Paleoclimate data identifies the impact that these missing slow feedbacks have in pushing temperatures higher than expected. New research matching greenhouse-gas levels with the Earth's temperature over the last 450,000 years has established the climate sensitivity with slow feedbacks to be 6 degrees. Fifty-five million years ago, in the Arctic, temperatures were 11 degrees warmer than the ECS models would predict - which also suggests that other feedback mechanisms were at work.

A paper published in the June 2005 issue of Nature supports the theory of even higher climate sensitivity. It describes research led by Meinrat Andreae of the Max Planck Inst.i.tute for Chemistry in Germany, which used climate models and various aerosol-cooling a.s.sumptions to find the 'best fit' for the data involved in a climate sensitivity in excess of 6 degrees. By studying the planet's climate history over the last 50 years and fitting it to various climate-model options, they concluded that the effects of airborne particle pollution (or aerosols: soot and exhausts from burning fossil fuels, industrial pollution, and dust storms) and climate sensitivity are both much higher than generally a.s.sumed. They say that greater pollution controls and 'clean air' legislation will remove much of the aerosol cooling, and that if carbon dioxide levels are double their pre-industrial levels by 2100, a rise of 6 degrees can be expected. When this understanding is combined with predictions that parts of the natural carbon cycle after 2050 will reverse from being net absorbers to net emitters of carbon, they say that warming by 2100 may be as high as 10 degrees.

These findings have enormous implications. A long-term climate sensitivity of 6 degrees would mean that we have already pa.s.sed the widely advocated 2-degree threshold of dangerous anthropogenic interference with the climate. It would, therefore, require us to find the means to engineer a rapid reduction of current atmospheric greenhouse gas even to restrict global warming to below 2 degrees - a target which we believe is, in any case, far too high.

A key question is whether the slow feedbacks have started to operate. In the case of the Greenland and Antarctic ice sheets, the data is already disturbing. One of the most important slow feedbacks to be considered is the reversing of the carbon cycle - as the oceans and soils take up less carbon dioxide - and the significant amounts of methane and carbon dioxide that are released by the permafrost.

Understanding how the carbon cycle works and how changes in the cycle will affect global warming are important in understanding the scale of action required to avoid catastrophic climate changes.

The carbon cycle is the flow and exchange of carbon in its various forms (including carbon dioxide, methane, and calcium carbonate) between the planet's four large, interconnected carbon reservoirs: the atmosphere (carbon dioxide); the oceans (carbon dioxide dissolved in seawater, carbon incorporated in living and non-living plants and animals, and methane trapped under pressure on the ocean floor); fossil organic carbon (coal, gas, and oil); and the land-surface biosphere (including soils, plants, and freshwater systems). Larger amounts are stored in the Earth's crust as rock carbonates, but these are relatively immobile.

The greatest carbon reservoir is the ocean, which contains about six times the amount of carbon that is stored in plants and soils. The fossil-fuels reservoir is of similar size to the land surface biosphere, while the atmospheric sink is the smallest.

Carbon flows between these reservoirs are driven by a variety of biological, physical, and chemical processes. Examples include extracting and burning fossil fuels; animal respiration; the exchange between the atmosphere and the oceans; drawing down carbon from the atmosphere by plant photosynthesis; and destroying forests by fire, land clearing, or decomposition.

Carbon reservoirs that absorb more carbon dioxide than they emit are called carbon sinks (as opposed to carbon sources, which emit more carbon dioxide than they absorb). The ocean is a carbon dioxide sink that responds rapidly to rising levels of atmospheric carbon, but not rapidly enough to meet the present need. The ocean water soaks up some of the additional carbon dioxide, and calcifying marine organisms absorb some of it (with subsequent burial in sea-floor sediments). Forests and gra.s.slands also absorb some carbon dioxide by photosynthesis. Much of the carbon dioxide, however, remains in the atmosphere.

Many sinks governed by living organisms become less effective as the environment heats up. Though it has long been expected that the capacity of the Earth's carbon-drawdown mechanisms would decrease due to human activity and as a consequence of higher temperatures, changes already observed suggest that this is happening earlier than antic.i.p.ated. The fraction of total human-caused carbon dioxide emissions that remain in the atmosphere has increased slowly with time - which implies a slight weakening of sinks, relative to emissions.

But some sinks may get to a point where they stop drawing down carbon and start emitting it instead. In 2000, a landmark study led by Peter c.o.x, then at the UK's Hadley Centre, found that about half the present emissions are being absorbed by the ocean and by land ecosystems. But this absorption is sensitive to the climate, as well as to atmospheric carbon dioxide concentrations. These two factors are creating a feedback loop, so that, under a 'business as usual' scenario, the terrestrial biosphere will only act as an overall carbon sink until about 2050, when it will fail and revert to being a carbon source. This slow feedback will increase temperatures by another 1.5 degrees by 2100.

Research published in October 2007 by Joseph Canadell, the executive director of the Global Carbon Project, confirmed that significant contributions to the growth of atmospheric carbon dioxide arise from the slow-down in the rate of absorption of natural sinks, or from 'a decrease in the planet's ability to absorb carbon emissions due to human activity'. According to Canadell: 'Fifty years ago, for every tonne of carbon dioxide emitted, 600kg were removed by land and ocean sinks. However, in 2006, only 550kg were removed per tonne and that amount is falling.' The data suggests that from 19592006 there was an implied decline of 10 per cent in the efficiency of natural sinks. Of the recent acceleration in the rise of atmospheric carbon dioxide levels, 18 per cent is attributed to the decreased efficiency of natural sinks.

Another key factor in this decreased efficiency has been identified by Peter c.o.x, now at Britain's Centre for Ecology and Hydrology in Dorset, who says that while plants are absorbing more carbon dioxide (because photosynthesis speeds up with warming), warming also encourages plant material in the soil to break down and release carbon dioxide. A lag between these events has seen the rise in carbon dioxide levels slowed for the last two decades; but science writer Fred Pearce says, 'Soon the biosphere will start to speed it up'. According to c.o.x, a possible surge of carbon dioxide into the atmosphere in 2003 is the first evidence of this process.

c.o.x spent years researching carbon cycles while at the Hadley Centre in Exeter, which has one of the world's most highly regarded climate-modelling systems. A summary of some of the centre's modelling work, published in 2005, included two startling graphs. In one, the amount of total carbon stored in the Amazon forest and soils shows a drop from around 70 billion tonnes of carbon in 2000 to just 20 billion tonnes of carbon by 2100. The second, using the same technique, compares vegetation and soil carbon levels in 2100 to those in 1850. While vegetation carbon had increased by about 60 billion tonnes of carbon by 2100, the amount of soil carbon had decreased by 130 billion tonnes.

The Amazon hosts a quarter of the world's species, and accounts for 15 per cent of land-based photosynthesis, as well as being an engine of regional and global atmospheric circulation and regional rainfall. Yadvinder Malhi of the Environmental Change Inst.i.tute in Oxford led a team that concluded that the Amazon is warming at 0.25 degrees per decade, a rate twenty-five times faster than the temperature increase at the end of last ice age. There has already been an observed drying. Periods of recent drought in parts of the Amazon have increased the frequency of forest fires. With a total bioma.s.s store of 120 billion tonnes of carbon and predictions of large-scale drought in the eastern Amazon, the release of stored carbon by wildfires would be catastrophic.

Professor Guy Kirk of Britain's National Soil Resources Inst.i.tute calculated that since 1978, the carbon lost by Britain's soil has increased by 13 million tonnes of carbon dioxide per year - more than the 12.7 million tonnes a year that Britain saved by cleaning up its industrial emissions as part of its commitment to the Kyoto Protocol. The loss is likely to be due to plant matter and soil organic material decomposing at a faster rate as temperatures rise - an effect that is expected to compound as temperatures increase. 'It's a feedback loop,' says Kirk. 'The warmer it gets, the faster it is happening.' It is thought that the terrestrial carbon sink will begin to convert to a carbon source at an increase of 23 degrees.

Bristol University researchers also argue that the previously unexplained surge of carbon dioxide levels in the atmosphere in recent years is due to more greenhouse gas escaping from trees, plants, and soils. Global warming is making vegetation less able to absorb the carbon pollution pumped out by human activity. Wolfgang Knorr believes that 'we could be seeing the carbon cycle feedback kicking in, which is good news for scientists because it shows our models are correct. But it's bad news for everybody else'. Another bad sign comes from Canada's Manitoba region, where a study of a one-million-square-kilometre area of northern boreal forest found that the area is now releasing more greenhouse gases than it absorbs, because of an increased incidence of forest fires. This is consistent with predictions that climate change, by producing hotter and drier conditions, would lead to more fires. 'Those wildfires have caused this transition in the boreal forest from a carbon sink to a carbon source ... Climate change is what's causing the fires; if it was left unchecked, it could become a feedback,' says Tom Gower of the University of Wisconsin. A further consequence of wildfires is that more sunlight reaches the ground. This increases the rate of decomposition of organic matter, releases more carbon dioxide and, perhaps, contributes to the melting of the underlying permafrost.

Burning rainforests are also emitting hundreds of millions of tonnes of carbon dioxide each year. During the 200506 Amazon drought, thousands of square kilometres of land burned for months, releasing more than 100 million tonnes of carbon. Philip Fearnside of the National Inst.i.tute for Research in the Amazon says that 'the threat of a "permanent El Nino" is to be taken very seriously ... Disintegration of the Amazon forest, with release of the carbon stocks in the bioma.s.s and soil, would be a significant factor in pushing us into a runaway greenhouse'. Daniel Nepstad, head of the Woods Hole Research Center's Amazon program, says: [It is] not out of the question to think that half of the basin will be either cleared or severely impoverished just 20 years from now ... The nightmare scenario is one where we have a 2005-like year that extended for a couple of years, coupled with a high deforestation where we get huge areas of burning, which would produce smoke that would further reduce rainfall, worsening the cycle. A situation like this is very possible. While some climate modellers point to the end of the century for such a scenario, our own field evidence coupled with aggregated modelling suggests there could be such a dieback within two decades.

In October 2007, there were more than 10,000 points of fire across the Amazon, most of them having been set by ranchers to clear land. 'These fires are the suicide note of mankind,' says Hylton Murray-Philipson of the London-based charity Rainforest Concern.

A survey on tipping points, led by Tim Lenton of the University of East Anglia and published in early 2008, found that leading researchers estimated that there was a medium risk that the Amazon would be largely destroyed by 2050. (Regarding other potential tipping points, they also estimated a medium risk of the Indian summer monsoon destabilising within one year; the West African monsoon collapsing in 10 years; and the Arctic boreal forest dying in 50 years.) Total carbon emissions from tropical deforestation are estimated at 1.5 billion tonnes of carbon a year, including illegal fires in Indonesia's vast peatlands, the haze from which regularly blankets Sumatra and Malaysia. Indonesia's peat swamps contain 21 per cent of the Earth's land-based carbon, and are now subject to increasing clearing, drying, and burning. During the 1997 El Nino event, an estimated 0.81 2.57 billion tonnes of carbon was released to the atmosphere as a result of burning peat and vegetation in Indonesia. This is equivalent to 1340 per cent of the mean annual global-carbon emissions from fossil fuels. This burning also contributed greatly to the largest annual increase in atmospheric carbon dioxide concentration ever detected since records began in 1957.

New a.n.a.lysis of two decades of data from more than 30 sites also indicates that the ability of forests in the frozen north to soak up man-made carbon dioxide is weakening.

The melting of permafrost (permanently frozen soil, or soil below the freezing point of water) is another 'slow' feedback that is adding to global warming. As the Arctic warms, permafrost in the northern boreal forests, and further north in the Arctic tundra, is starting to melt. As it melts, its thick layers of thawing peat trigger the release of methane and carbon dioxide, both greenhouse gases.

With less than 1 degree of warming, Arctic ground that has been frozen for 3000 years is melting and producing thermokarst (a land surface that forms as ice-rich permafrost melts). Even under scenarios of modest climate warming, this could affect 1030 per cent of Arctic lowland landscapes, and severely alter tundra ecosystems. As the permafrost thaws, lakes form and microbes convert the soil's organic matter into methane. The methane bubbles through the surface water into the atmosphere. In dry conditions, the warming soil also releases carbon dioxide.

A 2006 study found that Siberia's thawing wetlands are a significant, underestimated source of atmospheric methane, with lakes in the region growing in number and size, and emission rates appearing to be five times higher than previously estimated. The NCAR in Boulder predicts that half of the permafrost will thaw to a depth of 3 metres by 2050. As glaciologist Ted Scambos says: 'that's a serious runaway ... a catastrophe lies buried under the permafrost.'

The western Siberian peat bog is amongst the fastest-warming places on the planet, and Sergei Kirpotin of Tomsk State University calls the melting of frozen bogs an 'ecological landslide that is probably irreversible'. One estimate puts methane releases from the current area of melting bog at 100,000 tonnes per day.

Russian Arctic climate researcher Sergei Zimov frames the gravity of the situation well: 'Permafrost areas hold 500 billion tonnes of carbon, which can fast turn into greenhouse gases ... The deposits of organic matter in these soils are so gigantic that they dwarf global oil reserves ... If you don't stop emissions of greenhouse gases into the atmosphere ... the Kyoto Protocol will seem like childish prattle.'

The ocean carbon-cycle feedback is also a significant slow-feedback contributor. Part of the decline in sink capacity comes from a decrease of up to 30 per cent in the efficiency of the Southern Ocean sink over the last 20 years. This decrease has been attributed to the strengthening of the winds around Antarctica, which enhances ventilation of natural, carbon-rich deep waters. Lead author Corinne Le Quere of the University of East Anglia says: This is the first time that we've been able to say that climate change itself is responsible for the saturation of the Southern Ocean sink. This is serious. All climate models predict that this kind of 'feedback' will continue and intensify during this century. The Earth's carbon sinks - of which the Southern Ocean accounts for 15 per cent - absorb about half of all human carbon emissions. With the Southern Ocean reaching its saturation point, more carbon dioxide will stay in our atmosphere.

This finding follows pioneering work by CSIRO marine research scientists, including Stephen Rintoul and John Church, that seeks to understand how the Southern Ocean influences the climate system, its patterns of circulation, and the region's role in the global ocean-circulation system.

Measurements of the North Atlantic taken between the mid-1990s and 2005 found that, in the course of that decade, the amount of carbon dioxide in the water had reduced by half. It is suggested that warmer surface water was reducing the amount of carbon dioxide being carried down into the deep ocean. Lead researcher, Andrew Watson of the University of East Anglia, concludes: 'We suspect that it is climatically driven, that the sink is much more sensitive to changes in climate than we expected ... if you have a series of relatively warm winters, the ocean surface doesn't cool quite so much ... so the carbon dioxide is not being taken down into the deep water'. He warned that the process may fuel climate change: 'It will be a positive feedback, because if the oceans take up less carbon dioxide then carbon dioxide will go up faster in the atmosphere and that will increase the global warming.'

Satellite data gathered over the past ten years shows that the growth of marine algae, the basis of the entire ocean food chain, is being affected adversely by rising sea temperatures. Algae, the microscopic plants that permeate the oceans, remove up to 50 billion tonnes of carbon dioxide per year from the Earth's atmosphere. This system is as effective in removing carbon dioxide from the air as all plant life on the planet's land surface.

Jeff Polovina of Hawaii's National Marine Fisheries Service laboratory says that satellite imagery shows that green colouration (indicating chlorophyll life) in the middle of the ocean is fading away: 'The regions that are showing the lowest amount of plant life, which [are] sometimes referred to as the biological deserts of the ocean, are growing at roughly 1 to 4 per cent per year.' While such areas expanding are consistent with global warming scenarios, the rates of expansion already observed greatly exceed recent model predictions.

Increasing ocean acidification will also weaken marine life. This occurs as some of the carbon dioxide absorbed by the ocean reacts with water molecules to produce carbonic acid, which lowers the ocean's pH. The oceans are already 30 per cent more acidic than they were at the beginning of the Industrial Revolution, more than two centuries ago. If emissions continue at 'business as usual' rates, carbon dioxide levels in the oceans will rise so high that, by 2050, the ocean will be so acidic that current US water-quality standards would have to categorise it as industrial waste. Stanford University chemical oceanographer Ken Caldeira states that, if unabated, this could potentially cause the extinction of many marine species: 'What we're doing in the next decade will affect our oceans for millions of years ... carbon dioxide levels are going up extremely rapidly, and it's overwhelming our marine systems.'

Waters around the Great Barrier Reef are also acidifying at a higher-than-expected rate. Ecosystem collapse caused by acidification will likely reduce marine bioma.s.s and, therefore, the capacity of the oceans to absorb carbon dioxide. Professor Malcolm McCulloch of the Australian National University says that, contrary to previous predictions, this acidification is now taking place over decades, rather than centuries: '[T]he new data on the Great Barrier Reef suggests the effects are even greater than forecast.'

Acc.u.mulating evidence suggests that slow feedbacks from oceans, soils, and permafrost are already affecting the climate system.

CHAPTER 6.

Most Species, Most Ecosystems.

Martin Parry, co-chairman of one of the three IPCC working groups, told his audience at the launch of the full 2007 IPCC report on the impacts of global warming: 'We are all used to talking about these impacts coming in the lifetimes of our children and grandchildren. Now we know that it's us.' He said that destructive changes in temperature, rainfall, and agriculture were now forecast to occur several decades earlier than expected - and that means a huge threat to biodiversity.

As global temperatures rise, many species have to migrate towards the poles to stay in their habitable zones. If they can't migrate at sufficient speed, many species will be lost, and many ecosystems will degrade. During rapid change, such as the deglaciation and warming that occurred after the last ice age about 15,000 years ago, some widespread and dominant species became extinct when temperatures rose 5 degrees over a span of 5000 years. That is a rate of increase of 0.01 degrees per decade - 20 times slower than today's rate of change.

Cagan Sekercioglu from Stanford University says that the IPCC's worst-case scenario to 2100, combined with extensive habitat loss, would result in the extinction of around 30 per cent of land bird species. With warming, birds will try to move to higher alt.i.tudes. Once the top of a mountain is reached, there is nowhere left to go. In the lowland tropics, where most bird species live, there can be no significantly higher slopes to which they can retreat.

The rate of change in temperature is also very important in determining the impact it will have, because many ecosystems and species are sensitive to small temperature changes. A study by Rik Leemans and Bas Eickhout found that if a 2-degree impact builds up slowly over 1000 years, most affected ecosystems are likely to adapt (most often by moving); but if the same rise happens in 50 years (0.4 degrees per decade), many ecosystems will deteriorate rapidly.

At 0.4 degrees per decade, the isotherms (bands of equal temperatures) will be moving towards the poles at about 120 kilometres per decade; at this rate of temperature change, most ecosystems will be torn apart. Interestingly, Australia's birds are moving south at a rate of 100150 kilometres a decade with only half this rate of warming. Very fast-moving species will migrate with the temperature changes if they can survive in the ecosystems into which they move. Slow-moving species will not be able to keep up with the movement of their preferred temperature band and, unless they are tolerant of high temperatures and not dependent on species that have moved on, they will die out. At 0.4 degrees of change per decade, the isotherms are moving so fast that virtually all ecosystems will not be able to survive, and very large percentages of the dependent species will die out; yet this is the rate antic.i.p.ated in some of the IPCC scenarios by mid-century, and few scenarios antic.i.p.ate rates of less than 0.3 degrees per decade.

A 2007 study of the IPCC report's low- and high-emission scenarios, led by Dian Seidel of the NOAA in Washington, found that up to 39 and 48 per cent, respectively, of the Earth's terrestrial surface may experience novel and disappearing climates by 2100. Work published two years earlier projected the effects on 1350 European plant species under seven climate-change scenarios, and found that more than half could be vulnerable or threatened by 2080. The risk of extinction for European plants may be large, even in moderate scenarios of climate change.

Over the past 25 years, the area defined as 'climatologically tropical' has expanded to the north and south, away from the equator by about 2.5 degrees of lat.i.tude in each direction. This is equivalent to a rate of 110 kilometres per decade, and is greater than the IPCC's worst-case scenario of a total predicted shift of 2 degrees of lat.i.tude by 2100. This will disrupt the tropicaltemperate geographic transition of ecosystems and, if maintained over a century timescale, it suggests that few of the affected ecosystems would adapt at the implied warming of greater than 0.3 degrees per decade.

Seidel and her team also found that the expanding equatorial belt has 'potentially important implications for subtropical societies and may lead to profound changes to the global climate system'. They argue that the pole-ward movement of large-scale atmospheric circulation systems such as jet streams and storm tracks 'could result in shifts in precipitation patterns affecting natural ecosystems, agriculture and water resources'. Of particular concern to them are subtropical dry belts that could affect water supplies, agriculture, and ecosystems over vast areas of the Mediterranean, the south-western United States, northern Mexico, southern Australia, southern Africa, and parts of South America.

For wooded tundra, an average of 27 per cent of the ecosystem would remain in place for a warming of 3 degrees in 100 years - or 0.3 degrees per decade, over a century timescale. At that rate, IPCC lead authors Rik Leemans and Bas Eickhout found that 'only 30 per cent of all impacted ecosystems ... and only 17 per cent of all impacted forests' can adapt. If the rate were to exceed 0.4 degrees per decade, all ecosystems would be quickly degraded, opportunistic species would dominate, and the breakdown of biological material would lead to even greater emissions of carbon dioxide. This would, in turn, increase the rate of warming.

With emissions already tracking higher than the worst scenario of the IPCC, we must conclude that 'business as usual' would see the destruction or degradation of most species and most ecosystems by mid-century.

CHAPTER 7.

The Price of Reticence.

Roger Jones is a CSIRO princ.i.p.al research scientist. On 10 December 2007, in Melbourne's Herald Sun, he issued this call for scientists to overcome their aversion to risk taking: Often, scientists do not like to release their results until they are confident of the outcome. Important decisions need to be made now and cannot wait another five to seven years. Scientists will have to leave their comfort zone and communicate their findings on emerging risks, even when scientific confidence in those findings may be low... Sometimes, it is worth taking some risks in the short term to avoid worse risks down the track. We have spent too long being risk-averse about short-term costs and ignored the benefits of avoiding long-term damages.

If only the IPCC would adopt such an att.i.tude. Those turning to the 2007 IPCC reports for an up-to-date, authoritative view on global warming will find little of the real discussion of the events in the Arctic with which we started our story. The 2007 report is the IPCC's strongest call yet for governments, businesses, and communities to act immediately to reduce greenhouse emissions. But it is not enough, because it is based on outdated and incomplete data sets. The IPCC's four-year schedule for producing reports requires a submission deadline for scientific papers that is often two years, or more, before the report's final publication. What happens if there is significant new evidence, or dramatic events that change our understanding of the climate system, in the gap between the science reporting deadline and publication? They don't get a mention, which means that the IPCC report - widely viewed as the climate-change Bible - is behind the times even before it is released, though some new data is presented at forums.

On 28 January, just days before the release of the first of the IPCC's 2007 reports, the science editor of The Observer, Robin McKie, told of a serious disagreement between scientists over the report's contention that Antarctica will be largely unaffected by rising world temperatures: [M]any researchers believe it does not go far enough. In particular, they say it fails to stress that climate change is already having a severe impact on the continent and will continue to do so for the rest of century. At least a quarter of the sea-ice around Antarctica will disappear in that time, say the critics, though this forecast is not mentioned in the study. One expert denounced the [IPCC] report as 'misleading'. Another accused the panel of 'failing to give the right impression' about the impact that rising levels of carbon dioxide will have on Antarctica.

As McKie notes, the IPCC is, necessarily, a careful body.

Its reports involve the synthesis of many hundreds of pieces of research, and cooperation between many authors and contributors, such that only points that are considered indisputable by all of them are included: 'This consensus deflects potential accusations that the body might be exaggerating the threat to the planet. But the critics say it also means its doc.u.ments tend to err too much on the side of caution.'

Under intense pressure from global-warming deniers, the IPCC has adopted some methods that have gone beyond being 'careful' and are now simply conservative.

Fred Pearce, writing in New Scientist on 10 February 2007, tells of an IPCC review process that was 'so rigorous that research deemed controversial, not fully quantified or not yet incorporated into climate models was excluded'. Pearce wrote: 'The benefit - that there is now little room left for sceptics - comes at what many see as a dangerous cost: many legitimate findings have been frozen out.' After interviewing many of the scientists involved, he described the process as 'a complex mixture of scientific rigour and political expediency [that] resulted in many of the scientists' more scary scenarios for climate change - those they constantly discuss among themselves - being left on the cutting room floor'.

The peer-review process for experimental science is conservative, insisting on verifiable, reproducible results. Peer-review can significantly delay the full publication of new findings. When research produces a range of outcomes with differing probabilities or risks, there is a tendency for the general reader, and even policy-makers, to be drawn to the middle position - or even to the low end of the range, which requires less action.

Wider uncertainties in climate science and the vulnerabilities of species to fast rates of temperature change are good examples, because they drive us to consider the worst outcomes - not just the scenarios that have average effects. Some of the high-impact scenarios considered by the IPCC to be 'extreme' are now looking quite likely.

Barrie Pittock says that uncertainties in climate-change science are inevitably large, due to inadequate scientific understanding, and to uncertainties in human agency or behaviour. He says: [Policies] must be based on risk management, that is, on consideration of the probability times the magnitude of any deleterious outcomes for different scenarios of human behaviour. A responsible risk-management approach demands that scientists describe and warn about seemingly extreme or alarming possibilities, for any given scenario of human behaviour (such as greenhouse gas emissions), even if they appear to have a small probability of occurring.

This, he says, is recognised in military planning, and is commonplace in insurance; the lesson for climate policy is that the object of policy advice must be to avoid unacceptable outcomes, not to determine the most apparent, likely, or familiar outcome.

Michael Oppenheimer and three fellow scientists agree, arguing that the emphasis on consensus has put the spotlight on expected outcomes, which then become anch.o.r.ed via numerical estimates in the minds of policy-makers; however, with the general credibility of the science of climate change established, they say it is now equally important that policy-makers understand the more extreme outcomes that consensus may exclude or downplay.

In the case of the Arctic, for example, it is clear that this has not been done. James Hansen laments: For the last decade or longer, as it appeared that climate change may be underway in the Arctic, the question was repeatedly asked: 'Is the change in the Arctic a result of human-made climate forcings?' The scientific response was, if we might paraphrase, 'We are not sure, we are not sure, we are not sure ... Yup, there is climate change due to humans, and it is too late to prevent loss of all.' If this is the best that we can do as a scientific community, perhaps we should be farming or doing something else.

Pittock has described the limitations of the IPCC process: Vested interests harboured by countries heavily reliant on fossil fuels for industry and development, or for export, lead to pressure to remove worst-case estimates; scientists ... tend to focus on 'best estimates', which they consider most likely, rather than worst cases that may be serious but which have only a small probability of occurrence; many scientists prefer to focus on numerical results from models, and are uncomfortable with estimates based on known but presently unquantified mechanisms; and due to the long (four-year) process of several rounds of drafting and peer and government review, an early cut-off date is set for cited publications (often a year before the reports appear).

Inez Fung at the Berkeley Inst.i.tute of the Environment says that for her research to be considered in the 2007 IPCC report, she had to complete it by 2004. 'There is an awful lag in the IPCC process,' she says, also noting that the special report on emission scenarios was published in 2000, and the data it contains were probably collected in 1998. 'The projections in the 2007 IPCC report [using the 2000 emission scenarios] are conservative, and that's scary', she says.

There is a widespread view that the more extreme an outcome, in a range of possibilities, the less likely it is to occur. This can underestimate the role of feedbacks in a nonlinear world, and the evidence suggests that, in many cases, it is precisely the more extreme events that are coming true.

The data surveyed strongly suggests that, in many key areas, the IPCC process has been so deficient as to be an unreliable and, indeed, a misleading basis for policy-making. We need to look to processes that are not dogged by politics, and to a more up-to-date and relevant scientific knowledge base that integrates recent data and findings, expert comment, and the need to account for the most unacceptable, but scientifically conceivable, outcomes. On that basis we can build strategies that will at least give us a real chance to avoid the great dangers manifest in the climate system, of which humanity has become both master and victim.

The primary a.s.sumptions on which climate policy is based need to be interrogated. Take just one example: the most fundamental and widely supported tenet - that 3 degrees represents a reasonable maximum target if we are to avoid dangerous climate change - can no longer be defended. At less than a 1-degree rise, the Arctic sea-ice is headed for rapid disintegration; in all likelihood, triggering the irreversible loss of the Greenland ice sheet, catastrophic sea-level increases, and global warming from the albedo flip. Many species and ecosystems face extinction from the speed of shifting isotherms. Our carbon sinks are losing capacity, the seas are acidifying, and the tropical rainforests are fragile and vulnerable.

We have been lulled into a false sense of security by the stability of the climate during the Holocene period (the geological period that started 11,500 years ago, after the last glacial retreat, and which includes the whole period of human civilisation). Yet the period of ice ages and rapid deglaciations that occurred when the climate whipsawed between two states for millions of years is the usual mode. 'Abrupt change seems to be the norm, not the exception', says Will Steffen, head of the ANU's Fenner School of Environmental Science in Canberra. This is something we do not see, or do not want to see - and that incapacity means that, inevitably, abrupt changes, which our actions are now ensuring will occur, will be all the more devastating for our lack of foresight.

If we could start all over again, surely we'd say that we need to stabilise the climate at an equilibrium temperature that would ensure the continuity of the Arctic ice. This safe level has long since been pa.s.sed. We should have acted rapidly to restore and maintain the Arctic ice cap, with a safe margin for uncertainty and error, as soon as we knew there was a problem. But, given what has happened, what choices do we have now?

PART TWO.

Targets.

'We, the human species, are confronting a planetary emergency - a threat to the survival of our civilisation that is gathering ominous and destructive potential ... the Earth has a fever. And the fever is rising. The experts have told us it is not a pa.s.sing affliction that will heal by itself. We asked for a second opinion. And a third. And a fourth. And the consistent conclusion, restated with increasing alarm, is that something basic is wrong. We are what is wrong, and we must make it right.'

- Al Gore, n.o.bel Peace Prize acceptance speech, 11 December 2007.

CHAPTER 8.

What We Are Doing.

Something is wrong and we must make it right. This section explores the direction in which we must head to do so. How far must human greenhouse-gas emissions be reduced? What is a safe temperature zone? How do we get there?

In answering these questions, our first task is to understand what the greenhouse gases that we are pouring into the air have done, and what they are likely to do in the future.

In November 2006, The New Yorker reported on calculations by Ken Caldiera, from Stanford University, that 'a molecule of carbon dioxide generated by burning fossil fuels will, in the course of its lifetime in the atmosphere, trap a hundred thousand times more heat than was released in producing it'.

The quant.i.ty of carbon dioxide in the atmosphere and its persistence mean that it contributes more to global warming than any other product of human activity. Together with water vapour, methane, nitrous oxide, ozone, and trace gases, it maintains the Earth's greenhouse effect by trapping heat that radiates from the surface and, in doing so, keeps the surface temperature 33 degrees warmer than it would otherwise be.

Humans pour carbon dioxide into the air princ.i.p.ally by processing and burning fossils fuels (coal, gas, oil, and its derivatives), and through the burning and decay of large amounts of organic material (as a result of changing land-use patterns and de-afforestation).

Human activity has increased the level of carbon dioxide in the air by 38 per cent from the 1750 pre-industrial level of 280 parts per million: by 2008, it was at 387 parts per million. According to UNESCO, this is the highest carbon dioxide concentration recorded in the past 600,000 years and, probably, the highest in the past 20 million years. What's more, the rate of increase has been at least ten, and possibly a hundred, times faster than at any other time in the past 400,000 years. So our species is creating energy imbalances in the climate system that are pushing the rate of change far more rapidly than at any time since modern humans began to walk the planet.

When carbon dioxide is added to the atmosphere, oceans absorb some of it, vegetation (through photosynthesis) absorbs some, and some is trapped in sediments or by chemical reactions with eroding rock. The portion that remains in the atmosphere, however, is so stable and long-lived that it continues to produce its greenhouse effect for hundreds, even thousands, of years. It is generally understood that if we stopped adding carbon dioxide to the air, the carbon cycle would gradually draw down the amount of atmospheric carbon dioxide and, slowly, over time, the temperature would decrease; but this may not happen over the short time relevant to our current predicament.

New research presents a very sobering picture. Ken Caldeira and his Stanford University colleague Damon Matthews have used climate modelling to demonstrate that the portion of carbon dioxide that remains in the air produces a temperature increase that persists for many centuries. In the terms of their study, this means for at least 500 years - which was as far into the future as their model was run.

They showed that current human-related carbon emissions will produce a temperature rise of 0.8 degrees that will persist for more than 500 years. In plain language, the carbon dioxide that we emit will keep the planet heated for many centuries, and the more we emit the higher the temperature over that period will be. The unavoidable bottom line, according to Matthews and Caldeira, is that if we want to stabilise temperatures, we must eliminate all carbon dioxide emissions. They show that 'stabilizing global temperatures at presentday levels [which are 0.8 degrees above the pre-industrial level] required emissions to be reduced to near-zero within a decade [our emphasis].' This is an important result to which we will return later in the story.

Methane quant.i.ties in the atmosphere are also increasing. Since 1750, they have increased by 150 per cent, and about half a billion tonnes of methane are added each year, mostly as a consequence of human activity. When the full impact of methane is accounted for, its heating effect - including the results of its interaction with other gases to form ozone in the lower atmosphere - is at least half that of human carbon dioxide emissions. (At low levels in the atmosphere, ozone contributes to smog and is polluting. This is distinct from its role at levels in the upper atmosphere, where it creates the protective 'ozone layer'.) Drew Shindell of NASA's G.o.ddard Inst.i.tute for s.p.a.ce Studies estimates that methane may account for one-third of all climate warming from well-mixed greenhouse gases since the 1750s (carbon dioxide, methane, nitrous oxide, and halocarbons are known as well-mixed gases because their lifetime in the atmosphere is a decade or more): 'Control of methane emissions turns out to be a more powerful lever to control global warming than would be antic.i.p.ated,' Shindell concludes.