Scientists are working to understand why the lower atmosphere isn't heating up as fast as some global warming models predict.

by Patrick L.Barry

Spencer and Dr. John Christy, a professor of atmospheric science at the University of Alabama in Huntsville, are trying to account for the unexpected temperature patterns. By explaining the contrary behaviour of atmospheric and surface-level temperature trends, they hope to improve computer models used to simulate the world's climate. This would provide a better picture of how severe or mild global warming will be over the next century.

"I believe the data bolster the traditional scientific scepticism one must have when discussing predictions of the future," Christy said.

A recent National Research Council report states: "For the time period from 1979-1998, it is estimated that on average, over the globe, surface temperature has increased by 0.25 to 0.4 degrees C and lower to mid-tropospheric temperature has increased by 0.0 to 0.2 degrees C." These are stated as ranges because of measurement uncertainties in each. Current climate models predict that the layer of the atmosphere called the "lower troposphere" -- which extends from the surface to an altitude of about 5 miles -- would be warming at a slightly faster rate than the surface.

Global Warming Ranges

But satellite measurements of temperatures in the lower troposphere over the last 21 years don't agree with that prediction. Collected by the National Oceanic and Atmospheric Administration's TIROS-N series of weather satellites, the data show only a slight net warming of 0.05 degrees Celsius per decade.

What could possibly be causing these unexpected trends? Right now, there are only theories.

"Stratospheric ozone depletion, unknown effects related to the major volcanic eruptions and the infrared effect of aerosols have all been bounced around as ideas, but none has had serious work done on them," Spencer said.

Some scientists suspect that the record of surface temperature warming has been exaggerated by the so-called "asphalt effect," creating unrealistically high expectations for the warming of the troposphere.

Thermometers used to calculate the average surface temperature are usually located in areas easily accessible by people. In industrial countries, the thermometers are most often found at airports. It is not clear what fraction of the observed warming of the Earth's surface is due to the influence of "urban heat islands" (Heat islands are created when city growth alters the urban fabric by substituting manmade asphalt roads and tar roofs and other features for forest growth) on the measurements.

"I believe there are still urban warming biases in the global thermometer record that are exaggerating the global warming signal," Spencer said. "I don't think it will be possible to remove these biases since virtually all thermometer sites have experienced changes in their microclimate due to (humanity's) activities."

The satellites, on the other hand, sweep over almost the entire globe as they take their measurements, covering about 95 percent of the Earth's surface. Oceans and continents, forests and factories are all incorporated into the satellite figures.

Most of the current work at GHCC focuses on improving and expanding the body of data to provide a clearer, more detailed picture of the long-term temperature patterns of the atmosphere.

For example, Christy is trying to expand the atmospheric temperature record to before 1979 -- which is when the first of the TIROS-N satellites was put in orbit -- by using data from radiosonde balloons. Potentially, the balloon data could extend the record back to the late 1950s.

Also, a new version of the temperature sensors used by the satellites will improve the detail of the measurements taken. The new Advanced Microwave Sounding Unit (AMSU) sensors can detect temperatures in the upper stratosphere, which is the atmospheric layer above the highest layer that the older sensors could measure. The new sensors can also distinguish between more sub-layers of the troposphere -- the layer where most weather occurs.

The first AMSU sensor was launched into space in May of 1998 aboard the NOAA-15 satellite, and data from that sensor are already being incorporated into the daily temperature updates produced by Spencer. Another AMSU sensor is scheduled to launch in late August aboard the NOAA-16 satellite, and the Aqua satellite and the European Space Agency's MetOp series of polar-orbiting satellites will also bear the sensors.

"AMSU sensors are what's going to be providing our temperature information from satellites for the foreseeable future," Spencer said.

The improved temperature record will guide efforts to refine computer models of the world's climate so that the behaviour of the models more closely resembles the observed behaviour of the atmosphere.

Is Forest being eclipsed by Roads?

Current models suffer from several shortcomings.

For example, clouds are not well represented by the models. The resolution of current models is too coarse for features as small as clouds, Spencer said. Yet clouds clearly play a crucial role in climate due to their influence on humidity, precipitation and albedo (the percentage of solar energy reflected back into space as light).

"The role of clouds is still regarded as one of the biggest uncertainties in global warming predictions," Spencer said.

The ability of plants to remove carbon dioxide from the atmosphere and the role of soils have only recently been added to the models, and scientists aren't confident yet of how the models portray these factors, Spencer said.

"While we know that vegetation takes up some of the carbon dioxide we generate from burning of fossil fuels, how that sink of carbon will change in the future is still pretty uncertain," Spencer said.

Climate models are also limited by the computing power available.

"The global models would be much better if computers were much faster," Spencer said. "Instead, a lot of approximations are made to make the models simple enough to do climate simulations over the whole globe.

"Unfortunately," Spencer continued, "we know that many of the processes that are crudely represented are quite non-linear, and so have the potential to respond in unexpected ways."