The snows atop Kenya’s Mount Kilimanjaro-described by Ernest Hemingway as “unbelievably white in the Sun”-will be no more by 2020. Between 1912 and 2000, the ice fields have shrunk in area from 12 square kilometres to less than 2.4. And a similar pattern is unfolding with glaciers throughout the world.

“What you have throughout the Himalayas, and, in fact, throughout 46,000 glaciers that the Chinese are measuring, is a retreat of all glaciers in the recent past,” said Dr. Lonnie Thompson, a geophysicist from Ohio State University. Worse, he added, this rate of loss is accelerating, and, paradoxically, “under today’s conditions, all things being equal, the glaciers should be advancing-but they’re not.”

This means not only the loss of “the world’s watertowers,” as he termed them, but also of a valuable climate archive, which Thompson has worked to piece together for the last thirty years. He is based at Ohio State’s Byrd polar research centre.

On Monday, he delivered the 2005 J. Tuzo Wilson lecture to a packed Isabel Bader Theatre. Thompson has studied dozens of tropical glaciers, in Tibet, South America, and Africa, and is a world expert in the field.

His many expeditions consisted of drilling for-and bringing back-ice cores. These are long circular cross-sections cut through glaciers, each half a kilometre long. Examining an ice core is similar to studying tree rings, and offers a glimpse into the past. By looking at the amount and composition of trapped gases and dust in the ice, scientists glean clues about past climatic conditions.

Thompson described some of the challenges on these grueling forays. One such setting was Mount Huascaran, in Peru, the world’s highest tropical glacier. “Six tons of equipment goes up. Fifty-three days later, ten tons, including two frozen ice cores, come back down,” Thompson recounted.

In Tibet, the ice cores are cut up into pieces six metres long, each stuffed into a insulated box, and carried down the mountain on the backs of yaks. “You get 12 metres [of ice core] to a yak. We have quite a herd,” he quipped.

In his lecture, Thompson built up the case that something is amiss with our climate, drawing on his own glaciological work, and without assuming the answer.

He began with the observation that few ice core climate records extracted from tropical glaciers go back more than 25,000 years. In ice cores from Greenland and Antarctica, meanwhile, scientists have been able to peer back 400,000 years, or more.

To solve this riddle, Thompson drew on two observations: changes in the earth’s obliquity, which is the tilt of its axis from the horizontal; and precession-as the earth spins its axis, it also wobbles ever so slightly, like a spinning top. The former cycle repeats every 41,000 years; the latter every 21,000.

Though these wobbles in the earth’s rotation are very minute, they affect the longitude at which the so-called inter-tropical convergence zone falls-“rain and shine central” in a sense, the part of the planet that receives the most moisture and sun. For not only the height of the mountain determines whether a glacier will adorn its top; the amount of moisture, and hence snow, falling on it is critical.

Based on the length of ice core records of glaciers close to the equator, Thompson concluded that tropical glaciers shrink and grow, and disappear and reform cyclically.

“Asynchronous growth of low-latitude mountain glaciers may result from precession-driven changes in water vapour,” Thompson concluded. At present, the most rain falls 10° north of the equator-so glaciers in the northern hemisphere, such as Tibet, should be thickening and advancing. This happened in the early twentieth century, but the trend has since reversed.

“You see how unusual the last 20 years are-and that’s the same time as we’re losing all the glaciers,” he said.

The last significant climatic shift occurred about 5,200 years ago, and left an imprint in ice that Thompson and others have studied. Before then, levels of methane in the atmosphere, a greenhouse gas, were falling, but ice core samples from Mount Kilimanjaro show that methane levels turned around then, and began rising. Scientists are still struggling to explain the cause.

Some surmise this rise stemmed from agriculture. That period was roughly when the first cities were founded, and when agriculture was becoming more widespread. Dr. William Ruddiman, at the University of Virginia, has argued that the spread of agriculture released great amounts of carbon dioxide and methane gas in the atmosphere in the past.

He attributes the uptick in carbon dioxide, starting 8,000 years ago, to extensive deforestation in Europe and Asia, to make way for fields and pastures. A second uptick, this time in methane, roughly 5,000 years ago, may have been caused by the advent of “wet rice” farming, according to Ruddiman. So a first wave of human-caused climate change may have made civilization possible; now, scientists worry that a second bout might put an end to it.