different emissions scenarios and uncertainties over the sensitivity of the climate.
Scientific understanding has advanced considerably since the IPCC's report. Summarising the work presented at Exeter, the conference's steering committee reported "greater clarity and reduced uncertainty about the impacts of climate change". "In many cases the risks are more serious than previously thought," while "potentially disturbing" new impacts have come to light.
The steering committee points to growing evidence of "tipping points" beyond which there is a serious risk of large-scale, irreversible disruption to the climate. Events such as destabilisation of the Antarctic ice sheets and changes to the Gulf Stream become "more likely" once temperatures have risen by 3ºC above pre-industrial levels. Disruption to ecosystems and economies in developing countries also becomes highly significant above about 2ºC.
Many speakers at the conference offered support for limiting the average global temperature rise to 2ºC above pre-industrial levels. This figure has been endorsed by EU Environment Ministers and the European Commission. Support for the approach has also been voiced by Tony Blair (see p 35 ), BP chairman Lord Browne and by an international task force set up to advise governments on how to break the climate policy deadlock (see box ).
Achieving the 2ºC limit would represent a huge challenge - not least because we are already well on the way to meeting it. Last year, the International Energy Agency predicted that global CO2 emissions will increase by 63% between 2002 and 2030. The steering committee notes that "the world will, in the absence of urgent and strenuous mitigation actions in the next 20 years, almost certainly be committed to a temperature rise of between about 0.5ºC and 2ºC by 2050."
Acidifying the oceans
One of the most disturbing threats raised at the conference is not, in fact, driven by changes to the climate. The scientific community is increasingly alarmed that dissolved CO2 may directly alter the chemistry of the oceans - with drastic impacts on the entire marine food chain, and on the sea's capacity to remove CO2 from the atmosphere.
Carol Turley, head of science at Plymouth Marine Laboratory, told the conference that the oceans are an "enormous reservoir of carbon, greater than that on land or in the atmosphere." So far, the seas have taken up roughly half of the CO2 emitted by human activities.
However, the oceans' buffering capacity decreases as they take up CO2, which dissolves in seawater to form a weak acid. Dr Turley explained that the pH of surface waters has already reduced by 0.1 because of man-made CO2 emissions. A further reduction of up to 0.4 is expected by the end of the century even if atmospheric concentrations are kept to 560ppm - and pH could decrease by 0.77 if all known fossil fuel reserves are burnt.
"These geochemical changes are highly predictable," Dr Turley said. "Only the timescale is really under debate." Such dramatic changes have not occurred for more than 20 million years.
However, Dr Turley says, "there is surprisingly little research on the potential impact on marine and global ecosystems" - even though "marine productivity, biodiversity and biogeochemistry may change considerably."
A key concern is that reduced pH will reduce the availability of calcium carbonate, affecting the growth of calcifying species such as planktons and shellfish. Coral reefs - which are already expected to suffer badly from bleaching if global temperatures rise by 2-3ºC - will also be affected.
Dr Turley warned that acidification may also lead to "drastic changes" in the fixation of nitrogen and phosphorus - key nutrients for marine life.
So far, limited experiments have shown a range of adverse effects at atmospheric CO2 concentrations of 550ppm. Dr Turley says that "very little work has been carried out on whole ecosystems" - but warns that "the impact on the marine food web could determine whether we have cod or jellyfish in the North Sea."
Last year, a report for UNESCO's Intergovernmental Oceanographic Commission warned that the direct impact of CO2 on the oceans "has largely been neglected." However, the issue is shooting up the agenda - and in the spring the Royal Society will publish a major report on the issue.
A final - and highly significant - concern is that acidification will compromise the oceans' ability to act as a carbon sink by decreasing buffering capacity, reducing fixation by marine organisms and increasing stratification between ocean layers. The danger is that such "positive feedback" will increase the warming impact of future CO2 emissions.
Rising sea levels
The oceans also play a major role in two other worrying aspects of climate change - sea level rise, and the potential disruption of the Gulf Stream.
Over the last century, sea level rose by 1-2mm per year. The IPCC's 2001 report projected that mean sea levels are set to rise by 0.09-0.88 metres by 2100, driven largely by thermal expansion of the oceans and loss of mountain glaciers.
Dealing with this issue is particularly challenging because the huge thermal inertia of the oceans mean that sea levels will continue to rise for many centuries even once atmospheric temperatures are stabilised. Indeed, sea level will eventually rise by about 1 metre even if greenhouse gas concentrations were stabilised today.
Robert Nicholls of the University of Southampton stressed that well-planned adaptation strategies "can successfully manage many of the risks" for the next century at least. However, he suggested that this approach may have limitations in the longer term.
More rapid and dramatic sea level rise is a real possibility, however. The main concern is the potential melting of the Greenland and West Antarctic ice sheets which could increase sea level by a further seven and five metres respectively, albeit on a timescale of some centuries.
Bill Hare of the Potsdam Institute for Climate Impact Research told the conference that limiting warming to 3ºC above pre-industrial temperatures was expected to increase sea level by 2.7-5.1 metres by 2300. Of this, some 2-4 metres would be caused by melting of the Greenland and Antarctic ice sheets.
The slumbering giant stirs
Chris Rapley, Director of the British Antarctic Survey, told the conference that in 2001 the IPCC "characterised Antarctica as a slumbering giant in terms of climate change. I would say that this is now an awakened giant. There is real cause for concern."
Professor Rapley pointed to two recent findings. The first concerns the Antarctic Peninsula which could contribute perhaps 0.3 metres to global sea level rise. The peninsula is warming "strongly", and several of the surrounding ice shelves have disintegrated - most notably the Larsen B shelf in 2002.
The collapse of these ice shelves has, Professor Rapley said, been "like removing a cork from a bottle" - and has led to a "dramatic" acceleration and thinning of most of the Peninsula's glaciers. The concern is that loss of ice shelves elsewhere in Antarctica "could eventually cause accelerated and dramatic sea level rise".
The second new finding is that two major glaciers in the huge West Antarctic ice sheet have shown rapid thinning in recent years. It is not yet clear whether this is a result of natural instability or of climate change - but Professor Rapley fears the latter.
Warming of Greenland could also lead to a dramatic rise in global sea level. The conference heard that localised summer warming of 2.7ºC above present levels - associated with a global average rise of just 1.5ºC - is expected to trigger melting of the huge ice-cap. The issue was recognised in the IPCC's 2001 report but has only recently figured prominently on the climate policy agenda.
Jason Lowe of the Met Office's Hadley Centre told the conference that deglaciation of Greenland would probably be irreversible - and "may be triggered for even quite modest [temperature] stabilisation targets". Under a "pessimistic but plausible" scenario in which atmospheric CO2 concentrations rise to 1,100ppm, the ice sheet would melt almost completely in 3,000 years, with more than half melting in the first 1,000 years.
The Arctic ice-sheet is also under severe threat - although melting will have little impact on sea level because the ice is already floating on water.
Mark New of Oxford University presented the findings of a WWF-funded study showing that if the planet warms by an average of 2ºC - expected within a few decades on current trends - the Arctic will warm by 3.2-6.6ºC. The report warns that summer sea ice is melting at a rate of 9.6% per decade and is set to "disappear entirely by the end of the century" - with devastating effects on polar bears, other wildlife and the indigenous peoples of the region.
Gulf Stream on the edge?
Another issue which has attracted much attention - not least because of its starring role in the Hollywood epic The Day After Tomorrow - is the possible shut-down of the thermohaline circulation (THC).
This current, better known as the Gulf Stream, is responsible for maintaining Europe's relatively benign climate. The Hadley Centre says that abrupt shutdown of the THC could lower average annual temperatures in the UK by up to 5ºC "within a decade or two", with winter temperatures regularly falling below -10ºC. However, most models suggest that the cooling effect of a weakened THC would be more than offset by rising global temperatures.
In 2001, the IPCC concluded that a weakening of the THC could be expected by 2100, but that abrupt shutdown was unlikely. Beyond that date, complete and potentially irreversible shutdown was considered possible.
According to Mike Schlesinger of the University of Illinois, without action to reduce emissions "the likelihood of a THC collapse sometime over the next 200 years is more than two in three." Moreover, even "rigorous, immediate climate policy" would leave a 25% chance of THC collapse.
Peter Challenor of the UK's Tyndall Centre for Climate Change Research also presented preliminary results of a simplified model which suggest a 30% possibility of THC shutdown this century.
In contrast, Richard Wood of the Hadley Centre argued that the more complex global circulation models do not suggest a complete shutdown of the THC this century, although many show a significant weakening of the current. He told ENDS that "the conventional IPCC view, and mine still, is that abrupt shutdown is possible but of low probability."
Carbon cycle going backwards
Another potential "tipping point" in the climate system is the prospect that vegetation and soils could switch from being a net carbon sink to being a net source.
Peter Cox of the Centre for Ecology and Hydrology in Dorset explained that, at relatively low temperature increases, the sequestration of carbon is likely to increase because higher CO2 levels will enhance photosynthesis. But he warned that there is "a significant probability" that the land carbon cycle will become a net carbon source by 2100 under "business as usual" emissions scenarios.
Tropical forests, which currently cover 10% of the Earth's land surface, are an important carbon sink - despite the ongoing threat of deforestation - as they contain some 40% of the world's biomass.
However, Simon Lewis of Leeds University told the conference that tropical forests are under threat from drought and fire. Recent climate models suggest a "massive die-back" of the East Amazon forest later this century, he said. Tropical forests could become a major source of CO2 by 2100 - and may "greatly accelerate the rate of climate change over the coming decades".
Work at the Hadley Centre suggests that soils, currently an important carbon sink, are set to become an even more significant source than vegetation - and as rapidly as the middle of this century.
These positive feedbacks have profound implications for any strategy to reduce greenhouse gas emissions. Chris Jones of the Hadley Centre told the conference that "climate effects on the carbon cycle mean we may have to reduce emissions by around 20-30% more than previously expected."
The boiling frog
The focus on tipping points and on large-scale, high-impact events is crucial in defining dangerous climate change - but there is a risk that the numerous impacts caused by steadily increasing temperatures are overlooked.
The conference heard evidence that climate change threatens serious impacts on water resources, agriculture, and the spread of disease. Areas which are particularly vulnerable are Africa, central Asia and parts of China. One difficulty with all these issues is that society's ability to adapt is hard to predict, but is central to determining what level of climate change is "dangerous".
Summarising the evidence, Bill Hare of the Potsdam Institute said that food production in many developing countries is expected to suffer at all levels of warming. However, most developed countries benefit at temperature rises of up to 1ºC above pre-industrial levels. Above that point, "the risks begin to rise rapidly" - and become significant to severe in many parts of the world above 2-2.5ºC.
Water shortages could be a major threat, with perhaps 1.5 billion people being put at risk with a temperature rise of 1.5ºC. At higher temperatures, Dr Hare warns, a non-linear risk threshold may exist - posing a threat to megacities in India and China.
Ecosystems are also under growing stress - not least because fragmentation of habitats by human development will restrict the ability of many species to migrate.
John Lanchbery of the Royal Society for the Protection of Birds highlighted the large-scale breeding failure in seabird colonies around the North Sea in 2004. This appeared to be linked to a crash in the sandeel population caused in large part by a decline in plankton, the base of the marine food chain. Research suggested that a "regime shift" in plankton had been driven by increasing water temperatures - "almost certainly" a result of climate change.
Rik Leemans of Wageningen University in the Netherlands told the conference that "widespread ecological impacts of climate change are visible in every part of the world and in every ecosystem." He pointed to "numerous" recent studies which suggest that the IPCC's 2001 report underestimated the scale of future impacts - not least because ecosystem changes appear to be driven most strongly by changes in extreme weather rather than average conditions.
Professor Leemans called for a ceiling on global temperature rise of 1.5ºC - along with a limit on the rate of change. Mr Lanchbery argued that the inability of ecosystems to adapt meant that "the atmospheric concentrations of greenhouse gases can be considered to be already too high."
Crunching the numbers
So much for the impacts - but what are the implications for future emissions and the policy framework?
One key uncertainty is the sensitivity of the climate to increasing levels of greenhouse gases. The conventional measure is the expected warming arising from a doubling of CO2 concentrations - put at 1.5-4.5ºC by the IPCC.
However, a growing number of studies suggest that the climate could be considerably more sensitive. The most recent example is the climateprediction.net project, co-ordinated by Oxford University, in which 95,000 participants around the world helped to provide huge computing power for a climate simulation model.2The project confirmed that the most likely temperature increase for a doubling of CO2 is 2-4ºC - but also found a significant possibility that the response could be as much as 11ºC. Project leader Dave Stainforth told the conference that "we can't yet rule out the possibility of extreme responses...Even current levels could lead to dangerous climate change."
Malte Meinshausen of the Swiss Federal Institute of Technology set out the scale of the challenge involved in meeting the EU's target of limiting temperature rise to 2ºC above pre-industrial levels.
Stabilising greenhouse gas concentrations at the equivalent of 550ppm of CO2 would be "unlikely" to meet the target - and is equally likely to lead to an increase of more than 4.5ºC. Meeting this concentration would need a 10% cut in global emissions between 1990 and 2050.
Peaking at 475ppm followed by stabilisation at 400ppm CO2 equivalent is "likely" to keep temperature rise below 2ºC. Total greenhouse gas concentrations already stand at about 450ppm, although much of the warming impact is currently masked by the cooling effect of sulphate air pollution. This "global dimming" effect is expected to decrease over the next century.
Dr Meinshausen said that achieving stabilisation at 400ppm would require global greenhouse gas emissions to peak around 2015, and to be reduced by about 50% from 1990 levels by 2050. "A delay of just five years matters," he went on. "A delay of global action by 10 years nearly doubles the required reduction rates by around 2025."
A timebomb for policy-makers
The implications for greenhouse gas abatement policies are huge. The Government has won plaudits for adopting an aspirational target to reduce the UK's CO2 emissions by 60% by 2050 - a goal based upon stabilisation at 550ppm and allowance for increasing prosperity and emissions in the developing world.
In February, the left-leaning Institute of Public Policy Research - one of the key participants in an international task force on climate change (see box ) - concluded that the UK will need to reduce CO2 emissions by about 90% by 2050. The IPPR also called on the Government to match Germany's pledge of a 40% cut in CO2 emissions by 2020.
Prime Minister Tony Blair wants the UK's presidency of the G8 to deliver "a new global consensus" on climate change, with major nations "united in moving in the direction of greenhouse gas reductions". Mr Blair is treading carefully in the hope of luring the Bush administration back into the internationalist fold. The danger is that in doing so, he and other world leaders may be failing to acknowledge the daunting size of the mountain they are setting out to climb.