12.2. The number of weeks per year that sea ice jammed northern Iceland ports increased for several centuries but then dropped sharply during the 1900s.

But records of Icelandic sea ice and alpine glaciers speak for only a small part of Earth's surface, and local records may not represent the larger picture. For example, the winter of 1976—1977 was very cold by normal eastern United States standards, with harbors choked with ice south to New York and beyond. Two decades later, the winter of 1996 was the snowiest of the century for many eastern states. Yet these cold, snowy winters occurred during an interval when, on average, the planet has been gradually warming to unusually high levels. Even though the East Coast was having a few old-time winters (the kind your grandfather talked about), the planet as a whole was warming. Records from one or two regions do not necessarily reveal the larger picture.

As a result, some climate scientists expressed doubts about whether or not there really was such a thing as a Little Ice Age (or a medieval climatic optimum) on a global basis. This kind of challenge could be answered only by looking at detailed records of temperature changes from many regions. Unfortunately, highresolution historical records that stretch back many centuries (like the one from Iceland) are extremely rare. So instead, climate scientists used techniques to extract climatic signals from natural archives.

The most common natural archive of climate change is the width of rings added each year by long-lived trees. In regions where summers are generally cold and somewhat inhospitable, trees add thicker rings during favorable conditions in slightly warmer (or wetter) years. Scientists spend summers in mosquito-infested areas of the Eurasian and American Arctic finding these long-lived trees and (non-destructively) extracting pencil-thin cores. Back in the lab, they count the rings to date each year of growth history and measure the ring thickness (or other properties) to extract records of yearly temperature change during the life span of the trees. Records like these now exist from far-northern latitudes all around the Arctic.

Other kinds of records of annual temperature change are available from the midlatitudes and tropics. At very high altitudes, annually deposited layers of snow harden into mountain-glacier ice that contains many kinds of climatic records at


12.3. Reconstructions of Northern Hemisphere temperature during the last 1,000 years show an erratic cooling trend through the late 1800s followed by an abrupt warming during the 1900s. The light shading indicates uncertainties.


12.3. Reconstructions of Northern Hemisphere temperature during the last 1,000 years show an erratic cooling trend through the late 1800s followed by an abrupt warming during the 1900s. The light shading indicates uncertainties.

yearly or near-yearly resolution. Hardy scientists (like the legendary Lonnie Thompson) climb to altitudes of 6,000 meters (20,000 feet) or more, elevations near those familiar to professional mountain climbers, to drill cores into the ice. On the way in, they carry all their equipment and food. On the way out, they add hundreds of meters of newly drilled ice cores to their loads.

Not all of the work done to obtain annual records of climate involves such hardships. In the tropical oceans, some kinds of corals build their reef structures in annual layers that can be cored in warm, shallow, aquamarine waters. But most of the reefs that lie within easy reach of comfortable hotels have been cored by now, so that even this work requires travel in small planes to isolated islands that have few luxuries and in some cases host exotic tropical diseases.

From these natural archives, scientists have obtained climatic records many hundreds of years long at dozens of sites, mostly in the Northern Hemisphere. In 1999 atmospheric scientist Mike Mann and colleagues Ray Bradley and Malcolm Hughes used mathematical techniques to extract a record of estimated temperature change in the Northern Hemisphere over the last 1,000 years. The reconstructed trend (fig. 12.3) is known as the "hockey stick." The gradual cooling from AD 1000 through roughly 1900 is the handle of a hockey stick, and the much faster warming from 1900 through 2000 is its blade. Both the handle and the blade show significant shorter-term temperature oscillations, but the overall trend (at a quick squint) is much like a hockey stick. Many other reconstructions using records from different combinations of sites show the same basic trend.

At least in a broad sense, the reconstruction in figure 12.3 shows that the medieval interval around 1000 to 1200 was relatively warm compared to the gradual and somewhat erratic cooling in subsequent centuries. No one time stands out as the clear start of the Little Ice Age, but temperatures were certainly cooler by 1600-1900 than they had been earlier. If the warmer centuries of 1000-1200 are used as a baseline, the average temperature had cooled 0.1°C to 0.2°C by 1600 to 1900, and a few extreme decades showed changes as large as 0.4°C.

What could have caused this gradual, oscillating cooling trend after AD 1200? One factor might be the slow cooling of far-northern regions in response to orbital changes in solar radiation (chapter 10). Some parts of the Arctic have cooled by 1 to 2°C over 9,000 years. At that average rate, the cooling over 900 years would be 0.1 to 0.2°C, just about what the curve shows. But this cannot be the main explanation: the average drop in temperature across the entire Northern Hemisphere would have been much smaller than the amount at high latitudes (where responses are amplified), probably well under 0.1°C. Other factors must also be contributing to this cooling.

One likely factor is large volcanic explosions, which inject small amounts of sulfur dioxide more than 15 kilometers high into the atmosphere, well above the tops of clouds. The sulfur gas reacts with water vapor already present in the atmosphere and forms sulfuric-acid droplets that reflect some incoming solar radiation and keep it from reaching Earth's surface. As a result, the climate cools. Temperatures remain cooler for the two years or so it takes Earth's gravity to pull the tiny particles back into the lower atmosphere, where snows and rains can quickly wash them out. One eruption can cool climate noticeably for a year or two, but several eruptions bunched within a few years are required to keep climate cool for as long as a decade.

Tropical eruptions can cool the entire planet by spreading sulfur into both hemispheres, but eruptions to the north or the south of the tropics cool only the hemisphere in which they occur. Ben Franklin, taking note of the colorful sunsets and unusually cool summer temperatures in 1784 just after an eruption in Iceland, realized that volcanic explosions might have a cooling effect on climate. After the 1883 eruption of Tambora, crops failed in New England because of late and early freezes, and the locals referred to that year as "Eighteen Hundred and Froze to Death." Many of these earlier eruptions were much larger than the 1992 Mount Pinatubo eruption in the Philippines that gave us one summer of beautiful sunsets and cooled global climate by 0.3°C for a year.

The second factor proposed to affect climate over centuries or decades is the changing strength of the Sun (not to be confused with changes in the distribution of solar radiation caused by variations in Earth's orbit). Satellite measurements since 1981 have shown very small changes in solar radiation occurring with the same timing as changes in the number of dark spots (sunspots) on the Sun's surface. Although the dark spots reduce the amount of radiation streaming away from the Sun in the small areas where they occur, the overall relationship is just the reverse: the Sun emits more radiation when sunspots are most common. Stronger activity on other parts of its surface more than counteracts the localized effect of sunspots in reducing emissions.

Both the number of sunspots and the amount of solar radiation cycle from maximum to minimum size and back every 11 years. The 11-year cycle of sunspots is well recorded in telescope observations that go back into the 1700s, but almost no climatic records from Earth's surface show a convincing 11-year temperature cycle because the changes in radiation are so small. Yet this does not necessarily mean that changes in solar radiation are unimportant over longer intervals. Telescope observations from the Middle Ages show that clusters of 11-year sunspot cycles can vary greatly in amplitude, even disappearing altogether during intervals such as 1645 to 1715. Some (but not all) climate scientists believe that these intervals of less-frequent sunspots, and by inference weaker Sun strength, played a role in causing the cooler climates of the Little Ice Age.

What about greenhouse gases? What is their role in all this? Clearly the fast warming at the blade end of the hockey stick must at least partly be caused by greenhouse gases; both the rate and the amount of rise are unprecedented compared to the first 900 years of the record. Part 5 of this book will explore this recent warming trend. But what about the earlier record, and particularly the CO2 wiggles in figure 12.1? The most recent CO2 minimum (from 1500 to 1750) seems to match at least broadly the coolest intervals of the Little Ice Age, and higher CO2 values tended to occur during the slightly warmer medieval era between 1000 and 1200.

The conventional explanation for this correlation of temperature and CO2 is that changes in volcanic explosions and/or solar radiation are the "first cause" of both the Northern Hemisphere cooling and the CO2 decreases. In this view, CO2 values drop when Earth cools largely because of a basic law of chemistry: more CO2 can be absorbed in a cool ocean than in a warm one. If volcanic and solar changes cool Earth's climate, the cooler ocean takes CO2 from the atmosphere. If this view is correct, both the hemispheric temperature changes and the CO2 variations are simply two responses of the climate system to natural (solar and volcanic) changes.

But this plausible-sounding explanation has a serious flaw. Today's most advanced climate models are constructed with the goal of reproducing all the complex interactions among the atmosphere, ocean, land surfaces, vegetation, snow, and ice. These models attempt to simulate all of these interconnected responses, rather than analyzing them one-by-one in isolation. When scientists specify changes in the amount of solar radiation entering the climate system as a result of volcanic eruptions and/or changes in Sun strength, the models simulate the integrated response of the many parts of the climate system, including the changes in temperature and atmospheric CO2 levels.

Examining one of these modeling studies, I noticed a result that seemed to indicate a major flaw in the natural explanation. To match the largest (10 parts per million) decreases in CO2, the model had to cool climate by almost 1°C, yet the reconstructed temperature trend in figure 12.3 permitted a decrease of only 0.1-0.2°C between the warmer interval from 1000 to 1200 and the cooler one from 1500 to1750. Conversely, to stay within the bounds of the small temperature changes in this reconstruction, the models would allow CO2 changes of only 2—3 parts per million, compared to the 10 parts per million changes observed (fig. 12.1). Whichever way I looked at it, the CO2 changes were simply too large relative to the temperature changes. Something seemed to be seriously wrong with the conventional explanation.

The natural explanation also struck me as suspect for another reason. The rates of change during these CO2 drops looked very abrupt, in fact much faster than the natural changes that had occurred at the end of the last deglaciation, and during the previous three deglaciations as well. As shown in figure 9.1, CO2 concentrations during each of these deglaciations rose by almost 100 parts per million within a span of not much more than 5,000 years, for an average rate of about 2 parts per million per century. In contrast, the CO2 concentrations in figure 12.1 fell and then rebounded by 10 parts per million within only a century or so, a rate approximately five times as fast as the deglacial changes. Why would CO2 values change at a faster rate during a time of almost stable global climate than they had during the highly dynamic climatic changes that occurred at the end of the major glacial cycles? This didn't make any sense.

So once again, an explanation based on natural climatic changes had come up short. And again it seemed to me that the only solution to this apparent dead end must be an explanation lying outside the realm of "natural" causes, an explanation somehow linked to humans. The explanation had to lie in some kind of process that could reverse the slow deforestation and accompanying CO2 releases that had been occurring for thousands of years and could cause abrupt CO2 decreases that lasted for decades to a century or two.

Was this article helpful?

0 0

Post a comment