Storms of My Grandchildren (24 page)

Thus the estimated carbon dioxide forcing of 12 watts is, by itself, close to what is needed to account for the measured temperature change. Other long-lived greenhouse gases, specifically methane and nitrous oxide, are expected to augment the carbon dioxide forcing, that is, their atmospheric amount is likely to be greater when the planet is warmer. In addition, this is a good time to remind ourselves that the climate sensitivity of 3 degrees Celsius for doubled carbon dioxide was derived empirically from climate change in the late Cenozoic. At the warmest temperatures of the early Cenozoic it is likely that climate sensitivity was moving into a different, vitally important, climate regime, with higher climate sensitivity, as I will discuss in conjunction with data from the Paleocene-Eocene thermal maximum.

We must also note that the deep ocean temperature change defined by forams is not the same as global mean surface temperature change. The difference between the two must become large as the deep ocean temperature approaches the freezing point, because the deep ocean temperature does not go below the freezing point, while the surface continues to cool. However, even during the coldest increment in the entire Cenozoic curve (figure 18)—the time between the current interglacial period and the last ice age, when global surface temperature changed 5 degrees—there was a substantial deep ocean temperature change (3 degrees). But while global temperature change exceeded deep ocean temperature change in the late Cenozoic, the opposite is likely during the warm portion of the Cenozoic temperature curve. Why? Because high-latitude surface temperature change (which determines deep ocean temperature change) exceeds global mean temperature change—and ocean temperature was well away from the freezing point limitation in the early Cenozoic, so the amplified high-latitude temperature change was transmitted to the deep ocean. Thus overall, although there is necessarily uncertainty in the relation of global deep ocean temperature change and global surface temperature change, it appears that the total temperature changes over the Cenozoic era at the surface and in the deep ocean are comparable in magnitude.

Now let us turn to the question of why atmospheric carbon dioxide changed during the past 65 million years. First, note that the carbon dioxide causing the large climate changes in the Cenozoic era necessarily came from the solid Earth reservoirs (rocks or fossil fuels; see figure 15 on page 118). The alternative—oscillation of carbon among its surface reservoirs—is important for glacial-interglacial climate change, as a climate feedback, but it alters atmospheric carbon dioxide by only about 100 ppm, not 1,000 ppm.

The solid Earth is both a source of carbon dioxide for the surface reservoirs and a sink. The carbon dioxide source occurs at the edge of moving continental plates that “subduct” ocean crust. What does that mean? Continents are composed of relatively light material, typically granite. Ocean crust, that is, the solid Earth beneath the ocean water, is heavier rock, typically basalt. Both continents and ocean crust are lighter than material at greater depths, and they are slightly mobile because of convection deeper in the Earth. The energy that drives movement of the surface crust comes from the small amount of heat released by radioactive elements in Earth’s interior. As continents move, commonly at a rate of several centimeters per year (an inch or two), they can ride over ocean crust. Intense heat and pressure due to the overriding continent cause melting and metamorphism of the ocean crust, producing carbon dioxide and methane from calcium carbonate and organic sediments on the ocean floor. The gases come to the surface in volcanic eruptions and at seltzer springs and gas vents. This is the main source of carbon dioxide from the solid Earth to surface reservoirs.

The main carbon sink—that is, the return flow of carbon to the solid Earth—occurs via the weathering of rocks. Chemical reactions combine carbon dioxide and minerals, with the ingredients being carried by streams and rivers to the ocean and precipitated to the ocean floor as carbonate sediments. A smaller, but still important, carbon sink is the sedimentation of organic material in the ocean, lakes, and bogs. Some of this organic material eventually forms fossil fuels and methane hydrates.

A key point is that the solid Earth source and the solid Earth sink of carbon are not in general equal at a given time. The imbalance causes the atmospheric carbon dioxide amount to vary. The carbon dioxide source to the atmosphere is larger, for example, when continental drift is occurring over a region of carbon-rich ocean crust.

A qualitative explanation for the large Cenozoic climate change, and a picture of the solid Earth’s role in the Cenozoic carbon cycle, almost leaps out from figures 18 and 19. During the period between 60 and 50 million years ago, India was moving about 20 centimeters (8 inches) per year, which is unusually rapid for continental drift. India was heading north through an ocean region, now called the Indian Ocean, that had long been an area into which major rivers of the world had deposited carbon sediments. Undoubtedly, atmospheric carbon dioxide increased rapidly during that period as the carbon-rich sediments on that ocean floor were subducted beneath the Indian continental plate. Then, 50 million years ago, India crashed into Asia, with the Indian plate sliding under the Asian plate. The colliding continental plates began to push up the Himalayan mountains and Tibetan plateau, exposing a large amount of fresh rock for weathering. With India’s sojourn across the carbon-rich ocean completed, the carbon dioxide emissions declined and the planet began a long-term cooling trend.

A quantitative analysis of the Cenozoic atmospheric carbon dioxide history is carried out in our “Target CO2” paper described above. We calculated the range of carbon dioxide histories that can match the observed temperature curve (figure 18), accounting for uncertainties in the relation between the deep ocean and surface temperature. We estimated maximum carbon dioxide 50 million years ago as 1,400 ppm, with an uncertainty of about 500 ppm. The carbon dioxide amount 34 million years ago, when Antarctica became cold enough to harbor a large ice sheet, was found to be 450 ppm with an uncertainty of 100 ppm. This calculated carbon dioxide history falls within the broad range of estimates based on several indirect ways of measuring past carbon dioxide levels, as described in the “Target” CO2 paper.

A striking conclusion from this analysis is the value of carbon dioxide—only 450 ppm, with estimated uncertainty of 100 ppm—at which the transition occurs from no large ice sheet to a glaciated Antarctica. This has a clear, strong implication for what constitutes a dangerous level of atmospheric carbon dioxide. If humanity burns most of the fossil fuels, doubling or tripling the preindustrial carbon dioxide level, Earth will surely head toward the ice-free condition, with sea level 75 meters (250 feet) higher than today. It is difficult to say how long it will take for the melting to be complete, but once ice sheet disintegration gets well under way, it will be impossible to stop.

With carbon dioxide the dominant climate forcing, as it is today, it obviously would be exceedingly foolish and dangerous to allow carbon dioxide to approach 450 ppm.

What does the Cenozoic history tell us with regard to Administrator Griffin’s assertion that natural climate changes exceed human-made change?

Surely, nature changes carbon dioxide, and climate, by huge amounts. But we must look at time scales. The source of carbon dioxide emissions from the solid Earth to the surface reservoirs, when divided among the surface reservoirs, is a few ten thousandths of 1 ppm per year. The natural sink, weathering, has a similar magnitude. The natural source and sink can be out of balance, as when India was cruising through the Indian Ocean, by typically one ten thousandth of 1 ppm per year. In a million years such an imbalance changes atmospheric carbon dioxide by 100 ppm, a huge change.

But humans, by burning fossil fuels, are now increasing atmospheric carbon dioxide by 2 ppm per year. In other words, the human climate forcing is four orders of magnitude—ten thousand times—more powerful than the natural forcing. Humans are now in control of future climate, although I use the phrase “in control” loosely here.

Okay, I know, this is getting long, but for the sake of your children and grandchildren, let’s look a little more closely at another story in figure 18, one that is vitally important. I refer to the PETM, the Paleocene-Eocene thermal maximum, the rapid warming of at least 5 degrees Celsius that occurred about 55 million years ago and caused a minor rash of extinctions, mainly of marine species.

The PETM looks like an explosion in figure 18, and by paleoclimate standards it was explosively rapid. Carbon isotopes in the sediments deposited during the PETM show that there was a huge injection of light carbon into the atmosphere—about 3,000 gigatons of carbon, almost as much as the carbon in all of today’s oil, gas, and coal. It was injected in two bursts, each no more than a thousand years in duration.

The most likely source for such a rapid injection is methane hydrates. There is more than enough methane ice on continental shelves today to provide this amount of light carbon. The methane hydrate explanation is now broadly accepted, but it leaves open a vital question: What instigated the release of this methane? Was it an “external” trigger or a climate feedback? The answer holds enormous consequences for the future of humanity.

If the trigger for the methane hydrate release was external, such as the intrusion of hot magma from below or an asteroid crashing into the Arctic Ocean, then humans have no influence on whether the process will happen again. And the chances are remote that another such external event would happen in a time frame that most humans would care about. There have been several PETM-like rapid warming events in the past 200 million years. At that frequency, the chance of one beginning in the next hundred years is less than 0.00001 percent.

On the other hand, if the PETM and PETM-like methane hydrate releases were feedbacks, that is, if a warming climate caused the melting of frozen methane, then it is a whole different ball game. In that case, it is practically a dead certainty that business-as-usual exploitation of all fossil fuels would cause today’s frozen methane to melt—it is only a question of how soon.

Unfortunately, paleoclimate data now unambiguously point to the methane releases being a feedback. If the PETM were an isolated case, that interpretation would be less certain. But it has been found that several PETM-like events in the Jurassic and Paleocene eras were, as with the PETM, “astronomically paced.” Huh? That means the spikes in global warming and light-carbon sediments occurred simultaneously with the warm phase of climate oscillations caused by perturbations of Earth’s orbit. In other words, the methane releases occurred at times of natural warming events

So, why do methane hydrates produce a huge amplifying feedback in a small number of cases, while most “astronomical” warmings show little or no evidence of methane hydrate amplification? That mercurial behavior, in fact, is exactly what is expected for methane hydrates.

The largest volume of methane hydrates is on continental shelves, in the top several hundred meters of ocean sediments, although a smaller volume exists in continental tundra. The marine methane hydrates form in coastal zones with high biologic productivity. A sufficient rain of organic material onto the ocean floor yields a low-oxygen environment in the sediments, which causes the bacterial degradation of organic matter to produce methane. If the temperature is right, the methane is frozen into hydrates.

If a warming occurs that is large enough to melt methane hydrate, each liter of melted hydrate expands into 160 liters of methane gas. A small methane release may dissolve in the ocean, but a large release can bubble to the surface. Methane is a strong greenhouse gas, and on a time scale of about a decade it is oxidized to carbon dioxide, which will continue to cause warming for centuries. If the warming is large enough, most of the methane hydrate on continental shelves may be melted, as seems to have been the case in the PETM.

If Earth’s methane hydrate inventory is suddenly discharged, as during the PETM event, it requires several million years to fully reload the planet’s methane hydrate gun. Thus the next light-carbon methane hydrate event in the Paleocene, about 2 million years after the PETM, was only about half the strength of the PETM. This half-PETM was followed by still weaker and more frequent light-carbon warming spikes. These events occurred in conjunction with astronomical warming peaks during the time Earth was on its track toward peak warmth 50 million years ago, which suggests that the warmer Earth made the melting of hydrates easier and did not allow the hydrate reservoir to return to pre-PETM size.

Today, following global cooling over tens of millions of years, the methane hydrate reservoir is fully charged. The size of the hydrate reservoir is difficult to determine from spotty field data. However, methane hydrate models that are consistent with the limited data suggest a total inventory of about 5,000 gigatons of carbon in the form of methane ice and methane bubbles. Thus, unfortunately, not only is the methane gun now fully loaded, but it also has a charge larger than the one that existed prior to the PETM blast.

Let’s not jump to conclusions, however. We must glean more from the PETM before we discuss the likely fate of today’s frozen methane. Comparisons of the timing of carbon and temperature changes at many ocean sites show that a dramatic change in ocean circulation occurred at the time of the rapid PETM increases of light carbon and temperature. The ocean circulation change indicates that the main location where dense surface water sank toward the ocean bottom shifted from the region around Antarctica to middle latitudes in the northern hemisphere. Sinking water at the new location was also dense, but warmer and saltier. It is likely that this warmer water instigated the melting of methane hydrates. The methane, and carbon dioxide that formed as methane oxidized, provided an amplifying feedback that resulted in the large PETM spike in global temperature. Why ocean circulation changed is uncertain, but it is likely related to the global warming of 2 to 3 degrees Celsius that occurred just prior to the PETM event (figure 18).