For Additional Information:
Dr. John Christy, (256) 961-7763
Dr. Roy Spencer, (256) 961-7960
Phillip Gentry, (256) 824-6420
EMBARGOED: For release Thursday, Aug. 11, 2005
California group's answer to climate puzzler improves the accuracy of global climate data
A curious puzzle in the study of climate science has been solved, and that solution is helping scientists at The University of Alabama in Huntsville (UAH) make the satellite record of global climate change more reliable than it was previously.
Research published this week by Carl Mears and Frank Wentz of Remote Sensing Systems (RSS) in Santa Rosa, Calif., identifies a problem that kept the UAH group from accurately correcting one error caused by NOAA satellites drifting in their orbits over the past 26 years.
The net result of changes in how the data are analyzed added about 0.09 C (about 0.16 degrees Fahrenheit) of global warming over the past 26 years, with most of that previously unreported warming occurring in the tropics.
"This work helps us produce a climate record that is even more reliable than it has been," said Dr. John Christy, director of UAH's Earth System Science Center.
Previously, the long-term (December 1978 through July 2005) climate trend in the UAH satellite dataset showed average global warming at the rate of about 0.88 C (about 1.58 degrees Fahrenheit) per century. The new trend, which includes the extra warming in the tropics, shows average global warming at the rate of about 1.23 C (about 2.21 degrees Fahrenheit) per century.
The research published this week by Mears et al includes an RSS satellite dataset showing a long-term warming trend of 1.9 C per century between 1978 and 2003, although that study only included data for the portions of the globe between 70 degrees south latitude (which excludes most of the Antarctic continent and ice shelves) and 82.5 degrees north latitude (near the north shore of Greenland).
When Christy and Spencer ran the UAH dataset for the same sub-global sample and the same time span, they found a long-term warming trend of 0.125 C per decade. The puzzle that Mears and Wentz solved was why climate data gathered by microwave sounding units aboard NOAA satellites agreed less with climate data gathered by balloons in the tropics than it did with balloon data collected in other parts of the globe.
"After we build the datasets we use the balloon data as a way to test the accuracy of the satellite data," Christy said. "To the best of my knowledge, we are the only group to do a rigorous, station-by-station, apples-to-apples comparison of the satellite data to other data collected in the same layer of the atmosphere, instead of trying to compare the satellite data to something artificial, like computer model forecasts."
Station-by-station comparisons assure independence, since there is no method that can be used to adjust the satellite data to match data records from dozens of different locations.
"We've been pleased with the level of agreement between the two datasets, indicating a high degree of consistency," said Dr. Roy Spencer, a principal research scientist at UAH. "Over the years, however, we noticed that the satellite data didn't seem to agree quite as well with balloon data from the tropics as it did with balloon data from other parts of the globe.
"But reliable balloon data from the tropics is scarce and difficult to access, so we couldn't find enough data points to really make a good test. We noted the relative differences in our published datasets, but couldn't explain what might cause those differences."
In a study published this week in the electronic journal "Science Express," the RSS team says the difference was caused by techniques used to correct for orbital drift by satellites that carry the temperature sensors.
These satellites are launched into a pole-to-pole orbit that carries them over the equator at the same local time each day. Over time, however, the satellites tend to drift eastward. Over a period of several years this drift causes each satellite to go over the sunny side of the Earth later and later in the afternoon, so the air temperatures it "sees" are at first warmer and then progressively cooler.
To correct for these false signals, the UAH team developed a technique using parts of the data collected earlier and later in the day.
As each satellite orbits from pole to pole, its temperature sensor spins around a north-south axis every 25 seconds. It takes eleven temperature "snapshots" along a wide west-to-east swath each time that spin turns the sensor toward the Earth's surface.
On the daylight side of the globe that means the first low-angle readings from each spin look westward at a part of the globe that is earlier in the day than the point directly under the satellite, while the last part of each scan looks toward the east at an area that is slightly later in the day.
(Think of a satellite going south-to-north along the Rocky Mountains in the Mountain Time Zone at noon. The first readings on the west side might pick up part of the Pacific Time Zone at 11 a.m., while the last readings on the east might see part of the Central Time Zone at 1 p.m.)
Because air temperature changes through the day -- the diurnal cycle -- Christy and Spencer were able to develop a technique in which they subtracted the differences between the cool temperatures on one side of the satellite from the warm temperatures on the other to determine how much they should correct the data due to orbital drift.
"We didn't take into account the fact that there is a diurnal change between the readings just on each side of the swath, especially in the tropics," Christy said. "That is what the RSS team found, that there are diurnal temperature changes between individual readings on each side of the satellite. We hadn't taken that into account."
The UAH team has now incorporated a diurnal cycle correction from direct satellite readings of the latest, high precision satellite instruments, into the past 26 years of satellite climate data and will provide that updated dataset to the public today. A set of global maps showing 26-year climate trends in both the corrected and uncorrected datasets is available on-line at: http://climate.uah.edu/
"While some people might question the importance of a correction that changes the long term trend by only 0.035 C per decade, for us the most important thing is to produce a climate dataset that is as accurate and reliable as humanly possible," Christy said. "Roy and I would argue that the political debate over climate issues should always be driven by the data, not the other way around."
"It is fairly obvious that some portion of this warming is probably due to human influences," said Spencer. "What isn't clear is how much or which influences."
In addition to rising levels of carbon dioxide, a greenhouse gas, in the atmosphere, other human influences on the climate include deforestation, urbanization, irrigation and the widespread use of chemicals that are or that release powerful greenhouse gases.
"And you have to overlay all of this onto the climate's natural instability," said Christy. "In an environment with so much uncertainty, getting accurate data becomes that much more critical."
In a recent scientific comparison, the UAH team found that the UAH satellite dataset agreed exceptionally well with several long-term temperature datasets prepared by NOAA, the National Centers for Environmental Prediction, the Hadley Centre of the British Meteorology Office, and the European Centre for Medium-Range Forecasts.
The newest satellite dataset correction doesn't reconcile differences between climate trends in the lower layer of the atmosphere and those reported in some surface sensor networks, a major source of controversy since Christy and Spencer published the first satellite-based climate record in 1990.
"The interesting thing is that so much of this warming correction is focused on the tropics, where the long-term warming trend changes by half of a degree Celsius per century," said Christy. "That is a significant amount of additional energy we now see in the tropical atmosphere, but it isn't enough to bring the tropical warming up to what we see on the surface.
"Despite the added warming the 'new' tropical troposphere warming trend is only half of the warming that is seen at the surface in the tropics," he said. "As a general rule, the climate models predict that the tropical troposphere should be warming 1.3 times faster than whatever the surface is doing. And it is only in the tropics that the surface and the troposphere don't seem to follow what the models forecast."
