Parents trying to determine the temperature of a feverish child may stick a thermometer under the tongue and under the arm and get slightly different measurements. Scientists trying to learn if the Earth's temperature is rising also rely on measurements taken at various places – the Earth's surface and its atmosphere - which also often produce different readings.

In both cases, the mere fact that the measurements in different places give slightly varied answers does not necessarily mean that one or the other is wrong. This issue is investigated in two papers published in the Feb. 18 issue of the journal Science.

"Scientists have puzzled over the difference between the findings from the satellites, which measured temperatures in the atmosphere, and those measurements that were obtained from the surface observing systems," said Dian Gaffen, a research meteorologist with NOAA's Air Resources Laboratory in Silver Spring, Md, lead author of one of the papers. Benjamin Santer of the Lawrence Livermore National Laboratory in Livermore, Calif., is the lead author on the other paper; Gaffen is a contributor. Both papers involve scientists at a number of different institutes.

"Our findings support the idea that the difference is largely real and that it may be due to a combination of natural and manmade factors," said Santer.

Gaffen and colleagues used data from radiosondes (weather balloons carrying instrument packages) to obtain independent data on both surface and lower-troposphere temperature change. The troposphere is the area of the atmosphere where weather occurs, extending about seven miles from the Earth's surface to the next layer, the stratosphere. Their findings show that the temperature trends are consistent with the satellite and surface results since 1979.

"Previous work with the satellite data suggested little or no temperature trend in the past twenty years, while the surface data show marked warming. We find that a third observing system, radiosondes, shows the same pattern as the satellite and surface data in the tropics, where the surface and tropospheric temperature change show the largest differences," said Gaffen. "The radiosondes also show that the tropical atmosphere has become slightly more unstable since 1979, when the satellite data began. But when we look further back, to 1960, we see more consistent warming at the surface and in the lower troposphere, which tells us that the past two decades might not be representative of longer-term changes."

In the companion paper, Santer and colleagues found that the discrepancies can partially be attributed to the fact that the surface observing system does not cover some parts of the globe, unlike the satellite system. Computer model simulations of the climate system indicate that the remaining differential cannot be explained by the natural variability of the climate system or by climate change resulting from increases in greenhouse gases alone.

The Santer et al. paper shows that different climate "forcing factors" may have had quite different influences on surface and atmospheric temperature. Increases in greenhouse gases probably act to warm the troposphere more than the surface. In contrast, the combined effects of aerosols from major volcanic eruptions (such as that of Mt. Pinatubo in 1991) and human-caused decreases in stratospheric ozone may act to cool the troposphere more than the surface.

Climate models have often been criticized for their failure to reproduce warming of the surface relative to the troposphere. Such criticism is usually based on model experiments involving changes in greenhouse gases alone. "Our work shows that the correspondence between modeled and observed temperature changes is much closer if the model experiments include a combination of human-caused and natural climate effects over the past twenty years, and not just changes in greenhouse gases," said Santer.

Similar issues were addressed in a recent report of the National Research Council. Gaffen and Santer were part of the 11-member panel that dealt with the issue of why the satellite data apparently show little or no warming of the troposphere, while ground-based thermometers indicated marked warming of the Earth's surface. The panel concluded that this discrepancy "in no way invalidates the conclusion that the Earth's temperature has been rising." The research described in both Science papers supports this conclusion.