Enlarge / Indonesia’s Paluweh volcano sits in a region of colliding tectonic plates—perhaps the formula for glacial periods in geologic history.

Generally speaking, it’s easy enough to make sense of the last few million years of climate patterns—the world looked much as it does today, so changes in greenhouse gas concentrations or ocean circulation can be related to what we see now. But as you go back farther in time, you can find very different climates and a rearranged map of continents, and those require more creative thinking.

For example, the ice age periods in the recent past are not unique. But most of the last 500 million years have been much warmer—what has caused the climate to slowly drift toward warmer or cooler temperatures over millions of years?


In the grand sweep of Earth history, its climate has remained within a habitable temperature range—thanks in part to the moderating influence of feedback loops within the system. The weathering of silicate minerals in bedrock pulls CO2 out of the atmosphere, for example. In a warming climate, weathering can speed up, removing more greenhouse gas and stabilizing temperatures. Cool the planet and weathering slows, giving greenhouse gases more chance to accumulate.

(This plays out on a geologic timescale—it’s not going to help us in the next century. But it’s an important long-term thermostat.)

But the Earth has been like a house full of passive-aggressive roommates, with the set point on the thermostat seeming to change periodically. Invoking geology to explain these large-scale changes more or less comes down to two options: either the release of CO2 by volcanic activity changes, or something strengthens the processes sucking CO2 out of the atmosphere. In recent years, researchers have been comparing these things to the pattern of climate changes to see how they correlate.

A new study by Francis Macdonald at the University of California, Santa Barbara, makes a strong case for plate tectonics fiddling with the thermostat through collisions between lines of volcanic islands and larger continents—especially in the tropics.

The idea isn’t new. Events like these are good candidates for climate movers-and-shakers. When one of these “volcanic arcs” collides with a continent, the volcanoes die and cease their release of CO2. But more importantly, some of their igneous rock gets smashed up onto the continent. The researchers say that this exposed igneous rock is extremely effective at removing CO2 when it weathers. The tectonic squeezing continues pushing up the rock, too, preventing it from being buried and keeping it exposed to weathering.

In the tropics, the warmer temperatures and plentiful rainwater are ideal for maximizing the weathering process—even during glacial periods when vast ice sheets are found closer to the poles. The number of these collisions at any point in time—and their location—is pretty much a bumper-cars accident of plate tectonics. So if a couple large collisions take place in the tropics for a few million years, the effect on the amount of CO2 in the atmosphere could be considerable.

Maps of colliding volcanic arcs and continents 0, 250, 300, and 445 million years ago. Modern locations are shown in the top left map.
Enlarge / Maps of colliding volcanic arcs and continents 0, 250, 300, and 445 million years ago. Modern locations are shown in the top left map.

To test the idea across a good sample of Earth’s history, the researchers put together a map of reconstructed plate tectonic collisions (sometimes called “sutures,” because the colliding rocks will stick together) going back to the Cambrian period 540 million years ago. They compared this to the recorded evidence of ice sheets, noting how close to the equator ice coverage extended.

Strong correlation

Their hypothesis came out looking good. The correlation between volcanic-arc-and-continent collisions in the tropics and ice coverage is very good. The biggest glacial periods line up with times when these collisions stretched over 10,000 to 14,000 kilometers, exposing quite a lot of weatherable rock. At other, warmer times, the length of these tropical collisions dropped to 4,000 kilometers or less.

If you include collisions outside the tropics, the correlation is a little weaker. That’s not surprising, as rocks exposed by high-latitude collisions wouldn’t weather as rapidly.

The researchers argue that the overall picture supports the idea that removing CO2 through weathering is more important than changing the release of CO2 from volcanoes. If the volcanic release was the key, the global correlation would be at least as good as the correlation in the tropics, because it doesn’t matter where CO2 gets released.

Timelines of plate tectonic collisions ("sutures") and ice sheet extent. The top graph shows the total length of global collisions, while the middle graph shows collisions in the tropics only.

Timelines of plate tectonic collisions (“sutures”) and ice sheet extent. The top graph shows the total length of global collisions, while the middle graph shows collisions in the tropics only.

The researchers similarly compared ice coverage to a couple different types of volcanic activity. There’s some correlation with the length of continental lines of volcanoes (like the modern Andes or Cascades), but it’s not nearly as good. Another idea is that rare but massive eruptions—like the Deccan Traps lavas that cover much of what is now India—have spit up huge amounts of weatherable rock in the tropics. But the correlation with that type of activity actually looks pretty poor in this analysis.

This leaves the researchers with a tidy explanation for the really long-term climate swings in Earth’s past. “[O]ur model accounts for both the initiation and termination of ice ages,” they write. “This pattern has repeated at least three times throughout the [last 540 million years]—when there have been abundant tracts [of weatherable rock] being exhumed and eroded in the tropics, the Earth has been in a glacial climate state, and when not, the Earth has been in a non-glacial climate state.”

Science, 2019. DOI: 10.1126/science.aav5300 (About DOIs).

Source link



Please enter your comment!
Please enter your name here