Scientists Explain How They Attribute Climate-Change Data

12 May 2006Sharon Begley

For laypeople, one of the toughest things about science isn't understanding what science knows, but grasping how it knows what it knows. The process seems like a black box. Questions go in one end, answers mysteriously emerge from the other.

Case in point: climate change.

Earth has warmed 1.4° Fahrenheit over the past 100 years. Skeptics concede that. But they scoff at the claim that the warming is caused by greenhouse gases from, in particular, the burning of coal, oil and gas, contending that it is more likely due to the climate system's natural variability or to natural "forcings," such as a hotter sun. After all, glaciers and arctic sea ice melted long before the first smokestack or tailpipe existed.

True. But as climatologists refine "detection and attribution" studies, they are getting better at discerning the fingerprints of changes that are so physically or statistically anomalous that they couldn't be natural. "Different factors that affect climate -- human or natural -- have unique signatures," says climatologist Ben Santer of Lawrence Livermore National Lab, California.

One signature is the pattern of warming in the atmosphere. Decades of data from satellites and weather balloons show that the lower atmosphere, or troposphere, has warmed while the upper atmosphere, or stratosphere, has cooled. "If you turn up the sun's energy output, the atmosphere should warm from the stratosphere to the surface," says Dr. Santer. "That's contrary to what's observed. But greenhouse gases do not produce a uniform warming. They warm the troposphere and cool the stratosphere."

Another signature is the pattern of warming in the seas. Some 84% of the total heating of Earth over the past 40 years has gone into warming oceans about 1°. (The 84% comes from calculating the heat needed to melt glaciers and warm air and water; oceans, being so huge, suck up most of the heat.) In theory, a hotter sun could be the culprit. But the sun has increased its energy output less than 0.1% over that time, according to satellite data. That isn't enough to explain even a few percent of the warming, says marine scientist Tim Barnett of Scripps Institution of Oceanography, La Jolla, Calif., who led a 2005 study on ocean warming, published in Science.

Could the warming reflect natural oscillations? In that case only some oceans would warm, since if you add heat to one ocean you have to take it from another. "For all the oceans to warm, you need to add heat to the system, which rules out natural variability" on thermodynamics grounds, says Dr. Barnett.

Could the warming come from natural, geothermal heat? Only if the warming were greater on the ocean floor, says Dr. Barnett. It's not.

Scientists do consider the possibility that "stuff happens" -- that natural variability in the climate system somehow explains observed changes. To do that, they comb through eons of data on how much and how often the climate has changed thanks to such sources of natural variability as El Niño, the North Atlantic oscillation and the Pacific decadal oscillation.

It's like observing a zillion poker hands and counting how often players are dealt a flush in five-card stud. Once you know that probability (0.002), you get suspicious if someone is dealt two flushes in a row (probability 0.000004). It might have been a fair deal, but the numbers suggest otherwise.

That's why, with glaciers and sea ice melting and rainfall patterns shifting, scientists smell a stacked climate deck. "We have never seen natural variability on a global scale like we've had in the last 100 years," says atmospheric physicist Michael Oppenheimer of Princeton University.

A study released last week also found evidence of a stacked deck. Scientists analyzed historical barometric data to infer the strength of winds across the tropical Pacific. This "Walker circulation" consists of high-altitude westerlies and surface easterlies. It has weakened by 3.5% since the mid-1800s, Gabriel Vecchi of the Geophysical Fluid Dynamics Laboratory, part of the National Oceanic and Atmospheric Administration, and colleagues reported in Nature. That has yet-unknown consequences for marine life and regional rainfall.

Again, changes in the sun's output since 1861 are too small to have warmed the world enough to weaken the Walker circulation that much, the scientists calculate. Adds Dr. Vecchi, "We looked at 2,000 years of data and asked whether internal variability could produce the weakening. There is less than a 1% chance it did."

The debate over what has caused the increase in severe hurricanes centers on whether they're just something Earth kicks up from time to time or are the result of seas warmed by anthropogenic climate change. In a study presented at an American Meteorological Society conference, scientists noted warming in every ocean basin where hurricanes form. Natural variations tend to hit one basin at a time.

Researchers report in the Journal of Climate that the tropical Atlantic where hurricanes originate warmed several tenths of a degree Celsius in the 20th century. "Natural variability would have to be more than two times higher than we've seen to account for our results," says NOAA's Thomas Knutson.