Evidence Against the Iris Hypothesis

  Page 1Page 3

“The Iris Hypothesis is very exciting,” states Bing Lin, an atmospheric research scientist at NASA LaRC. “Everybody would like to see tropical clouds changing in response to surface warming and acting to stabilize the climate system. The problem is when we used measurements from the Clouds and the Earth’s Radiant Energy System (CERES) sensor, we got significantly different results (from Lindzen).”

Copies of the CERES sensor fly aboard both the NASA/NASDA Tropical Rainfall Measuring Mission (TRMM), launched in November 1997, and NASA’s Terra satellite, launched in December 1999. Additional CERES sensors will be launched aboard Terra’s sister ship, Aqua, in the spring of 2002. CERES is the most advanced space-based sensor ever launched for measuring Earth’s radiant energy fluxes on a global scale.


Map of Net Flux

Lin’s team took the measurements made every day by CERES over the tropical oceans and plugged them into the same model that Lindzen used. Instead of the strong negative feedback that Lindzen’s team found, Lin’s team found a weak positive feedback (Lin et al. 2001). That is, Lin found that clouds in the tropics do change in response to warmer sea surface temperatures, but that the cloud changes serve to slightly enhance warming at the surface. Specifically, whereas Lindzen’s experiment predicts that cirrus clouds change in extent to reduce warming at the surface by anywhere from 0.45 to 1.1 degrees, Lin’s experiment predicts that changes in the tropical clouds will help warm the surface by anywhere from 0.05 to 0.1 degree (Lin et al. 2001).

The difference between the two experiments can be summed up as follows. According to the Iris Hypothesis, for each square meter of tropical cloudy, moist area that disappears with increasing surface temperature, 70 watts of heat is lost from the planet—like turning off a 70 watt light bulb for every square meter of area. But CERES’ measurements of cloud properties tell a very different story—clouds are much more reflective (51 percent instead of 35 percent) and somewhat weaker in their greenhouse effect than Lindzen’s model predicts. So instead of turning off a 70-watt bulb for each square meter affected, it is as if a small 2-watt night-light bulb was turned on in every square meter. Hence, the slight warming found by Lin’s team instead of the very large cooling found by Lindzen’s team.

Lin says the reason his team’s findings differ so dramatically is because some of the initial assumptions made in Lindzen’s model are incorrect. He says that while he has many minor differences of opinion with Lindzen on this subject, he has three major disagreements. For starters, he says, the Indo-Pacific warm pool region does not serve as a model for the tropics all around the world. The waters there are, on average, much warmer than the rest of the tropics and so convection (warm, upward-moving air) is much stronger. Therefore, the area covered by deep convective cumulus clouds (thunderheads, basically) and cirrus clouds is not the same throughout the tropics. In the Indo-Pacific warm pool, these two cloud types cover about 13 percent of the region, whereas they only cover about 10 percent of the world’s tropics on a global scale (Lin et al. 2001). Lin’s team found that while tropical cloudiness does change as sea surface temperature changes, there is a large reduction in total cloud amount—roughly 10 percent cloud cover as compared to the 22 percent proposed by Lindzen’s team.

  The balance of energy the Earth receives from the sun versus the amount emitted as heat from the Earth is measured by the Clouds and the Earth’s Radiant Energy System (CERES). Scientists used these data, combined with estimates of cloud type derived from geostationary weather satellite data, to evaluate the results predicted by the Iris Hypothesis. The image at left shows net energy flux for February 2002 measured by CERES. The amount of energy retained within the Earth system ultimately determines its average global temperature. (Image by Reto Stöckli, based on data from the and CERES science team.)

Diagram of Clouds Energy Balance


Secondly, Lin disagrees with Lindzen’s proposed physical model of the clouds themselves. “Deep convective clouds very strongly reflect sunlight back to space,” he states, “but their relative area of coverage is small.” Cirrus clouds, on the other hand, are very extensive and cover large areas. They can be thin enough to allow sunlight to pass through, or they can also have a high reflectivity. Cirrus provides a much larger “canopy” over the tropics so, from a radiative perspective, those clouds are actually more important than deep cumulus clouds.

The third major disagreement between Lin’s and Lindzen’s experiments pertains to the amount of heat escaping from cloudy regions. CERES measurements reveal that 155 Watts per square meter escaped the atmosphere over cloudy, moist regions, which is significantly more than the 138 Watts per square meter that Lindzen’s team assumed (Lin et al. 2001).


Different types of clouds have different effects on the balance of energy received and emitted by the Earth. In areas covered by the cumulus towers of a thunderstorm’s convective core (left) almost all the Sun’s energy is reflected. The cold cloud tops radiate very little energy out into space. Cirrus clouds (the cloudy and moist region, center), on the other hand, reflect some shortwave energy, but let some through to the surface. Likewise, they emit some heat (longwave energy) but redirect some back to the surface. Clear and dry regions (right) are almost the inverse of convective cores— most of the solar energy is absorbed by the surface, much of which is eventually emitted as thermal infrared radiation back out to space. In the clear regions, reflected energy increases as low level clouds increase, while as humidity increases less longwave energy is emitted. (Image by Robert Simmon)

Table of Lindzen's versus Chambers and Lin's results

In summary, Lindzen’s team suggests that higher sea surface temperatures lead to less cloudy, moist skies and a corresponding increase in clear, dry skies. Lin disagrees with Lindzen’s interpretation of the cloud physics. In their paper, Lin’s team wrote that the much smaller albedo and lower outgoing heat flux assumed by Lindzen exaggerated the cooling effects of the outgoing radiation over cloudy, moist regions while minimizing the warming effects of incoming sunlight through regions covered by cirrus (Lin et al. 2001). Based upon CERES data, Lin’s team concluded that the reduction in cloudy, moist skies allows extra sunlight to warm the surface by up to 1.8 Watts per square meter—a small but positive net energy flux (Lin et al. 2001).

“Our results are based upon actual observations that are used to drive global climate models,” Lin concludes. “And when we use actual observations from CERES we find that the Iris Hypothesis won’t work.”

next Reconciling the Differences
back Does the Earth Have an Iris Analog?

In the graphic on the left, “L” refers to Lindzen’s team and “C” refers to Chambers and Lin. Both teams used the same equations to predict climate change, but they used different data sources and made different assumptions for the values of some variables that model the behavior of clouds. The table at left shows the contrasting values used by the teams. The most important differences were in the cloudy and moist region. Lindzen et al. used an albedo of 0.35 while Chambers et al. used an albedo of 0.47. Values of net flux for the region were 123 W/m2 for Lindzen and 46 W/m2 for Chambers. (Table by Robert Simmon)