Revisiting the Iris Effect

A study published last summer by MIT’s Richard Lindzen and Yong-Sang Choi showed a curious thing: 15 years of observations of the tropics revealed that the Earth responds to rising sea surface temperatures by dumping more radiation to space. According to the authors, this feedback — dubbed the “iris effect” — would prevent much of the expected warming due to man-made greenhouse gas emissions. Further, Lindzen and Choi suggested that 11 major climate models had gotten this effect wrong. The paper found the Earth’s climate not nearly as sensitive to global warming as commonly thought.

A number of scientists, at NASA and elsewhere, immediately wondered how they could have missed such a major discrepancy between models and real-world observations. Their response to the paper provides an example of the back-and-forth, checking and re-checking that makes up the scientific process.

NASA Langley’s Takmeng Wong was one of the scientists surprised by the Lindzen and Choi paper. Wong works with data from the Clouds and the Earth’s Radiant Energy System (CERES) on NASA’s Terra satellite and from the Earth Radiation Budget Experiment (ERBE) satellite. Lindzen and Choi used ERBE data for their analysis, so Wong was naturally interested.

“When you see a surprising result, the first thing you do is go to the paper and see what they have,” Wong said. “We tried to do that and to reproduce their results. It’s part of the scientific process.” Being able to reproduce a specific result is an important building block of scientific knowledge.

Wong and Colorado State University’s Chris O’Dell eventually teamed up with Kevin Trenberth and John Fasullo, climate scientists at the National Center for Atmospheric Research, to publish a response. Their examination uncovered a number of deficiencies in Lindzen and Choi’s method, and found their result to be unstable and fragile. That analysis appeared this month in Geophysical Research Letters.

“We went to the same model data, to see if the observations are going one way and the models are going the other way,” Wong said. “When the analysis is done properly using robust scientific technique, what you find is that the observations and the models are consistent to within the uncertainty of the data.”

Wong summarized a few basic problems that led to the surprising finding: 

1. Lindzen and Choi focused on a number of selective time periods. But if the beginning and end points of those time periods are adjusted only slightly, their result falls apart.

2. The paper also treated the tropics as a closed system. In other words, it did not take into account any outside influences on what was happening in the tropics, such as the large amount of energy transport moving in and out of the tropics on ocean currents and atmospheric waves during events such as El Nino and La Nina.“The tropics is not a closed system,” Wong said. “But they treated it as such in the study.”

3. Lindzen and Choi took their result from the tropics and applied it globally, instead of using global data to study the link between global temperature increases and global outgoing radiation to space.