PART II: NOW

THE METHANE CATASTROPHE THAT MAY AWAIT US TOMORROW

What happened at the end of the Permian is long, long ago, but not far, far away. The catastrophic release of methane from the seafloor is not just something which was long ago. It could -- unless we change our way of dealing with our planet, and change it fast -- happen again tomorrow.

For most people, global warming -- the gradual and almost imperceptible increase in the world's average temperature -- is something that is only a dim and distant threat far in the future...if even then. The amount of global warming that most of us are likely to experience in our lifetimes is on the order of a degree or so. Celsius or Fahrenheit (it happens to be Celsius, in which scale each degree is 1.8 times that of a Fahrenheit degree): it hardly matters. Such small temperatures changes make no difference in the lives of most people. After all, the variability of temperature over the course of a day -- which from where I write is about 17°C (30°F) -- usually means nothing to us, and whether this spring happens to be a bit warmer or cooler than last year is typically quite unnoticeable. Moreover, in the affluent parts of the globe, it is easy to compensate for such minor temperature differences: with more or less clothing, with a bit more heat during a cold winter or a bit more air conditioning during a hot summer. Just change the setting on the thermostat.

The fact that sea level is rising (though not by much: during the 21st century sea level is expected to rise less than 40 centimeters/16 inches, about 3/4 due to the thermal expansion of water, and the rest due to glacier and icecap melting: Raper and Braithwaite, 2006; though a new report suggests that a rise of between 50 and 140 centimeters -- 20 to 56 inches -- may be more probable: Rahmstorf, 2007), and that some areas of the world will eventually be flooded, is known to the better-read members of the population. But even for them, warming-induced sea level changes seem remote. Yes, parts of New York City will be flooded in a few centuries, along with parts of many other coastal cities, but none living today will be around to witness such flooding. Parts of Bangladesh will go under water, and some island nations may disappear beneath the waves, but, again, these events will occur in a distant future, to other people, in faraway lands.

In addition to the indifference caused by our ability to control our immediate environments, many people have been apathetic about global warming because they have been told that there is considerable disagreement among scientists regarding whether there actually is warming, and, if so, exactly how much is caused by -- and therefore can be controlled by -- human activities. Unfortunately, these people have been misled, and, in most cases, deliberately misled. There are always disagreements among scientists, especially those who are conscientious in their endeavors. That is simply the way that science works. Disagreement helps determine truth, because it is disagreement that drives scientists to look for the evidence that will decide the issue. Nature -- the "real world" -- is always the decision maker.

Regarding global warming -- based on the evidence -- the vast majority of climate scientists are in agreement about three things:
(1) global warming is real, and will continue to increase,
(2) most global warming is due to human activity, specifically the burning of carbon fuels, particularly fossil fuels (carbon fuels include wood, peat, and charcoal in addition to the fossil fuels: oil, natural gas, and coal), and
(3) as global warming continues, it will have increasingly adverse effects on human beings, our environment, our fellow creatures, and the global economy.

It does not seem worthwhile to review all the enormous amount of evidence for global warming here. There are many fine books and a constant stream of scientific papers on the subject. Some of these papers occasionally attract enough attention that their findings make the daily newspapers. A few points, however, should be made.

First, the relation between atmospheric levels of the greenhouse gas carbon dioxide and global temperatures is clear. During the Phanerozoic (the last 543 million years), when atmospheric carbon dioxide levels have been high, global temperatures have been high; when low, global temperatures have been low. The only exception to this correlation was in the Late Ordovician (about 440 million years ago), when there was an ice age despite apparently high levels of carbon dioxide. Other suspected exceptions seem to have been based on the erroneous presumption that certain fossils that help provide ancient temperature information were not altered by conditions during and after their burial, when, in fact, they were (Pearson, 2001; Schwarzchild, 2001). The Ordovician exception has attracted further investigation. But in the 440 million years since that time, there are no other exceptions (for a review of their own evidence and that of others, see Royer, 2004). High levels of atmospheric carbon dioxide increase global temperatures (see Berner and Kothavala, 2001, diagram in previous section) because carbon dioxide absorbs reflected infrared radiation, and thus helps the Earth retain heat.

Second, the current warming is not, as has been alleged, just part of the natural variation. The current warming, rather, is quite exceptional. And third, it is due to human activity. Scientists, who generally are by training (if not by temperament as well) extremely cautious about reaching conclusions, especially in controversial matters, used to be very careful about attributing global warming to human activities. With mounting evidence, that caution has dissipated. In the past few years, the tone of scientific papers about climate change has shifted. Now scientists refer to anthropogenic (human-caused) global warming as a matter of course, and many seem quite (and properly) concerned that their repeated warnings and calls for remedial action are not being taken seriously enough.

A single graph provides a compelling answer to those who still disbelieve the connection between human activity atmospheric carbon dioxide and global temperatures. It traces the amount of carbon dioxide in the atmosphere over a period of 350,000 years, based on samples of ancient air bubbles caught in the ice at the Vostok Station in Antarctica, the most physically remote place on the surface of the planet. There, as snow slowly accumulated over the ages (Antarctica is technically a desert, because it receives very little precipitation), the snow also trapped tiny bubbles of air. From ice cores obtained by drilling through the Antarctic ice, scientists carefully extract that air, and measure the percentage of carbon dioxide, and the isotopes of oxygen. The oxygen isotopes can help reveal the temperatures of long ago. Thus, from the same tiny bubble, scientists can know both the amount of carbon dioxide in the air, as well as the temperature.

Atmospheric carbon dioxide levels and global temperatures over the past 350,000 years. (kyr BP means thousands of years before present.) The scale at the left refers to the green line, and indicates the level of carbon dioxide in the atmosphere. Note that it varies from about 200 to 300 parts per million by volume (ppmv). Temperatures are indicated by the blue scale at the right, which refers to the blue line. Temperature variation over the past 350,000 years is about 11°C. (Temperature here is measured by the Kelvin scale, in which each degree is exactly the equivalent of a degree Celsius, or C. Each degree Celsius equals 1.8 degrees Fahrenheit.) Notice how temperature closely varies with carbon dioxide level -- except at the extreme right. Here, for emphasis, the green line for carbon dioxide has been changed to red, to indicate the dramatic increase in atmospheric carbon dioxide -- in just two centuries. That red line, incidentally, has reached 380 ppmv as of mid-2005. (Rahmstorf, 2004)

The graph shows atmospheric carbon dioxide (CO¸2) and global temperature rise and fall in lock-step for these hundreds of thousands of years, right up until the present (the last hundred and fifty years or so). Then, a vertical red line, marked "anthropogenic CO¸2," abruptly -- and to those concerned about global warming, alarmingly -- spikes upward. The blue line designating temperature as yet records no change in response to this surge in atmospheric carbon dioxide (Rahmstorf, 2004, figure 1), but it will.

(According to newer Antarctic ice core studies, the same pattern holds back as far as 650,000 years. Though the level of atmospheric carbon dioxide reached about 300 ppmv twice at times other than the present during that 650,000 year period, it certainly never exceeded that level until the present, and is now at 380 ppmv: Siegenthaler, 2005; Brook, 2005.)

Though the red "anthropogenic CO¸2" line spikes straight up, this reflects the fact that the graph covers a period of more than a third of a million years; hence, the graph compresses all detail into a vertical direction. This is useful for comparing the rise and fall of carbon dioxide and global temperature over long expanses of time, and in putting the current spike in carbon dioxide in its proper long-term perspective, but it does not provide insight into the annual changes in carbon dioxide. A different graph, showing the increase in atmospheric carbon dioxide over the past several decades, furnishes those details.

Four things should be noted about this graph. First, it only records the changes in atmospheric carbon dioxide over the past few decades. Anthropogenic global warming has been happening at least since the beginning of the industrial age, about 200 years ago, but scientists only have precise information from about the past 40 years. Thus, while the graph begins at less than 320 ppmv (parts per million by volume), the actual pre-industrial average of atmospheric carbon dioxide is 220 ppmv, as averaged over 420,000 years (Falkowski, 2000). (The immediate pre-industrial level of carbon dioxide, from about 200 years ago, may have been about 280 ppmv. This value is commonly employed in climate assessments. Over the past 1000 years, this value has remained almost constant, with a variation of no more than 10 ppmv.) This means that the graph omits a major part of the anthropogenic carbon dioxide increase over the past 200 years, but simply because we were not engaged in the precision monitoring of carbon dioxide prior to about 40 years ago. As noted previously, atmospheric carbon dioxide has now reached 380 ppmv.

Second, the measurements for the graph were taken at the Mauna Loa Observatory on Hawaii (by the recently deceased Charles Keeling, who figured out how to measure atmospheric carbon dioxide and is one of the discoverers of global warming). The Observatory is at the summit of the volcanic peak, which rises almost 3500 meters (more than two miles) above sea level. Mauna Loa is an excellent place from which to monitor atmospheric concentrations of carbon dioxide, as it is far from the most significant sources of that gas; indeed, it is probably one of the best places in the northern hemisphere from which to conduct such measurements. (In fact, increased atmospheric carbon dioxide and global warming were first detected there.) Third, the sawtooth pattern simply reflects small seasonal variations in the general trend, and so are unimportant. Fourth -- and quite important -- is that the general trend of the line curves slightly upward with time (look at the line from its lower left corner). This means that the problem is getting worse from year to year, not getting better or merely staying the same.

(The increase in the level of atmospheric carbon dioxide may in part be accelerating due to global warming itself. In the summer of 2003, Europe experienced one of the greatest heat waves ever. Temperatures soared some 6°C [10.8°F] above the average recorded since 1851, and rainfall was down 50% from its long-term average [Baldocchi, 2005]. This caused a drop in primary productivity -- plant growth -- of some 30%, and a major increase in the release of carbon dioxide in the affected region [Ciais, 2005].)

An additional graph looks at the relation between atmospheric carbon dioxide and temperature for all ice-core air samples from which we have data on both. (The data used in this graph also comes from the ice cores of the Vostok station.)

Relation between carbon dioxide and global warming or cooling over the past 350,000 years. Each data point (from the Vostok, Antarctica, ice cores) shows both a carbon dioxide measurement and one for temperature (using oxygen isotope levels). The data points have been divided into two groups, depending on whether the data was taken from a time when glaciation was increasing (black points) or decreasing (gray points). During cooling periods (glaciations), both temperature and carbon dioxide drop; during warming periods (deglaciations), both temperature and carbon dioxide rise. The arrows indicate the general trends. (Falkowski, 2002)

Each point on this graph displays the information provided from one air sample, and each is color-coded according to whether it came from a period of glaciation (black) or deglaciation (gray) during the most recent part of the Ice Age. Unlike the previous graph, which traces carbon dioxide and temperature over time, this graph plots the level of carbon dioxide against temperature, displaying how they have varied in relation to one another. The arrows show what has happened: the Glaciations arrow indicates that the fall of atmospheric carbon dioxide correlates with the fall of global temperature. During Deglaciations, it is the opposite: the rise of carbon dioxide is correlated with the rise of global temperature.

There is a curious and particularly disturbing aspect to this graph. That is the direction of the arrow labeled "Modern." Although the arrows labeled Glaciation and Deglaciation point in opposite directions, they still define the same linear orientation, rather as a highway defines a particular route, even though it goes in two different directions. (Like the hands of a clock -- taking the top of the page as the 12 o'clock direction -- they point in about the 2 o'clock and 8 o'clock directions.) But the "Modern" arrow veers off on its own (in about the 12:30 direction). It does so because carbon dioxide is rapidly accumulating in the atmosphere.

Global temperatures, however, have not caught up. They will, and the angle at which the "Modern" arrow veers off will most likely, but gradually, come to define the same trend as do the Glaciations and Deglaciations arrows. (In other words, the arrow, like the hand of a clock, will slowly rotate clockwise from its current 12:30 position towards 2 o'clock as global temperatures increase.) Unfortunately, unless we severely curb emissions from the burning of fossil fuels, that arrow will reach to over 560 ppmv of carbon dioxide (double the pre-industrial level) in the atmosphere before the end of this century.

Extended to 560 ppmv in the direction currently indicated, the arrow projects global temperatures may reach about 6.5°C higher than at present. That projection is equivalent to some of the higher estimates of the amount the planet will warm by the year 2100, so it does not seem unreasonable, though many climate scientists believe that warming is likely to be less. However, if the direction of the "Modern" arrow does with time indeed slowly rotate clockwise, the planet will be warmer, perhaps considerably warmer, than those estimates now forecast.

Ocean acidification

In addition to warming the planet, increasing atmospheric carbon dioxide will have a further major planetary effect. Much of the fossil fuel carbon dioxide will enter the ocean from the atmosphere. There it will acidify the ocean. Excess carbon dioxide is now entering the ocean, from the atmosphere, at the appalling rate of about a million metric tons per hour, leading to decrease in ocean surface pH, that is, to an increase in acidity, of about 0.1 pH (Cicerone, 2004). A negative 0.1 pH change may not seem like much, but it represents a 30% increase in acidity. By the end of this century, pH will have dropped by another 0.2 to 0.4, possibly increasing acidity by more than 150%. (Negative pH changes represent exponential [geometric] increases in acidity rather than linear increases.) Organisms with calcium carbonate skeletons will be hard pressed to survive.

A detailed examination (Orr, 2005) of the likely effects of ocean acidification notes that not only will the world ocean become more acidic, but that the level of carbonate ions (CO¸3^2­) will also fall. Both of these effects -- the increase in acidity and the decrease in carbonate ions -- will be bad news for marine organisms with calcium carbonate skeletons, many of which provide shelter and/or food for organisms higher on the food chain, including human beings. Corals, for example, offer shelter for numerous invertebrates and fish, and are sometimes consumed by them. Pteropods, which are mollusks, and therefore the distant relatives of oysters, clams, octopi, and squids, are generally small, thin-shelled creatures that furnish food for many fish, such as salmon, mackerel, herring, and cod. They also are food for baleen whales (which strain water for food particles). In polar waters, pteropods may be almost as important as krill (shrimp-like crustaceans) in the diets of higher marine organisms (Orr, 2005).

Both corals and pteropods, however, construct their skeletons from aragonite, a less stable crystalline form of calcium carbonate (CaCO¸3). This particular mineral is more vulnerable to dissolution by acid than is calcite, the more stable crystalline form. As the world ocean acidifies, coral and pteropod skeletons will begin to dissolve, and growing corals and pteropods will find it increasingly difficult to construct their skeletons. This ecological threat will first appear in the world's coldest waters, because cold water can hold more dissolved carbon dioxide (just like cold soda, remember?)(Orr, 2005).

Two different scenarios for the coming increase in atmospheric carbon dioxide suggest that the survival prospects for cold water corals and pteropods in the 21st century is not good. According to one "business-as-usual" scenario (the IPCC's IS92a scenario) which postulates that human dumping of carbon dioxide will continuously increase, atmospheric carbon dioxide will rise to about 788 ppmv by the year 2100. The point of doubling beyond pre-industrial carbon dioxide levels (560 ppmv), therefore, will be reached around 2050. At this level of atmospheric carbon dioxide, the Southern Ocean's surface waters will be "undersaturated" with regard to dissolved aragonite. That means there will not be enough aragonite available in surface waters for those organisms which depend on it (Orr, 2005).

Organisms which construct their skeletons from carbonate minerals other than aragonite will also suffer. Coralline red algae and echinoderms (sea urchins, starfish), which make their skeletons from calcite that contains magnesium, will be affected even before corals and pteropods, because this form of calcite is even more vulnerable to dissolution than aragonite. Organisms which use relatively pure calcite, such as foraminifera and coccolithophores, will fare better, but not for long: calcite undersaturation will follow that of aragonite undersaturation by 50 to 100 years (Orr, 2005). The dissolution of coccolithophore skeletons (which do contain some magnesium) will cause not only the demise of the coccolithophores, but what could be the beginning of a significant drop in the level of atmospheric oxygen. Coccoliths (the shorter version of the name) are photosynthetic organisms which contribute a substantial amount of oxygen to the air.

Thus, by 2100, the situation will be far worse: pH will have dropped by another 0.3 to 0.4 units (relative to today's oceanic pH), representing a 100 to 150% increase in ocean acidity. At that point, the entire Southern Ocean (not just the surface waters), will be undersaturated with regard to aragonite. The sub-polar Pacific Ocean will be similarly affected (Orr, 2005).

A second scenario (the IPCC's "stabilization" scenario, S650) projects that acidification will take place more slowly. It assumes that atmospheric carbon dioxide will rise less rapidly, and level off at about 560 ppmv (twice pre-industrial) at about 2100. Under this scenario, the same conditions will be reached about 50 years later than with the "business-as-usual" scenario (Orr, 2005).

Previous studies had indicated that the process of increasing oceanic acidification would take centuries, but these researchers find serious acidification will take only decades (Orr, 2005). Disastrous as this probability is, however, it just glimpses our near-future prospects, to about the end of the twenty-first century. As the Caldeira and Wickett diagram (above) indicates, acidification of the ocean will not arbitrarily stop at the end of the present century, but will continue for many centuries more. And all of these scenarios focus on carbon dioxide, without attempting to include the possibly greater threat: seafloor methane.

METHANE RELEASE; METHANE CATASTROPHE

With the warming of the planet will come the release of hydrate methane. Already, the West Siberian Peat Bog, containing perhaps a quarter of the world's inventory of methane hydrate that is locked into continental permafrost, has begun to melt and is releasing some of that store of methane. The slight warming the Bering Sea and the northern coastal region of Alaska has undergone presumably has started to free some of its own hydrate methane. The current slow release, however, will certainly accelerate as increased temperatures heat high latitude continental regions, and warm ocean currents, disturbing marine circulation patterns.

Even if hydrate methane release never reaches a stage that can fully be described as catastrophic, it could make the difference between quite serious and quite deadly global warming. As time goes by, we can expect that methane will make more and more of a contribution to what may become not mere global warming but global scorching. But at some unpredictable "tipping point," the accelerating release may become overwhelming.

One way that such an accelerated release could occur is via a major submarine landslide, which can happen in a matter of hours, and which could trigger additional slumping and further sudden methane release. The stability of the continental margins in which methane hydrates reside is unclear: "It is not known if future warming is sufficient to cause failure of continental slopes on a global scale, but isotopic evidence of rapid carbon release in the past is suggestive" (Buffett and Archer, 2004).

Another release mechanism is by the venting of vast quantities of free methane and dissociated hydrate methane over a period of decades, perhaps triggered by marine warming sufficient to allow the fracturing of seafloor sediments by the free methane gas that underlies the hydrate. Either mechanism has the potential to abruptly and massively release continental margin methane within a short period of time, say, within a hundred years, or a few centuries at most. These are events that can take place in what is essentially a geological eyeblink, and will be beyond our ability to prevent (except by ceasing to dump carbon dioxide into the atmosphere), remedy, or mitigate.

At some point, that is, the release of methane may indeed become catastrophic, with a surge of this powerful greenhouse gas that could make the mere carbon dioxide warming of the planet pale to insignificance. Although some scientists (Harvey and Huang, 1995, for instance) believe that a powerful release of methane may await us, but probably no sooner than thousands of years in the future (if then), when ocean warming finally penetrates to what was at the time (1995) thought to be standard hydrate depth. Nonetheless, they recognized that such projections could be in error if more rapid release mechanisms prevailed.

University of Chicago geophysicists David Archer and Bruce Buffett, who also study what will happen with the immense store of hydrate methane in the seafloor, have estimated that some 85% of that inventory will be released with a 3°C (5.4°F) warming. Their calculations indicate that there are about 3000 billion metric tons (Gt) of hydrate and some 2000 billion metric tons (Gt) of "bubbles" -- the free gas that underlies the hydrate -- in the seafloor worldwide (Buffett and Archer, 2004).

Though this about half of the usual estimate of 10,000 billion metric tons (Gt) for seafloor methane, Archer notes that the oxygenation level of the ocean bottom is a significant factor in determining how much methane may be contained within the seafloor. If the deepest water of the Arctic Ocean is anoxic --as it could be, because of its restricted circulation -- the methane hydrate inventory would be greater (Archer, 2006). In addition, the Buffett and Archer calculations of the quantity of seafloor methane rely on measurements of carbon rain, the organic matter dropping through the water column to the ocean floor. The discovery of giant larvaceans, with their ability to send packages of organic matter rapidly to the ocean floor, has, according to the discoverers, been overlooked in previous estimates of carbon rain, and led to underestimates of 50 to 100% (Robison, 2005).

While Buffett and Archer project a loss of 85% of seafloor methane with a warming of 3°C, a more conservative estimate finds that about 2000 billion metric tons (Gt) of methane could be released with a seafloor temperature increase of 5°C (9°F) (Hornbach, 2004). (That is one-fifth of the usual estimate of 10,000 Gt of methane hydrate in the world's continental margins.) That's more than 2 1/2 times the amount of carbon as in the atmosphere. (Buffett and Archer's estimated release of 85% of their calculated quantity of seafloor methane amounts to 4 1/4 the amount of carbon in the atmosphere, or 8 1/2 times that carbon if seafloor methane is 10,000 Gt.) The carbon in the atmosphere, moreover, is largely in the form of carbon dioxide, which is a far less powerful greenhouse gas than methane. Though this methane would quickly be oxidized -- to carbon dioxide -- once it reached the atmosphere, even its short-term presence would deliver a stunning jolt of heat to the planet. The derivative carbon dioxide will maintain that heat over a much longer term.

At the end of the Permian, carbon dioxide was initially injected into the atmosphere by the Traps eruptions. These eruptions were presumably episodic, spread out over hundreds of thousands of years, interspersed with long periods of dormancy. Nonetheless, at least one or more of these episodes was sufficient to release vast quantities of seafloor methane. As noted previously, the primary trigger for seafloor methane release was presumably the direct heating of the PaleoArctic continental margin by Siberian Traps volcanism, particularly by the emplacement of magmatic sills within the seafloor sediments. No doubt, however, volcanic carbon dioxide also played a role in warming the planet and releasing hydrate methane.

But while not as dramatic, our own releases of carbon dioxide from the burning of fossil fuels well exceed, on average, those of the Siberian Traps. While not episodic, nor as sudden, our releases are considerable, continuous, and increasing. In no more than 300 years, virtually all of the accessible fossil fuel carbon reservoir -- some 5000 billion metric tons (Gt) -- will have been transferred to the atmosphere, in the form of carbon dioxide, if we continue to burn up our fossil fuels.

To put this carbon dioxide release into perspective, it is roughly equivalent to the total amount of carbon dioxide injected into the atmosphere by the Siberian Traps volcanism. Estimating from several sources, total Siberian Traps CO¸2 may have been about 1/2 to 2 1/2 to 4 times the amount of CO¸2 that will be released from the burning of fossil fuels (estimates based on Leavitt, 1982 and Gerlach and Graeber, 1985, as cited by Beerling and Berner, 2002, and Javoy and Michaud, 1989, as cited by Grard, 2005). Moreover, this colossal quantity of carbon dioxide will not be released over some 900,000 ± 800,000 years, as it was at the end of the Permian (Renne and Basu, 1991), but in the course of just three centuries. That amounts to an anthropogenic rate of carbon dioxide release that is roughly tens to perhaps thousands of times faster than the average rate of the end-Permian.

This enormous release of carbon dioxide ought to be quite sufficient to warm air and ocean enough to liberate a vast quantity of methane from its icy seafloor muds.

A massive methane release, and perhaps a true methane catastrophe, is just around the corner, geologically speaking -- as well as in human terms. Assuming we continue to conduct business as usual, it is inevitable. When it will happen cannot be predicted, but it will likely begin between about a hundred and, at most, a thousand years from now. Once it starts -- or even well before it starts -- it will be irreversible. Each of these characteristics -- inevitability, magnitude, unpredictability, and irreversibility -- requires further elaboration.

Inevitability

The release of seafloor methane is inevitable because we are pumping unprecedented quantities of carbon dioxide into the atmosphere. This carbon dioxide will warm the planet, and, in fact, is already doing so. Though the amount of global warming thus far (that is, in the twentieth century) is minimal -- only about 0.6°C (about 1°F), plus or minus 0.2°C -- the warming will significantly increase during this, the twenty-first century. The most generally accepted projections for global warming, those from the Third Assessment Report of the UN-sponsored Intergovernmental Panel on Climate Change (IPCC), indicates that the world's surface will warm by about between 1.4°C and 5.8°C (2.5°F to 10.4°F) by the end of this century (Kerr, 2001). As of the IPCC's 2007 Fourth Assessment Report, likely warming is projected at 1.1°C and 6.4°C (2.0°F to 11.5°F), with warming very likely between 1.8°C and 4.0°C (3.2°F to 7.2°F: IPCC 2007). Though this warming estimate represents the consensus thinking of the approximately 2500 climate scientists worldwide, recent warming estimates indicate that it may be too conservative.

Typically, climate scientists make their projections of global warming by estimating the heating effect of a doubling of atmospheric carbon dioxide, which has been expected to occur by the end of the twenty-first century. (The amount of warming that will take place as a result of a doubling of atmospheric carbon dioxide is often referred to as "climate sensitivity," though this is not the precise meaning of the term: see Schlesinger and Andronova, 2002.) Using a different and complex approach, some scientists now believe that there could actually be somewhat less warming than projected by the IPCC.

But these same scientists believe that there is an even greater likelihood -- in fact, a much greater likelihood -- that warming could considerably exceed the IPCC's projection. According to their projections, warming by the end of the century will likely range between 1.0°C and 9.3°C (1.8°F to 16.7°F), with the upper estimate significantly higher than that of the IPCC. The scientists who reached these conclusions find them -- in a modest departure from the ordinarily unemotional language of science -- "a disquieting result" (Andronova and Schlesinger, 2001).

In January 2005, this disquieting result was confirmed as a real possibility by the largest computer climate simulation ever done. Employing computer time from almost 100,000 home computers, the study compiled the results from the climateprediction.net experiment. (The harnessing of huge amounts of home computer time has become a standard activity in certain branches of science which require such time for extremely complex calculations. Ordinary citizens can make important contributions to science by the donation of such unused home computer time. This endeavor is immensely valuable for climate scientists, and readers are strongly encouraged to assist. This project does not interfere with the ordinary use of home computers. Details are furnished at the website, www.climateprediction.net)

With a doubling of atmospheric carbon dioxide, the study found, the possible global warming (that is, the "climate sensitivity") could range from 1.9°C (3.4°F) to as much as 11.5°C (20.7°F). Nonetheless, the study also found that the most likely temperature increase would be about 3.4°C (6.1°F), just as the IPCC had (Stainforth, 2005). One of the study's co-authors, Robert Spicer, pointed out that the highest temperatures in "recent" earth history occurred some 100 million years ago (during the Cretaceous Period), but that global temperatures at that time were probably only about 6°C (10.8°F) higher than today's (Connor, 2005). If the highest likely temperatures projected by the climateprediction.net study were to come to pass, they would be without precedent in hundreds of millions of years, perhaps for the entire Phanerozoic (Royer, 2004).

The disquieting findings of Andronova and Schlesinger were echoed by Richard Alley of Pennsylvania State University, Chair of the National Research Council's Committee on Abrupt Climate Change, at the December 2001 meeting of the American Geophysical Union. Alley, who was discussing his committee's newly released report "Abrupt Climate Changes: Inevitable Surprises," stated that significant global warming could come much more rapidly than the IPCC projects. He warned that global temperatures could rise 10°C (18°F) in just a short time, "tripping the switch" towards abrupt climate change in only a few decades (Showstack, 2001).

Nonetheless, the IPCC's 2007 Fourth Assessment Report pegs climate sensitivity (here, the doubling of atmospheric carbon dioxide) at between 2.0°C and 4.5°C (3.6°F to 8.1°F), with its best estimate being 3.0°C (5.4°F). This is not very different from another new estimate that puts climate sensitivity at between 1.5°C and 6.2°C (2.7°F to 11.2°F), with the most likely figure being 2.8°C (5.0°F). This estimate is based on a comprehensive study of carbon dioxide levels and presumed temperatures over the past 420 million years (Royer, 2007). However, because the weathering of silicate rocks draws down atmospheric carbon dioxide and restores temperatures some 100,000 to 150,000 years after a carbon dioxide outburst, this study may well overlook short, extreme spikes of temperature.

The warming that the planet has already experienced is not restricted to the lower atmosphere, or to the sea surface. Scientists used to be puzzled as to where the heat was going, because there seemed to be more heat being produced by global warming than could be accounted for, based on atmospheric and sea surface measurements. This is a conundrum no longer. As many scientists had previously suspected, the heat is going into the oceans. But they were only able to suspect that the oceans were taking up the heat because they lacked the ability to measure it. That has changed.

Based on the systematic investigation of millions of temperature records from various ocean depths worldwide, it is now clear that the "excess heat" is indeed going into the oceans. In fact, more than 90% of the heat from global warming has gone into the ocean, with the remaining heat having gone into the melting of polar region ice and mountain glaciers, and the atmosphere (Levitus, 2000). The oceanic temperature increase (0.06°C, or about 0.1°F) is minute -- only about a tenth of the temperature rise in the atmosphere -- but it represents an enormous amount of heat, because of the vast ability of the ocean to hold heat. The Atlantic, Indian, and Pacific Oceans all record the increase, and all indicate similar heat variations with time, over a forty year period from 1955 to 1995. All oceans show a similar increase trend (Levitus, 2000):

Ocean warming, 1955-1995. The heat is measured in joules (J), but the specific units are far less important than the general trends, obvious in all oceans. The red lines and the red figures indicate the approximate warming that has taken place, based on millions of measurements. (Levitus, 2000)

Perhaps most importantly, the warming has penetrated to deeper parts of all oceans, at depths from 300 to 1000 meters (about 900 feet to 6/10ths of a mile), and in the North Atlantic, even below the 1000 meter level. The total temperature increase of 0.06°C is the average of temperature readings down to 3000 meters (Levitus, 2000), emphasizing just how deeply the warming has penetrated. Such a warming of the deep ocean in such a short time was previously not thought possible. The North Atlantic data are possibly the most startling, because that ocean seems to be highly vulnerable to global warming, and most able to impact climatic conditions on its periphery and worldwide, because of its major role in driving global thermohaline circulation.

As a result of the warming, the ecology of the North Atlantic seafloor may be changing. The population of small (5 centimeters/2 inches long) marine creatures, the sea cucumbers (holothurians, a large group of echinoderms and thus the cousins of starfish, sand dollars, and sea urchins) has increased dramatically. Their numbers have jumped more than a thousand-fold since just 1996, an increase attributed to an influx of dead algae. Though the cause of the population surge has yet to be determined, climate change is a likely suspect (Krieger, 2004). Most likely, warming sea surface temperatures in the North Atlantic have resulted in a major die-off of plankton, furnishing the ocean-bottom-dwelling sea cucumbers with an unexpected -- and unsustainable -- bounty of food.

The salinity of the Atlantic Ocean is also changing (Curry, 2003). Over the four decades between the 1950s and the 1990s, water closer to the poles has become fresher, and that of the tropics has become saltier. The more poleward water has freshened because of increasing melting of the Greenland and Antarctic ice caps and Arctic and Southern Ocean sea ice. The increase in tropical water salinity is due to increased evaporation: an additional two meters (yards) of water evaporated over the four decades. This represents a five to ten percent rise in the evaporation rate in just 40 years (Curry, 2003).

Both the poleward and tropical changes are presumptively due to global warming. Indeed, between 50°S (about the latitude of southern Argentina) and 60°N (the latitude of the southern tip of Greenland), upper ocean temperatures in the western Atlantic (near the Americas) have risen about 1.0°C (1.8°F). An increase in equatorial precipitation near the African coast, possibly also due to global warming, has also been noted (Curry, 2003).

Hurricane Katrina and Global Warming (The Saffir-Simpson Hurricane Scale divides hurricanes into five categories, according to wind strength. A tropical storm becomes a hurricane when its winds exceed 119 kilometers/74 miles per hour. Category 1 hurricanes have winds from 119 to 153 kilometers/74-95 miles per hour; Category 2, 154-177 km/96-110 mph; Category 3, 178-209 km/111-130 mph; Category 4, 210-249 km/131-155 mph; Category 5, over 249 km/155 mph.) Hurricane Katrina as it approached the Gulf Coast on August 28, 2005 (NOAA). Hurricane Katrina crossed the southern tip of the Florida peninsula on August 25, as a Category 1 hurricane, and gathered strength to Category 5 during the next three days in the Gulf of Mexico. It then lost some of that strength, to Category 4, as it made landfall in Louisiana on August 29. Passage to the north over land quickly dissipated its strength, so that by August 30, it had become, once again, a tropical storm, just as it had been on August 24 before hitting Florida. Hurricane Katrina was a particularly nasty hurricane, which dropped enough water into Louisiana's Lake Pontchartrain to breach the levees surrounding the City of New Orleans. Inadequate maintenance of these levees, together with shoddy emergency preparations resulted in the flooding of about 80% of the city, and major destruction along the nearby Gulf Coast. The total cost of the hurricane is esttimated to be between $125 and $150 billion dollars, making it the costliest natural disaster in American history. Shortly after this colossal natural disaster, the Wall Street Journal ridiculed the idea that there was any connection between Hurricane Katrina and global warming. Let's examine the facts. First, there is no way to attribute any particular unusual weather event to a specific or single source. One need only to consider the accuracy of television and newspaper forecasts to recognize that if the causes and effects of particular weather events were well understood, we would never be caught without our umbrellas when it rains. The significant inaccuracy of such forecasts, as well as their inability to project more than a few days into the future, ought to convince us that scientific weather prediction accuracy is still an elusive goal (whatever the folks at the Old Farmers Almanac may think). General patterns, however, are often known to a fair degree of accuracy. A given area may typically have about two dozen days each year when temperatures exceed 38°C (about 100°F). Over many years, this pattern may hold. Nonetheless, there may be exceptional years when the number of hot days is considerably greater than two dozen, or, indeed, when there are no hot days at all. Careful study of weather patterns, together with mathematical modeling, can even give us some sense of how likely it may be for a given year to depart from the general pattern. Both trends and cycles may also alter an established pattern, and again, these departures from the initial pattern may suspected or predicted. One hurricane cycle has recently been identified. According to the US National Atmospheric and Oceanic Administration (NOAA), the number of hurricanes has been increasing as part of a natural cycle. The cycle seems to run for forty to sixty years; the number of hurricanes has been above average since about 1995. Indeed, as part of that cycle, the number of hurricanes forecast for 2005 is twice that of normal. So as many hurricanes may follow Katrina as preceded it, before the hurricane season ends on November 30. But even if the 2005 hurricane season is part of a natural cycle, it is characterized by a number of exceptional events. The year 2005 marks the first time there have been four named storms as early as July 9th since record keeping began in 1851. (Tropical storms get names when their wind speeds exceed 62 kilometers (39 miles) per hour.) The 26 named storms as of the end of the 2005 hurricane season exceed the previous record number of 22, and the naming system, for the first time, has had to extend into the use of Greek letters as hurricane designates. In addition, the year 2005 also marked the earliest Category 4 hurricane (Hurricane Dennis) on record. The 14 tropical storms whose wind speeds achieved hurricane status are also an annual record (the previous record was 12). Finally, Hurricane Wilma brought more than 162 centimeters (64 inches) of rain to Isla Mujeres, an all-time 24-hour rain record for the country of Mexico. (Though 2005 is likely to prove to be an exceptional year for hurricanes, it may nonetheless be part of a general trend to more numerous and ferocious hurricanes. The real significance of the 2005 hurricane season will only be clear from the perspective of a decade or more hence.) Is any of this unusual activity attributable to global warming? One of the predictions by those who study global warming is that the number and severity of unusual precipitation events will increase as the planet warms. Hurricanes are unusual precipitation events. Although it is too soon to assess whether the exceptional number of hurricanes predicted for 2005 and thereafter are truly part of a natural cycle, or whether they are some of the early effects of global warming, it may not be premature to suggest that the general severity of hurricanes is indeed increasing due to that warming. Hurricanes, after all, are colossal heat pumps, transporting vast quantities of tropical heat to more temperate regions. Intensity of precipitation events and temperature The data on this bar graph comes from 100 weather stations worldwide. Some stations (51 of them) are represented in blue; these are stations where temperatures range from ­3°C to 19°C (26.6°F to 66.2°F). Other stations (37 of them) have temperatures in the 19°C to 29°C (66.2°F to 84.2°F) range; these are shown in pink. The hottest stations (12), with temperatures between 29°C and 35°C (84.2°F to 95°F) are indicated in red. All stations have an average of about 230 millimeters (about nine inches) of seasonal precipitation. The daily precipitation at these stations has been divided according to amounts in tens of millimeters (each millimeter = 1/25th of an inch). Thus, the first set of bar graphs on the left shows how much of the total precipitation (of the 230 millimeters or nine inches) falls in events where the amount is about 10 mm (about 0.4 inch). For the coolest stations (blue), over 30% of total precipitation falls in such events. But for the warmest stations (red), less than 20% falls in these limited precipitation events. Now look at the right side of the graph. Here is the symbol >100, meaning that these are precipitation events where more than (>) 100 millimeters falls in a day. For the coolest stations (blue), only a percent or two of total precipitation falls in such events. But for the warmest stations (red), a much higher percentage (4-5%) of total precipitation falls in such precipitation events. The message of the graph is this: that in warmer conditions, even with the same total amount of precipitation falling, precipitation tends to be concentrated into bigger precipitation events. In a warmer world, therefore, we ought to expect more severe precipitation events, even if the total amount of precipitation remains the same. But, of course, a warmer atmosphere will hold more moisture, and the severity of extreme precipitation events will be even greater (Karl and Trenberth, 2003). A new computer model of projected extreme weather events for central North America during the next century seems to confirm these forecasts. Using five months worth of supercomputer time and enormous amounts of data, the resulting maps show in great detail the anticipated changes in extremely hot and cold events, as well as in extreme precipitation events. While the number of extremely cold events drops significantly, as would be expected in a warmer world, both the number of extremely hot events and the number of extremely wet events increases, in some cases quite significantly. For example, the desert Southwest will experience more than five times as many heat waves, and they will be of as much as five times longer duration than those of today. The northeastern US will experience heat waves typical of the summer's two hottest weeks -- but extending over two months. Even those parts of the US which usually have the fewest extremely hot events will have twice as many. The Gulf Coast, battered by hurricanes in the past few years, will not only have more heat waves, but also more rain and stronger precipitation events. Assuming that the model's projections can be extended to the nearby Gulf waters, that is a prescription for more, and more intense, hurricanes. Extreme weather event changes in central North Ameica over the century. The upper map shows projected changes in the number of extremely hot days. Some parts of the Southwest will have almost three months more heat waves than at present. The lower map indicates the projected increase in extreme precipitation events, as a fraction of today's extreme precipitation events. (Thus, 0.08 represents an 8% increase in extreme precipitation events; 0.16, a 16% increase, and 0.24. a 24% increase.) Worst hit will be the Gulf Coast (Diffenbaugh, 2005; Boutin, 2005). That the general severity of hurricanes may be increasing is the conclusion of a study published early the same month as Hurricane Katrina ravaged New Orleans. (Hurricane Katrina was not an exceptionally severe hurricane, though it was exceptionally large and struck in a particularly vulnerable and inadequately prepared area. However, a later hurricane of the 2005 season, Wilma, briefly became the most intense Atlantic hurricane ever recorded. It went from a tropical storm, with maximum winds of 119 kilometers/74 miles per hour to a category 5 hurricane, with winds over 249 km/155 mph, in a single day.) According to the study (Emanuel, 2005a), the severity of hurricanes, as measured by wind speeds and duration, has increased by some 50% over the course of the last 30 years. (This study found no evidence for an increase in the number of hurricanes.) Things could be worse: the same study indicates that storms in the northern Pacific have increased in strength by 75% during the same period of time. Though this study has its critics (see Landsea, 2005, and Pielke, 2005), its author stands by his conclusions, noting that they are based on about 100 times more data than those of his critics, rather than just hurricane wind speeds at landfall (Emanuel, 2005b). A second study confirms that the percentage of intense hurricanes has increased over the past 35 years, even while the total number of hurricanes, and their maximum wind speeds (at about 290 kilometers per hour/180 mph), has remained roughly the same. In those 35 years, the number of the most intense hurricanes (categories 4 and 5), has almost doubled, and that number, as a percentage of all hurricanes, has risen from about 20% to about 35%. The same period has seen a decline in the number and percentage of weaker (categories 1, 2, and 3) hurricanes (Webster, 2005).   Hurricane intensity, 1970-2004. This graph shows the changes in hurricane intensity for all oceans (in some parts of the world, hurricanes are known as cyclones or typhoons). The yearly data has been grouped into "pentads," that is, groups of five years. This is a statistical method used to reduce the influence of exceptional years on the results, and to make general trends more visible (Webster, 2005). Some scientists challenge these conclusions. But as scientists, it is their duty to demonstrate that the conclusions do not follow from the evidence, or that there is contravening evidence. At this point, the conclusions stand. And there may be good reason that they should. The increase in intensity of major storms over recent decades seems closely tied to the increase of ocean surface temperature during the same time. According to Kerry Emanuel, the lead author of the study, "'The total energy dissipated by hurricanes turns out to be well correlated with tropical sea surface temperatures. The large upswing in the past decade is unprecedented and probably reflects the effects of global warming'" (quoted in Verrengia, 2005). Storm intensity is measured by wind speeds and storm duration. [It should be noted that both the hurricane intensity study and Emanuel's comments date from several weeks before Hurricane Katrina struck.] The strong link between hurricane intensities and sea surface temperatures (as opposed to other possible influential factors) received additional confirmation in a study published in March of 2006. The link was found in all oceans that have such storms (Hoyos, 2006). Over the past 40 years, the upper ocean (down to about 300 meters/yards) has warmed -- as a result of the heating of the atmosphere by greenhouse gases -- by an average of 0.5°C (0.9°F: Barnett, 2005). This warming is present in all oceans to a greater or lesser degree, according to independent studies with millions of data points (Levitus, 2000; Levitus, 2001; Levitus, 2005). Warmer oceans contain more heat, which drives hurricanes; they also evaporate more water, adding to the severity of precipitation events. So there is a specific cause and effect relationship between oceanic warming and hurricane intensity. Consequently, it would be surprising if hurricane intensity had not increased as a result of oceanic warming. One would have to wonder why the predicted effect had not occurred when the causative agent was present. It would be rather like dropping a stone, only to have it float in the air. So, did global warming specifically contribute to the size and ferocity of Hurricane Katrina? Most likely, yes. Both the North Atlantic, in which Katrina formed, and the Gulf of Mexico, where it grew to a major hurricane, have been warmed by the action of anthropogenic greenhouse gases. The Gulf of Mexico, for example, was 2°C to 3°C (3.6°F to 5.4°F) warmer in early August than is usual for that time of year (Schiermeier, 2005c); seawater temperatures there were unprecedented. This allowed Katrina to suck "so much heat from the gulf that water temperatures dropped dramatically after it had passed, in some regions from 30°C to 26°C" [86°F to 78.8°F] (Schiermeier, 2005c). Both Katrina and the next major hurricane to hit the US mainland in the 2005 hurricane season, Rita, seem to have drawn their increased strength while passing over a "loop current" or its giant eddies in the Gulf of Mexico. This current carries warm water from the Caribbean through the Yucatan Channel (the passage between the Yucatan Peninsula and western Cuba), into the Gulf, and then out into the Atlantic via the Straits of Florida (between Florida and northern Cuba). The loop current and its eddies provide a deep (to water depths of 100 meters/yards or so) source of heat upon which passing tropical storms or weaker hurricanes can build into major hurricanes (Revkin, 2005). Sea Surface Temperatures on August 27, 2005, as Hurricane Katrina approached the Gulf Coast. Hurricanes require minimum temperatures of about 26°C (78.8°F) in order to form. Sea surface temperatures can only provide a general indication of hurricane building potential, however, because hurricanes draw their strength from waters as deep as many tens of meters/yards (NOAA). Deeper water temperatures and Katrina. Although rising sea surface temperatures are generally correlated with increased hurricane intensity, hurricanes draw their strength from more than just the surface of the ocean. Deeper ocean warmth is required. This satellite image shows why Katrina evolved into a major hurricane (and also why it lost strength just before reaching the Gulf Coast). Instead of showing sea surface temperatures, it displays the relative height of the ocean surface. Reds and yellows indicate a somewhat greater height to the sea surface, as measured in centimeters (2/5 of an inch). Because water expands as it warms, a higher sea surface indicates that the water below is generally warmer than in areas where the sea surface is lower (and the water below is therefore cooler). The image clearly shows the loop current which swings out of the Caribbean to the south, into the Gulf of Mexico west of Cuba, and then moves east between Cuba and Florida. A large blob of warm water, called a warm-core ring (WCR) has spun off the loop current. Katrina's path (from right to left in this image) took it over the loop current, and then over the warm-core ring, which enormously strengthened the storm's intensity. (TS = tropical depression; TS = tropical storm; please disregard the larger circles with crosshairs, which refer to another diagram. Scharroo, 2005). Without global warming, Katrina may still have formed and devastated the Gulf Coast, but its size and the force of its winds would likely have been reduced. The massive storm surge which swept over and obliterated several towns along the Mississippi coast would not have been as high as it was (about 10 meters/31 feet in places), and would have swept less far inland. The general evidence suggests that if Hurricane Katrina was typical of more recent hurricanes, its devastating force was some 50% greater as a consequence of global warming than it would have been otherwise. That's no small difference. The 2006 Hurricane Season: Despite the horrific hurricane seasons of 2004, when three major hurricanes hit the United States mainland, and that of 2005, when another three -- in addition to Katrina -- hit the mainland, the 2006 season was unexpectedly quiet. Not a single hurricane came ashore, and storm activity extended only as far westward as the central North Atlantic. What happened? One or both of two factors may have been involved. The first is called wind shear, and refers to the shearing force that is exerted when winds move at greatly varying speeds or directions at different altitudes. Because hurricanes are weather structures which extend from the ocean surface to at least sixteen kilometers (ten miles) high, wind shear has the ability to tear them apart. And, just as hurricane intensity is predicted to increase in a warming world, so also is wind shear. The same cause, therefore, can result in conflicting atmospheric phenomena. Nonetheless, according to predictive models, wind shear seems likely to increase at a slower rate with global warming than will hurricane intensity. Whether the predicted increase in wind shear will be sufficient to rip apart many future hurricanes, or reduce their intensity, remains to be seen. A second factor possibly impacting hurricane number and intensity may be dust from the Sahara Desert. Most North Atlantic hurricanes do begin their existences off the coast of northwestern Africa, as tropical low-pressure centers ("tropical depressions"). As they move westward across the North Atlantic, they can gain intensity from the warm water of the ocean. If the ocean water is cool, however, the conditions for increasing storm intensity are not met (Lau and Kim, 2007; Kerr, 2007). In 2006, unusually large quantities of dust from the Sahara were carried over the North Atlantic. This dust had the effect of blocking sunlight and reducing the solar heating of the ocean, dropping its temperature. In addition, the arid winds carrying the dust may have helped dry the marine air and increase wind shear. The 2006 season of low storm intensity may have been the consequence (Lau and Kim, 2007; Kerr, 2007). What will the future bring, as the competing influences of globally warming sea surface temperatures, wind shear, and dry, dusty winds vie for supremacy over the southeastern North Atlantic? Stay tuned.

As we have been repeatedly warned by geochemist and climatologist Wally Broecker (for example, Broecker, 2001), the increasing freshening of poleward waters, and the increasing salinity of tropical waters, can slow or shut down the Great Oceanic Conveyor (for a discussion of thermohaline circulation, see APPENDIX 3: THERMOHALINE CIRCULATION). There is, in fact, some evidence of just such a slowdown, in the subpolar North Atlantic during the 1990s (in contrast to the late 1970s and 1980s), though its cause is unclear. Lacking needed data from prior to 1978, when satellite monitoring of the oceans began, it cannot be determined if the slowing circulation is part of a normal decade-long cycle, or due to a warming of the water involved in that circulation (Häkkinen and Rhines, 2004).

But measuring the flow of the various Atlantic currents which contribute to the Great Ocean Conveyor has proven quite difficult. One study, relying on direct ocean measurements rather than those by satellite, found a significant (50%) reduction in the southward flow of North Atlantic Deep Water (NADW), a critical component of oceanic thermohaline circulation. Based on five oceanographic surveys made between 1957 and 2004 (1957, 1981, 1992, 1998, 2004), the study also found that half of the Gulf Stream's water failed to reach the Atlantic's far north (Bryden, 2005; Kerr, 2005).

Such surveys, using data from instruments which measure water density lowered from oceanographic ships, are nonetheless subject to quite high levels of uncertainty, and indeed, have failed to be confirmed. A new system of satellite-linked oceanographic buoys, strung across the mid-North Atlantic, now shows that marine circulation is much more variable than previously suspected. The new data suggest that it may be decades before a clear pattern emerges (Kerr, 2006). Nonetheless, an IPCC prediction indicates that the thermohaline conveyor may slow by as much as 25% by 2100 (Schiermeier, 2006).

It had been thought that a slowdown or shutdown would have enormous consequences for human populations living on both sides of the North Atlantic through the cooling that would ensue when warmer waters did not reach the far north (Broecker, 2001). Though such cooling could not induce a new ice age (though some have mistakenly thought so), it could nonetheless be precipitous. In particular, it was noted that in one Ice Age episode, Greenland temperatures fell by 10°C (18°F) in just a decade (Kerr, 2004). But now that we have a better understanding of the rapidity with which global warming will overtake us, the North Atlantic deep freeze scenario has now been dismissed as unlikely. It will not be cold about which those in northern Europe, eastern Canada, and the northeastern United States will have to be concerned, but warming itself, just like everyone else.

As the planet continues to warm, so will the oceans. The Southern Ocean, at mid-depths, has warmed by 0.17°C (about 0.3°F) since the 1950s (Gille, 2002). New data from the North Pacific confirms that even at great depths, 5000 meters (three miles) and below, seawater temperature has risen by 0.005°C (0.009°F). This is a tiny change, but it occurs where there should be no change at all. And it occurs across the entire North Pacific, a distance of many thousands of kilometers (several thousand miles), from off Washington state to Japan, as surveyed mostly along latitude 47°N. (The distance is approximately one-quarter of the circumference of the globe at the latitude surveyed.) And the warming occurred in just 14 years, between 1985 and 1999 (Fukasawa, 2004).

There should be no change in deep-ocean temperatures because the deep ocean is quite isolated from the upper ocean, at least on short time scales. This deep ocean water should not have been in contact with the surface in 800 years (Davidson, 2004). Textbook time for ocean mixing -- the amount of time it takes for a mass of ocean water to blend with other ocean water -- is 1000 years. This slow rate, and the isolation of deep water from the surface, means that there should have been no warming whatsoever in the period studied. The fact that there has been such warming is a matter of surprise and concern -- and even alarm -- to oceanographers.

The oceans are the roach motel of global warming. In the case of the oceans, it is heat which checks in but doesn't check out. Water has an extraordinary heat capacity. In other words, it holds heat better than almost anything else. That means that when the oceans warm, they lose that heat only slowly and reluctantly. Ocean heat stays around for a long time. And most of the heat from global warming -- over 90% -- goes into the oceans.

And the oceans hold 99% of the world's supply of methane hydrate; the rest is in permafrost. Inevitably, in a warming world, hydrate methane will also be released from permafrost, but its quantity pales to insignificance compared to that in the oceans. Moreover, it is not merely hydrate methane that will be released, but also free methane from below the BSR.

The New BSR

It used to be thought that the depth of the methane hydrates within the ocean floor sediments would delay and perhaps prevent methane release with global warming. This thinking was based on the amount of time it would take for warmth from the overlying ocean to penetrate the sediments, and upon scientific understanding of the physical contours of the methane hydrate deposits. The bottom simulating reflectors (BSRs) that marked the boundary between the overlying hydrate and the underlying free methane closely mimicked the contours of the ocean floor (hence their name), though generally several hundred meters deeper.

These contours were fairly smooth and gently rolling (see the sonar image of the "old" BSR, above, in the Methane and Methane Hydrates, Part 2, section), and the upper boundary of the hydrate was similarly presumed to be relatively smooth. Because this hydrate methane was thought to be located well below the seafloor, it was assumed that heat from global warming would take a long time to penetrate that far, and that any significant release of margin methane would take place at least hundreds (Kvenvolden,1988a), and perhaps thousands of years in the future. Even employing "worst case estimates," little methane release was presumed to be possible, because about 98% of methane hydrate is found in sediment conditions which would require a 4°C (7.2°F) warming to dissociate (Harvey and Huang, 1995).

Such findings could provide consolation to those concerned about the possibility of a near-term methane catastrophe, were it not for the qualifying statement heading a list of conclusions: "In the absence of fracturing or sediment failure..." (Harvey and Huang, 1995). If fracturing or sediment failure were indeed possible, the comforting conclusions would not be valid. In fact, however, fracturing and slumping (sediment failure) turn out to be the major modes by which large quantities of hydrate methane can be released.

New, more detailed sonar images have completely changed our understanding of BSR topography. Instead of being fairly smooth and gently rolling, the BSR surface is now known to be punctuated with sharp, needle-like peaks, columns, and discontinuous knife-edge ridges, which may extend all the way up to the top of the sediment. These features apparently represent escape routes for free and dissociated methane, which at times follows "chimneys" (the needle-like peaks and columns), and at others follows faults (the knife-edge ridges) through the consolidated sediments (Wood, 2002; Pecher, 2002). [Sonar image from Wood, 2002.]

These escape routes are apparently in regular use. If they were not, we would expect that the free gas below the hydrate would gradually accumulate and build up pressure. Eventually, the pressure would lead to a blowout, and explosive release of the gas from its trap under the hydrate. Though such blowouts may occur occasionally, they are probably not common, because the faults and chimneys allow methane to escape from below the hydrate. The escape routes serve as safety valves.

A careful examination of the pressure of the free gas below the hydrate on several passive margins shows it to be essentially identical to the pressure needed to pry open overlying faults (Hornbach, 2004). This pressure is referred to as the critical pressure, and the amount of free methane under the hydrate is, in passive basins, at critical pressure. Free gas tends to remain at critical pressure because it forces open the safety valve when the pressure of the gas exceeds the critical pressure. When the free gas is at less than the critical pressure, the safety valve remains closed.

Seafloor faults thus seem to be highly responsive as pressure safety valves. In those active margins where oceanic plates are subducting beneath continents, the pressure being applied to the wedge of sediments piled against the continent (the accretionary wedge) forces fluids up through the hydrates. This process, called hydraulic fracturing, creates temporary pathways through the sediments, and allows free and dissociating methane to escape (Zühlsdorff and Spieß, 2004).

Because many active continental margins are being compressed by the forcing together of tectonic plates, it is not surprising that faulting and fracturing should be common there. On passive margins, by contrast, faulting should be much less. Therefore, there should be less free gas, and less pressure, under hydrate on active margins than on passive ones. And, indeed, that is the case: on the Blake Ridge off the Florida-Carolina coast (a passive margin), the free gas column is 200 to 250 meters (yards) thick. Other passive basins display similar free gas column heights. But the free gas columns in active margins are much less: typically about 35 meters thick, they rarely exceed 50 meters (Hornbach, 2004).

Sonar images of an active margin area, off the west coast of Canada's Vancouver Island, reveal the same sort of wipeout zones (transparent to sonar: here they are referred to as "blanking zones"; Zühlsdorff and Spieß, 2004), as reported elsewhere (Wood, 2002). The seafloor surface exhibits a pockmark, a familiar sign of fluid venting, with massive gas hydrate lying at only 3 to 8 meters (yards) depth (Riedel, 2002). It is presumably the process of hydraulic fracturing which allows a rather free flow of methane out of the hydrate zone, and prevents the buildup to critical pressures found in passive margins and basins. The process, in fact, may be important for methane release in all margin settings (Zühlsdorff and Spieß, 2004).

Thus free methane, as well as hydrate, extends in places to, or almost to, the seafloor, where it can be -- and presumably is -- released into the ocean. This means that at least part of the oceanic methane reservoir is much more accessible to warming than previously thought. The gas chimneys and faults that serve as escape routes for methane are therefore likely to be highly responsive to the oceanic warming that is accompanying global warming generally. Consequently, assumptions about the relative remoteness and inaccessability of oceanic hydrates will have to be scrapped. Clearly, methane hydrates and the free methane below them are considerably more vulnerable to warming than was previously presumed. Most dismayingly, there may be little lag time -- or even no lag time whatsoever -- between ocean bottom warming, and the initiation of seafloor methane release in response.

Inevitably, then, with the amount of global warming projected to occur in the next century or so, most or all seafloor methane will be released. Exactly how fast that release will take -- and how catastrophic that release will therefore be -- is something that can at best be only dimly glimpsed. Clearly there are some critical parameters which are unknowable or must be guessed at.

Magnitude

Initially, methane from hydrate will slowly trickle out of the sediments as the oceans warm. In fact, some undoubtedly is trickling out right now, contributing (along with other sources such as the increasing number of ruminants and increasing rice cultivation) to the slow rise of methane detected in the atmosphere. At some point, however, the gradual mode of methane release is likely to shift to a pattern of more abrupt, episodic releases, as oceanic warmth penetrates more deeply into the sediments. There will be rapid depressurization of hydrate at the base of the hydrate stability zone because of the release of free gas through warmed chimneys, leading to hydrate dissociation and release. Or there will be a submarine landslide, triggered by the melting of the hydrate and the consequent destabilization of deep sediment, or by an earthquake once the hydrate has been brought to the point of destabilization by the warming.

(A triggering earthquake could be just an ordinary quake, which are common on active continental margins -- hence their description as "active" -- or could actually be the result of the warming, which would increase the weight of the overlying water on continental slopes. Warming increases the water weight on the continental margins, including the slopes, because it causes the thermal expansion of water. Thermal expansion means that the volume of the water increases, but not its total weight, which remains the same. But since this increased volume is proportionately greater over the shallower portions of the ocean, the weight of water there increases, while the weight decreases in the deep ocean where the volume increase is proportionately smaller. See the diagram of the thermal expansion of water in the Methane and Methane Hydrates section.)

During the initial phases of more episodic releases, the rate of methane release will sharply increase. If the release is caused by a submarine landslide, most of the associated methane will be released in less than a day. If close to shore, the slide may produce a significant tsunami.

Tsunamis

A tsunami is just another unpleasant possible effect of a submarine landslide. In 1998, a 7.1 earthquake caused about 4 cubic kilometers (a cubic mile) of sediment to slide down a 25° seafloor slope a short distance offshore from the southwestern Pacific island of Papua-New Guinea. A 7 to 10 meter (yard) high tsunami (perhaps as high as 15 meters/50 feet in the area hardest hit) inundated the shore just moments later, sweeping away several villages and over 2000 coastal inhabitants (Chang, 2002).

Tsunamis caused by submarine landslides are not uncommon. (The December 2004 tsunami which devastated countries around the Indian Ocean was not caused by a submarine landslide, but by a magnitude 9.0 earthquake off the Indonesian island of Sumatra. The earthquake caused a huge movement of the seafloor, and it was this movement which produced the tsunami.) Usually their effects are confined locally, though this depends on the magnitude of the slide and its proximity to the coast, among other factors. Larger landslides do cause bigger tsunamis, other things being equal. And depending on their size, tsunamis may have regional or even hemispheric rather than merely local effects.

In November 1929, a 7.2 magnitude earthquake south of the coast of Newfoundland caused a significant undersea slump which cut a dozen transatlantic telecommunication cables from North America to Europe. It is estimated that the slump carried between 300 and 700 cubic kilometers (about 70 to 170 cubic miles) of sediment.

Originating on the continental slope, the Grand Banks slide tore asunder some half dozen cables, and the turbidity current -- a slurry of seawater and sediment -- that it engendered ripped apart an additional six. Every cable was broken in at least two places more than a hundred and sixty-five kilometers (a hundred miles) apart, indicating both the great width of the slide and turbidity current, and its speed, estimated at its origin as eighty to one hundred kilometers (about fifty to sixty miles) an hour. Despite continuously depositing its load of sand, mud and silt, the turbidity current still retained enough force to sever a final cable over 800 km (500 miles) seaward (Heezen, 1952).

The height of the resulting tsunami was 7 meters (22 feet). But bays and harbors, because of their constricted shape, tend to channel tsunamis to more destructive heights as they funnel inland. This fact is reflected in the name "tsunami," a Japanese word composed of "tsu" meaning harbor, and "nami" meaning wave. In the case of the Newfoundland tsunami, the channeling effect caused wave heights to run up to as high as 13 meters (40 feet) in some bays, destroying fishing vessels and harbor buildings and killing about 28 people (Ruffman, 2001).

Though the Grand Banks slide was probably the result of the earthquake shaking of waterlogged sediments, some submarine landslides very likely or certainly have involved free methane gas and methane hydrate, as did the Storegga slide mentioned previously. The East Coast of the United States holds its own methane-related hazards, and as scientists have come to know the marine world better, these hazards have become clearer. The relic of a major slide, roughly equivalent in volume to that of the 1929 slide off Newfoundland, has been found off the coast of North Carolina. Estimated to have taken place about 20,000 years ago, the Albemarle-Currituck slide is of ice age vintage, though it presumably was not triggered either by the weight of the glaciation or by glacial rebound, as there was no continental ice sheet within hundreds of miles.

Recently marine geologists discovered cracks on the outer edge of the continental shelf (at the "shelf-slope break") along the Virginia-North Carolina coast to the north of the Albemarle-Currituck slide. Similar cracks have also been found on the edge of the continental shelf off New Jersey. On the North Carolina-Virginia shelf edge, these cracks have caused the shelf edge to slump down as much as 50 meters (160 feet)(Driscoll, 2000). On closer inspection, however, the cracks have proven not to be simple cracks at all, but elongated craters as much as two by five kilometers (about 1.2 by 3 miles) in extent (Simpson, 2000).

These craters may be evidence of the rapid, or even explosive, expulsion of methane-laden fluids from the upper slope sediments, and could contribute to slope failure, resulting in major submarine landslides. Such landslides could result in significant tsunamis along the central section of the East Coast, the discoverers of the Virginia-North Carolina shelf edge cracks have warned (Driscoll, 2000). A rapid slope failure similar in volume to that of the Albemarle-Currituck slide could set loose a tsunami up to several meters (yards) high, equivalent to storm surges from major hurricanes.

As with major storm surges, the actual devastation that an East Coast tsunami would cause would depend on the topography of the coast where it hit, together with the height of the tide at the time. (The northern part of the American East Coast may be particularly vulnerable because it has numerous estuaries -- Chesapeake Bay, Delaware Bay, New York Harbor -- which can channel water to major low-lying population centers.) But in important respects tsunamis differ from the water rise of a storm surge. First and perhaps most important, there would not be the lengthy warning that accompanies the approach of a hurricane.

Second, tsunamis are often preceded, by several minutes to about a half hour, by an actual drawdown of sea level of a few to several meters (yards). This surprising precursor to a tsunami caused several deaths when it occurred along the French Riviera in response to a submarine earthquake about thirty years ago. Curious but unsuspecting bathers, lured by the withdrawal of the Mediterranean and the exposure of sea bottom normally unseen even at the lowest of tides, wandered out far from shore. There they were caught when the actual tsunami waves, several meters high, came rolling in a short time later. This third and last difference between tsunamis and storm surges, that of the rapid buildup and breaking of walls of water against the coast, rather than the more gradual rise that accompanies hurricanes, may come several minutes to an hour after the sea has withdrawn (Driscoll, 2000).

Any continental slope landslide, however triggered, has the ability to virtually instantaneously release most of the methane from the landslide material itself. After such a landslide, residual methane from the landslide scar would be released for weeks to months afterwards. As with earthquakes and aftershocks, an initial landslide event can trigger additional slides, generally smaller than the initial slide, but with the non-negligible possibility of an even greater slide. The amount of the release, obviously, depends on the quantity of methane that lay below the slide area.

There is no way to predict how large the methane release from an initial slump might be. This depends on initial conditions on the seafloor: the stability of the methane hydrate-containing portion of the continental shelf in one place versus another, the amount of oceanic warming in specific regions of the ocean, the amount of hydrate and free gas in the affected area. Oceanic areas close to the poles will be more vulnerable to methane release for several reasons. First, their stores of hydrate extend to shallower depths. Second, at shallower depths, sediments are usually of more recent origin, are less compacted and therefore less stable. Third, many of these areas are still rebounding from the release of the weight of the great continental ice sheets that lay upon them during the ice age. Fourth, global warming is projected to warm high-latitude (near polar) areas more than mid- or low-latitude (temperate or tropical) areas, and indeed, it is already doing so.

By the time of an initial methane-related slump, however, the amount of carbon dioxide in the atmosphere and its resultant global warming will have well exceeded that needed to trigger the initial event. In addition, the slump's methane and its successor gas, carbon dioxide, will contribute to further global warming, and the warming of the oceans may insure that additional slides follow. Indeed, the slides may continue intermittently over a period of hundreds to thousands of years. Their number and severity will depend, as with the initial slide, on factors such as slope stability, sediment consolidation and water content, and regional fault activity, together with the rate of oceanic warming and changes in global thermohaline circulation and local currents.

Eventually the slumping must slow and stop, because the warming of the sediment takes increasing time with increased depth (Nisbet, 1990). As slumping slows, methane input to the atmosphere will also decline, reducing and eventually removing the source of the warming. By that time, of course, the use of fossil fuels will have long ago ceased, if not from the desire to prevent further damage to the planet, then because fossil fuels will have been completely exhausted. At current rates of use, most of Earth's petroleum will be gone in fifty years, natural gas in sixty, and even coal, the supply of which is expected to last some 300 years at current use rates, will be a distant memory.

Unpredictability

The two possible modes of catastrophically rapid methane release, slumping and massive dissociation (combined with the release of the free methane gas which lies below), are each quite unpredictable, with regard to when each might begin, how long the process might continue, and how much methane could be released. The problem is compounded by the probability that the two modes could be combined in many circumstances.

As oceanic temperatures rise, the warmth will first begin to liberate the free gas and hydrate closest to the sediment surface, in places (the "escape routes") just a few tens of meters down. Reopening the escape routes (chimneys and faults), will allow the further escape of free gas that is buried more deeply under the main bodies of hydrate, hundreds of meters below. This in turn will depressurize the hydrate at its base, allowing its dissociation and release. Finally, this increasing dissociation of hydrate will cause the destabilization of the entire sediment pile, causing slumping.

Just as there is no way to predict how much methane will be released in an initial seafloor slump, there is no way to determine when an initial slump will occur. This depends, in part, on how close particular areas are to threshold conditions. A threshold is literally the sill of a doorway; it is the point where one enters or leaves a house or room. A threshold condition therefore is the place where change begins, from one condition to another. The conditions are often quite different. In the case of continental shelves, the threshold is between stability and instability, between "just sitting there" and sliding.

Basically the situation with oceanic slope stability is no different from snow avalanches, except that we know vastly more about how snow behaves. Consequently, we employ ski patrols to assess how close mountain snow is to the threshold where it will let go, and we warn recreational skiers off slopes when those slopes approach threshold conditions. With oceanic continental slopes, however, we know vastly less about threshold conditions, vastly less about the quantity and location of methane hydrate and free methane that lie within the slopes, and have vastly more area that would require assessment. Because of the many uncertainties and high cost of such assessments, it seems highly improbable that they would ever be done.

Irreversibility

The slumping of continental margins due to the warming of methane hydrate differs in another, extremely important way from snow avalanches. With snow avalanches, with sufficient warning, not only can we warn people away, but we can also trigger the avalanche ourselves by the use of explosives. In many mountainous areas there are even artillery emplacements, from which shells can be fired to cause avalanches to occur when they may do so most safely. In other areas giant gas burners are used to preemptively trigger avalanches. These measures can provide some control over avalanches, at least in those places we monitor and where we take remedial action. With hydrate-related submarine slides, we have no such option, even if we had the advance warning and the actual power to trigger such slumps, which we do not. Advance triggering would only cause the problem we would be trying to avert: the release of methane from the seafloor. Our one option is prevention: stopping the slumping before it starts. And that requires that we stop warming the planet.

Once massive dissociation and/or slumping begins, there will be no way to stop them. Indeed, there will be no way to stop these processes even from well before they start, because of the lag time between increasing atmospheric carbon dioxide and the warming of the globe and the seafloor sediments in which the methane hydrates and free methane reside. As NASA climate modeler James Hansen has pointed out, "even if rising concentrations of greenhouse gases could be stabilized tomorrow, gases that have already accumulated [in the atmosphere] will push surface temperatures up another half degree or so" (Kerr, 2000), an assessment which is supported by others (Miehl, 2005).

This amount of additional warming is known as the "warming commitment" because it represents the amount of warming we are committed to, even without additional carbon dioxide being dumped into the atmosphere. Some climate modelers believe the present warming commitment is even greater than Hansen and Miehl do, and that the planet could warm by another 1°C (1.8°F) over the next twenty years (Wetherald, 2001) or more (Wigley, 2005). As time goes on, and carbon dioxide continues to be dumped into the atmosphere, the amount of the warming commitment will continue to increase.

At some unknown point -- another threshold -- the amount of carbon dioxide in the atmosphere will be sufficient to eventually induce the warming of continental slopes enough to trigger enhanced methane venting, hydrate dissociation, and sediment slumping. Then there will be a period of time -- more lag time -- before massive dissociation and slumping actually begin. During this time the atmosphere will be warming, and the oceans will be warming, and the sediments on the seafloor will be warming. Of course, that is exactly what is going on now, so it is possible that we have actually crossed the carbon dioxide threshold needed for these processes to occur, and that some considerable seafloor methane release is now inevitable.

It is also possible that we have not crossed that threshold, and may not for many more decades, or even centuries. The time, however, is short. Commenting on the determination that climate sensitivity (with a doubling of atmospheric carbon dioxide) may range from 1.9°C to 11.5°C (Stainforth, 2005), Oxford University physicist Myles Allen indicated that "uncertainty over global warming may mean that no such [safe] threshold [for atmospheric carbon dioxide levels] may exist... 'The danger zone is not something in the future,' he says, 'We're in it now'" (Hopkin, 2005).

Certainly, based on numerous projections of global warming, the world will be considerably warmer by the end of this century. Although various estimates of the size of the temperature increase differ (the Stainforth study suggests that a 3.4°C [6.1°F] increase is most likely with a doubling of atmospheric carbon dioxide, and the best 2007 IPCC estimate is 3.0°C [5.4°F]), virtually all estimates project an increase. Moreover, it is important to remember that although most estimates of carbon dioxide release and consequent global warming project only until the end of the current century (or to a doubling of atmospheric carbon dioxide from pre-industrial levels, which most models assume will occur about that time), there is no reason to believe that the anthropogenic warming of the planet will cease at that point, and, indeed, every reason to believe it will continue.

The existence of lag time, but of unknown length, means that we will not suffer the consequences that can ensue when there is immediate feedback. When someone puts a hand on a hot stove, the message is received immediately. But when negative consequences do not immediately follow, there is a tendency to continue behaving the same way as in the past. It is likely, therefore, that carbon dioxide emissions will continue until there is clear, dramatic, and unambiguously negative feedback. That is, until catastrophe. Of course, at that point we will have dumped sufficient amounts of carbon dioxide into the atmosphere that serious warming will continue for hundreds, thousands, or even tens of thousands of years, into the future -- even if we were to stop the dumping immediately.

"The added carbon dioxide declines in a markedly non-exponential manner [that is, not in a smooth, geometric curve]", state the authors of a section of the 1990 IPCC report (Shine, 1990). "There is an initial fast decline over the first 10 years period, followed by a more gradual decline over the next 100 years and a rather slow decline over the thousand year time-scale. The time period for the first half-life [during which half the added carbon dioxide will be gone] is typically around 50 years, for the second [during which half of the remaining added carbon dioxide will be gone], about 250 years" (Shine, 1990).

Calculations such as this have led many to believe that the carbon dioxide we add to the atmosphere will be mostly gone in just a few centuries, or in a thousand years at the most. Even the US Environmental Protection Agency (EPA) states that the lifetime of carbon dioxide is up to 200 years. According to geophysicist David Archer, however, such projections are in error. His own calculations indicate that "about 7% of carbon [in carbon dioxide] released today will still be in the atmosphere in 100,000 years." He further states, "A better shorthand for public discussion might be that CO¸2 sticks around for hundreds of years, plus 25% that sticks around forever" (Archer, 2005). Not quite forever, perhaps, but long enough that it could remain a problem far, far into the future. The carbon dioxide we have already dumped into the atmosphere, and that which we will dump, in other words, will not easily go away. Neither will the warming it produces.

Undoubtedly the warming that the planet has experienced in the twentieth century has already caused the release of additional methane from permafrost and the continental margins. There is no way that it could be otherwise: warming releases methane. Furthermore, the rate of methane release will continue to increase as we continue to warm the Earth. Just when this release will shift from a more gradual to a more catastrophic mode (from chronic to acute, to use the medical terms), is unpredictable. This shift certainly depends on just how close to threshold conditions continental margin methane is, which is something we have no way of knowing. (Permafrost methane, although it potentially could make a major contribution to global warming, is likely to only be released gradually, though the rate of increase could change rapidly.)

When? How long?

A methane catastrophe can be divided in two parts, the initial or onset stage, together with that stage's immediate consequences, and the longer-term consequences.

Onset

For a methane catastrophe to occur, methane must be released in a short period of time. As mentioned previously (at the start of the Methane Catastrophe section), scenarios which allow for methane release over long periods (as many tens of thousands to millions of years) cannot produce catastrophic consequences because the excess methane released into the ocean would be consumed by normal atmospheric and marine oxidation processes, and by expanded populations of marine methanotrophs, other essential nutrients being present. To produce catastrophic consequences, the duration of the initial methane release must be a thousand years or possibly a lot less, precisely to avoid these oxidation processes and the possibility of a methanotroph population explosion.

It is important to note that once in the atmosphere methane is oxidized in less than ten years, though large quantities will temporarily overwhelm the oxidation system and allow for more extended atmospheric lifetimes. Nonetheless, even with a sudden and massive release the initial methane will be around for only a limited period of time. Its successor, carbon dioxide, will remain in the atmosphere considerably longer, but even the level of carbon dioxide will decline as it is taken up by photosynthetic organisms, the weathering of silicate rocks, and the ocean.

There is a second consideration which suggests that the onset stage of methane catastrophes cannot be protracted in length. Microbes can reproduce at quite extraordinary rates when nutrients are present and conditions favorable. Under optimal conditions in the laboratory, for example, the gut bacterium E. coli can double in number about every twenty minutes. Bacterial methanotrophs presumably have similarly high reproductive rates. While some methanotrophs are not bacteria, and are -- being archaea -- extremely difficult to cultivate in laboratories, their potential reproductive rates are also probably quite high. A gradual, extended increase in ambient oceanic methane, therefore, would presumably be easily consumed by the methanotrophs.

A methane catastrophe wreaks its havoc via four primary killing mechanisms -- oceanic anoxia, acid rain, and global warming. For these mechanisms to be maximally effective, they must operate over a very limited period of time. (As mentioned previously, rate is critical.) With the level of excess atmospheric carbon dioxide and acid rain starting to decline just as soon as they are produced (see Watson, 1990, diagram, above, for carbon dioxide decline), and the global heat the carbon dioxide engenders following closely -- though more slowly -- upon, the faster methane delivers its wallop, the more powerful it is.

The work of numerous scientists has set an upper limit on the possible length of the Paleocene-Eocene methane release by their determinations of the duration of the carbon isotope excursion. Dickens (1995) found it took less than 10,000 years; Bralower (1997), about 6000 years; Katz (1999), less than 5000 years; Norris and Röhl (1999) "a few thousand years or less." Kennett and Stott's 1991 finding that Southern Ocean temperatures jumped about 8°C in only 2000 years probably further constrains the length of the Paleocene-Eocene methane release. Considering that the faster it is, the stronger the punch, therefore, an upper limit of a few centuries to possibly a thousand years for the initial methane release is not unreasonable. To the extent that seafloor methane was released (at least in part) at the end of the Paleocene and the end of the Permian via the intrusion of magmatic sills (which must be emplaced on a time scale of decades: Svensen, 2004), the release time might have been shorter still.

It should be emphasized that this is simply the initial, catastrophe-producing release: the initial jolt. As projected here, the altered climate and oceanic conditions can last for millions of years, as they did after the initial catastrophic methane release of the end-Permian. This is because the initial jolt, with its anoxia, acid rain, and global heat, so reorganizes the global climate and ocean system that dramatically changed conditions can persist for great lengths of time before and as recovery occurs. As modeled by Dickens (1997), for example, the residual heat from a major methane release might last for two million years or so, though the main warming would quickly follow the release, and only minor warming would persist longer.

Triggers, Present and Past

The same "the faster it is, the stronger the punch" logic that governed the Paleocene-Eocene warming applies as well to the methane release triggering mechanism, increasing atmospheric carbon dioxide. With our profligate burning of carbon fuels, however, carbon dioxide seems to be entering the atmosphere at a rate vastly faster than with any normal natural process.

We have already consumed almost half of the world's supply of petroleum. The estimates of the total amount of recoverable petroleum differ, and some petroleum experts believe that it will be another decade or two before we reach that halfway point. Nonetheless, within another fifty years, most of the rest of the world's petroleum will be gone. That means that we will no longer have to worry about carbon dioxide emissions from petroleum-derived gasoline and other fuels. (Petroleum is used for other purposes than fuel, such as for plastics, paraffin, lubricants, and asphalt, but its primary use is to be burned for energy.) All the carbon dioxide from those emissions will already have been dumped into the atmosphere, although much will enter the ocean thereafter. (Fortunately for those who cannot live without the internal combustion engine, gasoline can be made from coal, as it was by the Germans during World War II.)

Not only will most of the world's petroleum be gone in some fifty years, in sixty to seventy years, most of its ordinary natural gas (that is, excluding that hydrate methane currently locked in permafrost or in continental margins) will also be exhausted. Each of these carbon reservoirs (petroleum and natural gas) is estimated to hold about 500 billion metric tons (each roughly equivalent to an imperial, or American, ton) of carbon (Kasting, 1998). Burned for their energy and injected into the atmosphere as carbon dioxide, that's more than enough carbon than is needed to double the atmosphere's current carbon reservoir (now approaching 800 billion metric tons).

The Earth's reserves of coal, the most plentiful of the fossil fuels, will last a bit longer: for 220 years, according to some estimates; for 300 or more, according to others. Obviously, however, and within two centuries, most of the planet's coal will also have found its way -- as carbon dioxide -- into the atmosphere and oceans. And there's an estimated 4000 billion metric tons of carbon (Kasting, 1998) that awaits its release via the burning of coal. That's more than five times the amount currently in the atmosphere.

For most climate change modeling purposes, future global warming is estimated on the basis of a doubling of pre-industrial atmospheric carbon dioxide. Such a doubling is projected to occur, at current rates of carbon dioxide emissions, before the end of the current century. (Measurements from the Mauna Loa Observatory now indicate the doubling period may be closer to sixty than to one hundred years.) Possibly because of the uncertainties associated with current projections of global warming, few models examine what could happen as the dumping of carbon dioxide continues beyond the point where atmospheric carbon dioxide has doubled, but based on our current use patterns, such dumping is almost certain to continue.

As paleoclimatologist (one who studies ancient climates) James Kasting has noted (1998), we have the ability not only to double the pre-industrial level of atmospheric carbon dioxide once, but to double it again, and double it yet again. (And, it should be added, we could actually come close to doubling it a fourth time!) And all in a period of perhaps just over two centuries, or three at most. In short, in addition to the carbon dioxide we have already dumped into the atmosphere, we have the ability to dump lots more, and probably will. Enough to thoroughly and rapidly warm the planet, and trigger a methane catastrophe.

The past 500 million years have recorded only a small number of methane catastrophes. Those of the end-Permian, the end-Triassic, the Jurassic (Toarcian and Oxfordian), the Paleocene-Eocene, and perhaps a few of lesser importance in the early Cretaceous may comprise the entire list. Probably only these satisfy the prerequisite of a rapid and significant warming trigger.

For the greatest of these events -- those of the end-Permian, the end-Triassic, the Paleocene-Eocene, and the Toarcian -- the warming trigger was presumably twofold. The most important trigger for each of these events was probably the direct intrusion of seafloor sediments by volcanic magma during the eruptions which created large igneous provinces: the Siberian Traps, the Central Atlantic Magmatic Province (CAMP), the North Atlantic Igneous Province (NAIP), and the Karroo Igneous Province, respectively (Vermeij and Dorritie, 1996; Dorritie, 2002; Svensen, 2004). In addition to directly heating the hydrate-bearing sediments, the magma also warmed the marine sediments and, thereby, the ocean. Where the ocean basin was small and restricted in its circulation, as with CAMP, the NAIP, and the Karroo Igneous Province, oceanic warming would have been more effective than if the massive volcanism took place in a less enclosed setting. But colder regions (as with the Siberian Traps) would have been more seriously impacted than warmer ones, other things being equal.

Each of these eruptive sequences probably had a large underwater component, although some may have had a major subaerial (under the air, that is, terrestrial) component as well. But all of these eruptive sequences injected large quantities of carbon dioxide into the atmosphere (the Central Atlantic Magmatic Province and the Siberian Traps most), the submarine eruptions acidifying the ocean locally as the carbon dioxide rose through the water column. This injection of carbon dioxide into the atmosphere caused a general warming of the surface of the planet, much as we are doing today. Except that our own (anthropogenic) carbon dioxide release is probably proceeding considerably faster than the volcanogenic carbon dioxide injections of the past. [My own "back-of-the-envelope" calculations indicate that we are currently injecting carbon dioxide into the atmosphere at a rate that is somewhere between ten and hundreds of times faster than the average rate of volcanic degassing during the eruption of the Central Atlantic Magmatic Province (CAMP degassing estimates from Beerling and Berner, 2002). The anthropogenic injection rate, moreover, is constantly increasing.]

Longer-term Consequences

It is certain that the methane that is now being released from permafrost and the seafloor will contribute to global warming, initially as methane, with a greenhouse capacity that vastly exceeds that of carbon dioxide, and then, upon oxidation, as carbon dioxide itself. With a greenhouse warming ability more than 20 times carbon dioxide, methane that reaches the atmosphere (and much that is released is likely to do so, because it will exceed the current capacity of the marine methanotrophs to consume it) has the potential of doing far more damage to the planet's climate and biosphere than all the carbon dioxide that has been and will likely be released into the atmosphere by the burning of fossil fuels.

Simple arithmetic shows why. The total amount of carbon locked up in fossil fuels is about 5000 billion metric tons (each slightly more than an imperial ton). As mentioned, in less than 100 years, almost all the carbon from petroleum and natural gas will have been converted to carbon dioxide; by about 300 years from now, all the coal will be gone as well. So in about 300 years, much of 5000 billion metric tons of carbon will have moved into the atmosphere as carbon dioxide, and about half may then wind up in the ocean, where it will no longer help warm the planet, but will continue to acidify the ocean. (There are other natural "sinks" for carbon dioxide besides the ocean, but research indicates that the absorptive capacity these sinks may change greatly over time. Consequently, we should not be depending on such natural sponges for soaking up the carbon dioxide we are pumping into the atmosphere.) This carbon dioxide -- alone -- contains an amount of carbon equivalent to almost ten times the amount of carbon that was in the atmosphere at the beginning of the industrial age, and roughly seven times that of the present.

By contrast, there may be more than 10,000 billion metric tons of methane carbon that can be released from the seafloor. Though this quantity is only twice that of the carbon in fossil fuels, it possesses more than 40 times the short-term warming potential of carbon dioxide. A 'mere' 250 billion metric tons of methane carbon -- less than 1/40th (2.5%) of the estimated total seafloor methane carbon reservoir -- has the warming capacity of all fossil fuel carbon. Looked at another way, a release of just 1% of seafloor methane (somewhat more than 100 billion metric tons) has a warming potential several times greater than the amount of anthropogenic carbon dioxide which scientists project will enter the atmosphere in the next 60 to 100 years.

The geological record provides us with only minimal guidance as to what might be expected from major methane releases. Effects of the known methane catastrophes (at least those which rise to the level of "catastrophe," which involves a several degree warming of atmosphere and ocean, and probably at least transient ocean anoxia), vary considerably. At the lower end of the range (represented by the oceanic anoxic events of the Toarcian, 183 million years ago, and Aptian, between 116 and 112 million years ago), there was limited global warming and transient deep ocean anoxia. More serious oceanic anoxia and warming seems to have occurred at the end of the Paleocene (the LPTM). Finally, in the most catastrophic event, at the end of the Permian, there was a stunning global warming, a euxinic ocean that lasted for millions of years, a significant drawdown of atmospheric oxygen, and the most massive of all mass extinctions.

No doubt variables such as the geographic location of particular methane-releasing submarine slumps, the general state of the global climate at the time, and the configuration of continents and oceans played a role in determining the severity of these catastrophes, but probably the most significant factors were the amount of and rate at which methane entered the atmosphere, and the length of time it remained there. While we have no way of estimating how much methane could enter the atmosphere as a result of current anthropogenic global warming, as noted above, even the release of a minute proportion of that which is available could wreak havoc on our planet.

There are too many factors of unknown size to allow any prediction of the long-term consequences of a future methane catastrophe. The end-Permian will not serve as a model, because the world was warmer then, even before the catastrophe. Furthermore, during the past twenty million years or so, the globe has cooled considerably. Many climatologists attribute this to the movement of Antarctica to its current position directly over the South Pole: if they are correct, the current cooler period of Earth's history will continue until Antarctica moves off the pole, an event at least tens of millions of years in the future.

The relatively greater warmth of the Permian period may have been a contributing factor to the oceanic anoxia of the Early Triassic. It is possible that in today's cooler world, thermohaline circulation could be restored quickly, ventilating the deep ocean, and allowing for rapid recovery. Even so, a "rapid" recovery could take many millennia, perhaps tens or even hundreds of millennia. If the jolt to the climate system is great enough, however, and the ocean becomes largely anoxic, methane presumably would continue to be produced by marine anaerobes until thermohaline circulation, and therefore deep water oxidation, are restored.

When will the input of methane overwhelm the global climate system?

Again, the unknown factors involved preclude a simple answer. With present-day carbon dioxide, we know the significant sources and have a fair sense of their coming likely concentrations in the atmosphere. There is no reason to expect that we will be surprised by a major release of carbon dioxide from an unanticipated source. As destructive as carbon dioxide is -- and will be -- for the Earth and its inhabitants, at least we have a fairly good understanding of it, and can -- if we choose -- have some control over it.

This is not the case with continental margin methane. Although we have a rough estimate of its global quantity, we have little idea of the details of its worldwide distribution. We have no idea of how close continental margins may be to the slumping threshold, which presumably varies from place to place. We do not know how long it will take for the oceans to warm as the atmosphere does. In fact, we are enormously surprised that they have warmed as much as they have (Fukasawa, 2004). We thought we knew how fast it would take for oceanic warming to reach the region of the methane hydrates (Harvey and Huang, 1995); we were wrong. Ocean warming will release continental margin methane much faster than we previously thought (Pecher, 2002; Wood, 2002; Zühlsdorff and Spieß, 2004).

The major factors involved in continental margin methane release seem to be the following (assuming no major near-term change in global thermohaline circulation):
1. How fast the oceans are heating up, particularly to the depth of the deepest methane hydrates (about two kilometers, or 1.2 miles).
2. How fast that heat takes to penetrate the sediments to the base of the gas hydrate stability zone (BGHSZ). (The BGHSZ, as previously mentioned, is identical with the BSR, the bottom-simulating reflector detected by sonar.)
3. How much methane will be released through the venting of free gas and dissociating hydrate as the oceans warm.
4. How close continental margins are to their slumping thresholds.

While we cannot expect to know, except in retrospect, how close continental margins are to their slumping thresholds (item 4, above), or how rapidly free methane and dissociated hydrate methane can be released into the ocean and atmosphere (item 3, above), we may be able to obtain estimates for the first two factors. The amount of time it would take for ocean warming to penetrate the sediments to reach the hydrates (item 2, above) has in fact been estimated. With the findings about the "roughness" of the BSR, and the closeness of some hydrate and free gas to the seafloor surface (about 15 meters; Wood, 2002), the sediment penetration time may be as short as 55 years for a 6°C heat pulse (Pecher, 2002). At the other extreme, it could take thousands of years for a heat pulse to penetrate to the deepest hydrates.

The warming of the oceans may be a second factor that can be estimated, though we do not yet possess enough data to do so. We do know that down to a depth of 3000 meters (about 2 miles), the oceans have warmed 0.06°C (about 0.1°F). This is a minuscule amount of warming, but it has been determined with great precision based on millions of measurements (Levitus, 2000), so it is reliable. We also know that the globe as a whole warmed 0.6°C (1.1°F) plus or minus 0.2°C (about 0.4°F) during the 20th century (Levitus, 2000). That's 0.06°C ocean warming for 0.6°C global warming. Because the atmosphere warms first (it has the greenhouse gas) during global warming, and then it warms the ocean, ocean warming lags atmospheric warming. We therefore can state with reasonable assurance that the oceans will warm a minimum of about 0.1°C for every 1°C (that's the same ratio as 0.06°C is to 0.6°C) that the globe warms, at least in the near future.

If we then take the IPCC's maximum estimate for global warming for the 21st century (remembering that it may be a conservative estimate), that is, 5.8°C (10.4°F), we can say that the maximum ocean warming in the 21st century may be 0.58°C (1.04°F). "May be" should be stressed. It could be less, just as the IPCC's 21st century warming estimate ranges from 1.4°C (2.5°F) to 5.8°C (10.4°F). On the other hand, it may well be more. Only when we have more data, in another decade or so, will we be able to know the rate of oceanic warming with some reasonable certainty.

This small amount of projected deep ocean warming (about 0.6°C/1°F in the 21st century) is unlikely to dissociate much hydrate. At such a slow rate of warming, it would indeed be several centuries, if not much longer, before even the free gas and hydrate closest to the sediment surface began to be released. That is, if it took a significant heat pulse (of say, 6°C, or 10.8°F) to release continental margin methane. But it does not. At least some free gas below the hydrates, remember, may be at threshold conditions, right now (Zühlsdorff and Spieß, 2004). That means that any warming whatsoever -- including the tiny amount of warming which has already occurred -- may be enough to trigger the release of at least some methane. Like the teapot on the stove in which the water is about to boil, any increase in global heat can set the whistle blowing -- or the methane flowing. How much methane will be released is something we will discover, but in view of the huge amounts of methane available in the continental margins, even a little may be sufficient to dramatically alter climate.

Global warming and methane release: A summary chart
Why the release may begin to arrive sooner than anticipated.

The release of methane from seafloor hydrate via the warming of seafloor sediments involves a number of processes. As atmospheric carbon dioxide increases, the atmosphere and surface of the Earth warm. This warming also warms the ocean. Eventually the ocean's increased warmth penetrates the sediments to hydrate depth. Estimates have been made about the duration of each of these processes. But more recent findings or modeling suggest that the processes may be proceeding more rapidly than the generally accepted views project:

Process	Generally accepted view	Cause for concern
Atmospheric carbon dioxide increase	About 100 years to a doubling of CO¸2 from pre-industrial levels (at the long-term rate of increase of 1.8 ppmv per year). Pre-industrial CO¸2 was 280 ppmv; the current level of CO¸2 is 380 ppmv.	Recent measurements from the climate observatory on Mauna Loa, Hawaii, indicate that the rate of increase may have accelerated during the past five years (Lean, 2004). At the current rate (2.2 ppmv), it will only be about 80 years to a doubling of CO¸2. There is no reason to believe, however, that CO¸2 accumulation in the atmosphere will stop at the arbitrary limit of only 560 ppmv, and every reason to believe that it will not.
Warming of atmosphere ("climate sensitivity": The usual way that climate sensitivity is estimated is with a projected doubling of atmospheric carbon dioxide. Climate scientists frequently use the formula DT¸2X to describe this sensitivity. The D means a change in; the T is for global temperature; the 2X subscript refers to a doubling of carbon dioxide. )	IPCC says that with a CO¸2 doubling, global temperatures at the end of the century will likely be in the range of 1.4 to 5.8 °C (2.5 to 10.4°F) According to Kerr, 2004 (Three degrees of warming), the general consensus among climate scientists now is that a doubling of atmospheric CO¸2 will most likely produce a 3.0°C/5.4°F warming.	Kerr, 2004, does note, however, that while climate scientists generally agree on a lower bound of 1.5°C/2.7°F for likely climate warming, and a most probable warming of 3.0°C/5.4°F, there appears to be little agreement on the upper bound: "The calculation of sensitivity probabilities goes highly nonlinear at the high end, producing a small but statistically real chance of an extreme warming." This uncertainty is greatly compounded by the realization that it seems highly improbable that the anthropogenic increase of atmospheric CO¸2 will cease with a mere doubling of CO¸2. In addition, climate modelers Andronova and Schlesinger (2001) foresee a warming of between 1.0°C and 9.3°C (1.8°F to 16.7°F) by the end of the century, and Alley has warned that global temperatures could rise 10°C (18°F) in just a short time, "tripping the switch" towards abrupt climate change in only a few decades (Showstack, 2001) The climate sensitivity projected by Andronova and Schlesinger has now received supprort from the largest climate modeling simulation ever done, the climateprediction.net experiment involving almost 100,000 home computers. With a doubling of atmospheric carbon dioxide, this study indicates that global warming could be as much as 1.9°C (3.4°F) to 11.5°C (20.7°F; Stainforth, 2005). Research on the Cretaceous Period, specifically from about 90 million years ago, may provide further evidence that increased atmospheric carbon dioxide produces higher temperatures than previously believed. Instead of the earlier estimates that the warmest Cretaceous tropical ocean temperatures were around 9°C higher than today's (at about 28°C), they may have been as much as 14°C higher. Contemporaneous carbon dioxide levels are estimated at between two and six times today's 380 ppm. Perhaps more startling is the suggestion that one of the most-relied-upon climate models (GENESIS, developed at Penn State University) may be seriously underestimating climate sensitivity: in order to obtain the proper match between Cretaceous tropical ocean temperatures and atmospheric carbon dioxide levels, the assumed level of atmospheric methane had to be increased to some thirty times what it is today. Though these results require confirmation, they may be pointing to major shortcomings in at least some climate models (Bice, in press, cited in Kintisch, 2006). An additional factor in the climate sensitivity models has now also changed. Surprisingly, it has changed because of something we are doing right: cleaning the air of pollutants. As we have reduced the level of aerosols (particularly those of sulfate, often found in coal, and thus released by coal-burning industries), their presence in the atmosphere has been reduced. But aerosols actually help cool the atmosphere, as we have known for some time. The amount of cooling, however, was only estimated. Now, as skies have become clearer, we have a better understanding of just how much cooling the aerosols contributed. It is considerably more than most computer climate models had assumed. When these new, observation-based figures are plugged into the climate models, they indicate "that future global warming may proceed at or even above the upper extreme of the range projected by the IPCC" (Andreae, 2005). According to the study's lead author, "'If our model is right, things could become totally uncontrollable in the second half of the century.'" Andreae's model, using relatively conservative input values, indicates that global temperatures could increase by 6° to 10°C (10.8° to 18°F) by the century's end (Schiermeier, 2005). That is even higher than the IPCC's highest estimate of 6.4°C (11.5°F: IPCC, 2007).
Warming of oceans	Deeper ocean to take about 1000 years to begin to warm	Levitus, 2000, found that the North Atlantic's temperature, averaged down to 3 km (about 2 miles), has increased 0.06°C, or about a tenth of a degree F, in 40 years. Similar warming has occurred in all oceans. Fukasawa, 2004, found a tiny but measurable temperature increase in deep waters (5 km/3 miles) in the North Pacific over a period of 14 years. At such depths, there should have been no warming at all. Finally, a slowing of global thermohaline circulation (such as contemplated by Broecker, 2001) could allow much faster warming in the North Atlantic and perhaps elsewhere.
Warming of sediments to hydrate release depth	Thousands of years, based on the former understanding of the BSR (Harvey and Huang, 1995, and Berner, 2002) Much faster, based on the new understanding of the BSR (Wood, 2002).	Pecher, 2002, suggests that heat from a warmed ocean could penetrate the 15 meters of sediment to reach the topmost methane hydrates in as little as 55 years. With much methane hydrate already at critical pressure, methane could begin to be released from sediment just as soon as it begins to warm (Hornbach, 2004). Some of the free methane underlying the hydrate, moreover, may be at threshold conditions, ready for release just as soon as the sediment begins to warm (Zühlsdorff and Spieß, 2004).
Sources, sinks, and feedbacks. Just as there are numerous sources of the carbon dioxide in the atmosphere -- that produced by the burning of fossil fuels being the major concern bceause it is disturbing the previous natural balance -- there are also what are referred to as carbon sinks. A sink is a way in which carbon dioxide is removed from the atmosphere. The burial of carbon debris in sediment at the bottom of the ocean, for example, is an extremely important sink, partly because of the amount of carbon which is removed from the carbon cycle, and partly because this removal is for geologically significant periods of time, and therefore is essentially permanent. Trees constitute another sink because they incorporate carbon from carbon dioxide in their trunks, branches, leaves, and roots, but this storage is only temporary because trees eventually die and rot, releasing that carbon. Over time, sinks may become saturated with carbon dioxide, and begin to release it instead of storing it. This change from sink to source is projected to occur with the terrestrial biosphere over the course of the 21st century (Cox, 2000).	Most computer modeling of climate change does not incorporate feedbacks from the biosphere (the realm of living things), which can enhance or reduce global warming. An example of of a biosphere feedback which could enhance warming is faster metabolism by soil microbes, which would force more carbon dioxide and methane into the atmosphere, increasing greenhouse gas warming. Such a feedback, because it adds to the initial effect, is referred to as a positive feedback. On the other hand, at least some plants photosynthesize faster in warmer conditions, thus drawing down atmospheric carbon dioxide, at least as long as those plants are alive. This is an example of a negative feedback, because the initial warming, in this feedback, reduces greenhouse gas and thus helps cool the atmosphere. The reason that most modeling fails to include biosphere feedbacks is that the models are enormously complicated even without incorporating such feedbacks, despite the recognition that they could have important effects on projected outcomes. In addition, the actual effects of many presumed biosphere feedbacks are only beginning to be understood, and introducing them into models increases uncertainty.	Some of the most important possible feedbacks are from increased photosynthesis by land plants, and from the changes in the carbon dioxide uptake of soils as a consequence of global warming. Although it was once believed that increased warmth would result in a greater uptake of carbon dioxide, it now appears that this uptake will be much less than previously projected (Cox, 2000; Sarmiento, 2000). The oceans are also likely to absorb less carbon dioxide than had been thought (Cox, 2000; Sarmiento, 2000). In other words, the oceanic and terrestrial biosphere sinks are likely to prove considerably less effective than had been assumed. This means that as atmospheric carbon dioxide increases, more will remain in the atmosphere. This extra carbon dioxide may result in an additional 2.5°C (4.5°F) greater global warming over land than projected by most climate change models (Cox, 2000; Sarmiento, 2000).

Worst Cases:

The foreseeable effects of a methane catastrophe -- an anoxic deep ocean, an acidic ocean, stunning global warming, desertification, increasingly acidified precipitation, and stronger precipitation events -- will be sufficient to inflict a colossal blow to the global environment, driving countless species to extinction and rendering the conditions of human life immeasurably more difficult. Unfortunately, there may be even worse consequences.

As global warming proceeds, thermohaline circulation will slow and the world ocean will become increasingly stratified. That is, the deeper ocean (below about 100 meters/yards), with its nutrients, will become increasingly more isolated and less interactive with the surface ocean, with its phytoplankton. In oceanographers' terms, the ocean will become less mixed. This will happen as a result of the warming of the polar regions, which will see a much greater temperature increase than the rest of the planet. (As Alaskans know, this is already happening.)

According to some newspaper reports and scientific projections, the "Northwest Passage," an ice-free marine route around northern North America from the Atlantic to the Pacific, may be open in just a few decades. The Northwest Passage achieved its fame in the sixteenth and again in the nineteenth centuries, when it was vainly (and sometimes fatally) sought after by several seafaring expeditions. The problem, we now fully recognize, was that this route through Arctic waters was blocked by sea ice. With global warming, however, Arctic sea ice will melt, and the Arctic will become largely ice-free for much or eventually all of the year.

The often gushingly enthusiastic projections of a much shorter maritime route between Europe and the Far East, however, ignore another consequence of the disappearance of the sea ice. When seawater freezes, seawater salt is left behind: the ice is entirely composed of fresh water. The remaining seawater is consequently more saline, and more dense. With global warming, as sea ice production slows or ceases, this "saltwater fractionation engine" also slows or stops.

If no sea ice is produced, highly saline water cannot be produced. This dense and frigid water, which today carries oxygen to the ocean floor and drives global thermohaline circulation, will cease to be. Its cessation will be the predominant one of the many causes of the oceanic stratification and anoxia that is to come.

As the deep ocean becomes increasingly dysoxic and then anoxic, the aerobic organisms which currently inhabit it will be replaced by anaerobic microbes, mostly sulfate-reducers. And, as in today's Black Sea, they will pump out hydrogen sulfide as a waste product. The deep ocean will become toxic as hydrogen sulfide attacks all available iron (that dissolved in the ocean, and that in living things) and sends it to the bottom as iron pyrite.

Some hydrogen sulfide will undoubtedly escape its confinement in the deep ocean and wreak havoc on neighboring coasts at least locally (as it does along today's Namibian coast) and perhaps regionally. But will it escape in sufficient quantities that it will produce a globally toxic atmosphere, as has been suggested for the end-Permian (Grice, 2005; Kump, 2005)? That depends on the efficiency of green and other sulfur bacteria.

As previously noted, green sulfur bacteria are anaerobic and anoxygenic (not using or producing oxygen) photosynthesizers that live at the boundary between the oxic shallow waters of today's Black Sea and its deeper anoxic waters. They consume most of the hydrogen sulfide produced below. As anoxia encroaches on shallower ocean depths, however, and allows the green sulfur bacteria (and their cousins, the purple sulfur bacteria) to rise higher in the water column, their activity presumably becomes more efficient: there's more light available as an energy source. That means that they (as well as their more distant cousins, the large sulfur bacteria, which do not require sunlight) ought to be able to consume more hydrogen sulfide. Other things being equal, therefore, the efficiency of the sulfur bacterial "lid" (on the hydrogen sulfide) ought to increase as anoxia reaches shallower depths. But other things are not equal.

At increasingly shallow depths, regrettably, ocean turbulence is greater, and the likelihood that this lid will be disrupted is similarly greater. And global warming is projected to increase the occurrence of extreme weather events, which churn the ocean surface. The closer the probably imperfect sulfur bacterial lid is to the surface, moreover, the shorter the distance though which the hydrogen sulfide must pass to make its escape into the atmosphere. In any case, the ultimate ability of the sulfur bacterial lid to contain the hydrogen sulfide in the deep is unknown. There is a distinct possibility, therefore, that large quantities of hydrogen sulfide could indeed breach containment, and deliver an additional, terrible blow to an already reeling biosphere. Thus,

in the following fashion:

There is a second worst case scenario. Deep ocean anoxia -- that is, anoxia extending as high as to within 100 meters (yards) of the ocean surface -- will also kill off the giant larvaceans (remember them?) which live at depths from 100 to 500 meters. The demise of the giant larvaceans will mean that the carbon rain they snag and deliver rapidly to seafloor will take many times longer to make that trip, increasing its chances of being decomposed along the way.

Were the demise of the giant larvaceans the only consequence of the changed oceanic conditions in a warmer world, it would eventually result in a serious depletion of atmospheric oxygen, as less carbon would be removed from the global carbon cycle. But in an anoxic ocean, decomposition will take place by anaerobes instead of aerobes. This decomposition is much less efficient, so much more carbon will make its way to the seafloor. Over time, this additional carbon burial should increase the level of atmospheric oxygen.

But there will also be unpredictable changes in total carbon rain from near the ocean surface, where most of it originates. On one hand, the carbon dioxide content of the ocean surface will be increasing, potentially allowing for an increase in phytoplankton activity, as well as an increase in the zooplankton which consume it. On the other hand, phytoplankton blooms are likely to be reduced by nutrient deprivation, brought on by the limited water column mixing that will accompany a stratified ocean. In addition, the coccolithophores, as noted previously, are likely to be hard hit by increasing ocean acidity, which will dissolve their skeletons. Coccoliths constitute a significant portion of the phytoplankton. The less phytoplankton, the less oxygen being released into the atmosphere. In addition, the less phytoplankton, the less carbon rain, and the less carbon being removed from the global carbon cycle.

The interaction of the disturbance of these various oxygen-carbon cycle components is difficult, perhaps impossible, to predict. Nonetheless, the potential for major, long-term changes to the composition of the atmosphere is certain. On balance, it seems more likely that oxygen levels may decline, especially with the increased concentration of methane and possibly hydrogen sulfide in the atmosphere. Creatures that require oxygen levels close to the present 21% -- like us -- would be hard-pressed to survive, because such a high level would not exist anywhere on the planet. If oxygen levels do rise -- a scenario that carries its own health risks, such as increased cancer -- people would be able to migrate to higher elevations, at which oxygen levels would be less. But there's less room up there as well, and a lot less agricultural land. In neither case is the future stability of the oxygen level assured: there could be great and even erratic departures from any general trends as marine biology and chemistry shifts.

There is even a third worst case scenario. Just as methane and hydrogen sulfide may have worked together at the end of the Permian to destroy stratospheric ozone, the same could happen in our future. Thus,

methane catastrophe

+ (plus)

hydrogen sulfide catastrophe

Æ (could lead to)

ozone catastrophe

With refrigeration gases having punched two polar holes in the ozone layer, we already have a leg up on this final scenario.

These worst case scenarios -- a hydrogen sulfide catastrophe, unpredictable changes in the atmospheric oxygen level, and the destruction of the ozone layer -- are not mutually exclusive, it should be noted. The first two arise from ocean warming and anoxia, and the third follows from the gases that are released as a consequence. As oceanic stratification and global warming proceed, all scenarios are possible, and, if anoxia takes does hold of the deep ocean, all may be inevitable.

A DEPLETED EXISTENCE

During a methane catastrophe (there seems no point in trying to describe an aftermath that lies thousands of years in the future), human beings will face a depleted existence. Global warming is currently projected to kill off, or "commit to extinction," between 15 and 37% of presently-existing species in just 50 years (Thomas, 2004); far more will be driven to extinction by the end of the century, and in the centuries following. Global warming (and its accompanying effects of acid rain, an acidic and largely anoxic ocean, and so on) will not be the only cause of the extinction of species: increasing encroachment by the rapidly growing human population on the habitats of other organisms, the unceasing exploitation of limited organic resources (as trees, fish) and the conversion of the planet solely to suit our own needs will also take their substantial toll. In addition to the extinction of other species, we will also face the destruction of our own global economy.

There are at least three outstanding examples of what happens when people ignore the limitations imposed by climate, or when they simply become the victims of natural climate change.

The Maya

At one time, the Mayan regions of Central America (predominantly in Guatemala and Mexico's Yucatan) hosted a population of from three to thirteen million. This was at the height of the Classical Maya, about 750 CE. The Maya, who had begun to construct cities about 150 CE, apparently suffered one setback about 250 CE, when they abandoned these cities, perhaps as a consequence of drought. But thereafter the Maya flourished, until the beginning of the ninth century. Within about one hundred and fifty years, first in more southern and central regions and later in the north, Mayan civilization fell into sharp decline, and its cities were abandoned (Haug, 2003).

The probable cause was a series of droughts, the first short episode occurring about 760 CE. Then, as the area became generally somewhat drier, three catastrophic multi-year episodes of drought seem to have happened, centered about 810, 860, and 910 CE. Interestingly, this timing seems to coincide with periods of intense cold in Scandinavia, possibly indicating a global rather than regional change in climate. The timing of the droughts also seems to coincide with a three-stage pattern of the abandonment of Mayan cities, though this suggestion apparently is controversial. Nonetheless, the episodes of drought likely gave rise to the cultural upheavals that characterized the Terminal Classical Period, and the final Mayan collapse (Hodell, 1995; Gill, 2000; and Hodell, 2001, as confirmed by Haug, 2003).

At its peak, Mayan civilization was presumably able to address the economic needs of its population. But that population, because of the very success of its economic and political system, probably had reached the limit of the ability of the Central American region's to provide for Mayan needs, particularly for food. The food supply depended on the availability of water, and an assortment of cenotes (natural pools in limestone), reservoirs, and water conservation strategies were employed to supplement the highly seasonal rainfall. When climatic conditions began to decline, however, the region's ability to support the Mayan population also fell, and during the times of multi-year drought, Mayans would have starved.

The Mayans had reached the carrying capacity of their environment; then, when conditions became decidedly less favorable, that carrying capacity plummeted, and with it went Mayan civilization. The term 'carrying capacity' is customarily applied to rangeland in the case of livestock, and a given ecological area in the case of wild animals. But its application seems quite appropriate here (it is used by Haug, 2003) in dealing with the consequences of adverse climate change on human cultures. Numerous other civilizations -- among them that of the Moche in Peru, the Anasazi in the US Southwest, that of Ubar in Oman, and many others at the edges of desert regions in Africa and Asia -- may have suffered similar fates. Significant heat or cold, or drought, can abruptly lower the carrying capacity of areas inhabited and developed by human beings, with consequences similar to those that happen to other creatures.

Iceland

The history of Iceland provides another glimpse of what a depleted existence could be like. Iceland was settled by the Viking farmers in about 870 CE, and within fifty years, most of the island's birch trees had been chopped down, for fuel, building, and clearing land for farming. The farmers had brought cattle, sheep, horses, goats, and pigs with them, but the goats tore apart the forests and the pigs ripped up the land in their search for food. Within about two hundred years, possibly because they had depleted their own food sources, the goats and pigs had all but disappeared, but the damage they had done remained. The island's trees were almost gone, and the soil, which requires centuries to replenish in the colder parts of the world, was rapidly being eroded away (Ogilvie and McGovern, 2000).

In the meantime, the global climate had turned colder, and the warm conditions that had allowed the Vikings to settle even in Greenland were ending. The Greenland settlements were abandoned, or their inhabitants simply starved, and Icelanders were pushed to the brink. Despite a lack of wood for fishing vessels, many Icelanders were forced to fish to supplement the meager production of their farms, which had been devastated by the loss of farm animals during increasingly harsh winters. Both people and natural climate change were therefore responsible for the serious economic hardship and even starvation that occurred in the fourteenth through the sixteenth centuries (Ogilvie and McGovern, 2000). Those with extensive land and livestock holdings were in a better position to survive: "It is not surprising that these rulers probably chose to ignore the early signs of climate change despite its long-term threats to Icelandic society as a whole: then as now, politics and ecology seem to have been closely interconnected" (Ogilvie and McGovern, 2000, p. 390).

Easter Island (Rapa Nui)

Another island may provide an even better indication of what happens when people ignore the constraints their environment imposes. Unlike Iceland, which is a large island located about a thousand kilometers (600 miles or so) from Scandinavia, Easter Island (Rapa Nui) is a tiny patch of land (about 160 square kilometers; 64 square miles) isolated by more than 3000 kilometers of South Pacific ocean from South America, the nearest continent. No ocean-going vessels intentionally made their way to Rapa Nui during most of its existence as an inhabited island (in contrast to Iceland, which was frequented by ships from Scandinavia), and indeed, for hunderds of years, the world never knew or even suspected that Rapa Nui existed. It has been described as "the world's most isolated scrap of habitable land" (Diamond, 1995).

Settled around 1200 CE (Hunt and Lipo, 2006; other scientists place the settlement date at the fifth century CE [Diamond, 1995] or about 900 CE [David Steadman, as cited in Bower, 2006]) by Polynesian seafarers, Rapa Nui, like Iceland, had a fertile volcanic soil. Like Iceland, it possessed a forest, composed, because of its position near the Tropic of Capricorn (23 1/2°S), of subtropical trees, bushes, and shrubs. Its isolation made it a haven and nesting site for numerous species of seabirds, as well as birds like owls, parrots and herons that made Rapa Nui home. Shellfish, fish, and dolphins were abundant locally. The Polynesians themselves also brought chickens (in addition, inadvertently, to rats), and grew sweet potatoes, bananas, sugarcane and the edible taro root (Diamond, 1995).

Once the inhabitants of Rapa Nui reached a level of economic stability and success, they were able to devote part of their attention to the construction of the massive stone figures for which the island is famous. Hundreds of these huge, multi-ton statues were erected after being carved out of the rock and hauled long distances (about 10 kilometers, or 6 miles); many hundreds more were abandoned at various stages of completion. The period of figure carving and statue erection seems to have occurred between about 1200 to 1500 CE, just as the island's population was peaking at about 7,000 to possibly as many as 20,000 (Diamond, 1995).

But the inhabitants were wreaking environmental destruction on their home. The trees that had been used for construction, fuel, canoes, and even rope were disappearing within several centuries of the Polynesians' arrival. By about 1400, the Easter Island palm, its largest tree and the source of its canoes, was extinct. Dolphins could no longer be hunted: they were too far out at sea. By the end of the fifteenth century, the forest was entirely gone. With the destruction of the native vegetation, the soil was eroded and crop yields declined. Every single species of resident bird was wiped out, and more than half the seabird nesting sites with them. The size of the population fell and social order succumbed to lawlessness and thuggery. Even cannibalism ensued. By about 1700, the population was declining to a mere fraction (1/4th to 1/10th) of what it had been at its peak. When the first Europeans arrived, on Easter of 1722 (giving the island its European name; Rapa Nui is the Polynesian name), they found an impoverished population eking out a subsistence living on a treeless desert island (Diamond, 1995).

As Jared Diamond (1995) has put it, "Easter Island is Earth writ small." Our own usually unobtrusive pace of environmental destruction and climate alteration makes people unaware of -- or, when aware, all too often indifferent to, or even contemptuous of -- the potential damage of their activities. But unlike the Easter Islanders, who would not "have noticed the felling of the last small palm" (Diamond, 1995), scientists have repeatedly warned that catastrophe awaits us unless we act now to curtail the reckless exploitation of our planet.

[Note: Since this section was written in 2004, Jared Diamond's book, Collapse: How Societies Choose to Fail or Succeed, has been published. The book has received excellent reviews in both Science and Nature. Though I have not yet had time to read it myself, I have enough familiarity with Diamond's work that I feel comfortable in highly recommending his book to the reader.]

The crust of the Earth has been compared to the skin on the top of a pudding, and ourselves to microbes that live on the skin's surface. Just as the microbes go about their business undisturbed by the nature of the skin and the pudding, so we also live on the Earth, taking it for granted, and largely unaware of its existence.

Below us, and around us, the Earth is in ceaseless motion. For the most part this motion occurs with extraordinary slowness, as the great plates of continents and oceans move about the planet's surface and jostle each other at the rate that fingernails grow. Only when earthquakes smash our buildings and other constructs, or volcanoes cover cities with ash (as Mount St. Helens did with Yakima, Washington, in its 1980 eruption), or floods drown cities, towns and farms (California, North Dakota, Danubian Europe, Germany) as in the late 20th century and early 21st, do we pay attention, at least for a bit. But these events typically do not affect us; usually they happen to others far away, intruding into our consciousness only by way of the morning newspaper or the evening television news.

Beneath our feet, however, and around us and overhead, the Earth goes about its own business, in as much ignorance and disregard for us as we have for it. But if we continue to abuse it, and in particular, if we continue to use the atmosphere as a sewer (as Stephen Schneider has put it), the Earth will suddenly and appallingly rise up against us, and we will pay a terrible, deadly price.

CONTINUE TO NEXT SECTION (PART II, Continued: Can science save us?)

RETURN TO CONTENTS