March 20, 2007

What Do We Know About Clouds?

Proponents of the catastrophic effects of global warming on the Earth’s climate often point (somewhat contradictorily) to either heat-related clear-sky drought or evidence of increased heavy rains when discussing global warming. Both of these phenomena would undoubtedly be closely linked to variations in cloudiness around world—either marked increases or marked decreases. The rather obvious question one might ask, based on these statements, is simply: “Have we actually observed changes in cloudiness around the world?” If so, we might have a relatively clear indicator of climate change. “Are there changes in cloudiness” would seem to be a rather simple question that can be answered in a straightforward manner.

Not so fast, say the authors of a recent 2007 study published in the respected science journal Geophysical Research Letters.

Amato Evan of the Cooperative Institute for Meteorological Satellite Studies at the University of Wisconsin-Madison and his colleagues critically examined the primary data source for cloud studies, the International Satellite Cloud Climatology Project’s (ISCCP) multi-decadal record of cloudiness. They conceived this project in part as response to the many studies (some published in such prestigious journals as Science) that have indicated the ISCCP cloud record demonstrates a widespread increase in surface solar heating, an associated decrease in the planet’s albedo, and that these trends linked to cloudiness may result from global warming.

Yet, Evan and his colleagues point out that one of the fundamental problems with ISCCP is that other records (specifically those involving surface observations and other satellite records) have simply not reflected these marked trends in total cloudiness seen in the ISCCP data. So in their article “Arguments against a physical long-term trend in global ISCCP cloud amounts,” the Wisconsin researchers took a very critical look at exactly how these well-regarded and well-used ISCCP data are computed and the potential problems inherent in using such data.

Before we can discuss the problems, it is necessary to give some background on this major cloud database that everybody uses. The International Satellite Cloud Climatology Project’s (ISCCP) multi-decadal record of cloudiness uses reflectance information from a series of geostationary satellites to create three-hourly maps of cloudiness that are then compiled for use in other projects such as numerical climate modeling. For the polar regions (not easily visible from the equatorial geostationary satellites), the project incorporates the microwave emittance data (from which cloud information can be extracted) taken from polar orbiting satellites. In their analysis of the ISCCP dataset, Evans and his colleagues meticulously examined all of the ISCCP data for extraneous influences. First, they removed any effects associated with the major climate phenomenon called El Nino/Southern Oscillation.

Then they set out to determine which regions of the world contribute most to these upward/downward global trends that have been reported in global cloudiness. In their words, “the most striking features [linked to these trends]…are the circular patterns centered in the Atlantic, Western Pacific and Eastern Pacific Oceans.” Intriguingly, these circular patterns correspond to the exact orbital locations of the geostationary satellites or “footprints” as Evans and his colleagues called them. Further analysis indicated that the observed cloud patterns (and consequently the trends) are the result of biases introduced by the orbital angles of the satellites. Specifically, as the viewing angle of the satellites is increased (in essence, ‘looking’ away from regions directly below the platform) a so-called “limb darkening” effect is created. What this means is that the satellite “sees” more clouds than actually exist for the more distant regions away from the satellite’s point over the Earth. This fact coupled with the realization that the geostationary satellites are located over ocean regions explains why many researchers had long maintained that the ISCCP time series consistently better represented ocean areas than land areas.

That was a troubling finding—large portions of the earth simply aren’t being measured correctly with regard to clouds. But was that all? Were there any other potential problems with the ISCCP data? The Wisconsin researchers re-emphasized their concern with the cloudiness data by determining if the changes in actual number of orbiting satellites over the length of record have had an impact on the overall trends in cloudiness. To do this, they first compared how precisely the changes in satellite geometries may be creating the ‘trends’ in the ISCCP data by creating a time series involving the satellite zenith angle. The satellite zenith angle is the angle between the local zenith and the line of sight to the satellite. They then examined that time series of angles in relation to specific cloud amount and found that major abrupt changes in cloud amount correlated with the introduction of new geostationary satellites or the repositioning of old satellites.

What was their basic “take-home” message? They found that if the spurious “limb darkening” effect is removed from the ISCCP dataset, “the ISCCP would exhibit no significant long term upward … and downward … trends in cloudiness.” And what does this mean for long-term climate studies? According the Wisconsin researchers, “Although the ISCCP data [are] very appropriate for many applications, clearly its use in global multi-decadal studies is troubling.”

Why is this the case? Why should we be concerned if trends in one of the most-used datasets on clouds are simply wrong? Clouds are some of the most basic meteorological phenomena in our earth/atmosphere system. Yet for all of their basic nature, researchers for many years have acknowledged that we do not have an adequate conceptual modeling representation of clouds in the many climate models that attempt to simulate global warming. This study by Evan and colleagues dramatically points out that we even have major difficulties in observing cloud from space over many years. As Evan and his colleagues warn, if we continue to use such flawed datasets to examine questions involving global warming, we may expect to encounter the infamous GIGO rule: “Garbage In, Garbage Out.” Fundamentally, it is very hard to reach good, viable conclusions about climate change when the basic input data are inherently bad.


Evan, A.T., A.K. Heidinger and D.J. Vimont, 2007. Arguments against a physical long-term trend in global ISCCP cloud amounts. Geophysical Research Letters, 34:LO4701.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress