Overview - Climate Change and Troposphere Temperatures

The evidence for global warming has been building steadily on a number of fronts for decades. But many questions remain. Until recently one of the more troublesome ones was the disagreement between tropospheric temperature measurements made since 1979 by NOAA satellites carrying Microwave Sounding Unit equipment, and the atmospheric temperature predictions of the best general circulation models. These satellites measure the brightness temperature of the lower atmosphere (that is, the temperature that corresponds to its total radiation emissions, most of which are at microwave frequencies). The best of state-of-the-art general circulation models of the earth/atmosphere system have consistently predicted that the lower to middle troposphere (the first 7 to 10 km above the earth's surface) should be warming at equal or greater rates than the surface. But for over two decades the expected degree of warming in this layer was not observed in these brightness temperature datasets. Attempts to independently validate these measurements with direct thermometer measurements by weather balloon borne radiosonde packages produced results that were also lower than expected in many regions. Skeptics claimed that this disproved global warming.

In fact, the issue is far more complicated. In 2000, the National Research Council published a pivotal review of the state of troposphere temperature trends at that time. Today, great strides have been made. While many questions still remain, the bulk of evidence is indicating that the observed discrepancies are at least partly spurious and related to known factors, including the short length of the data set relative to time scales for significant anthropogenic climate change, qualitative differences between the kind of temperatures modeled and those observed by satellites (e.g. level specific vs. tropospheric averaged), signal processing errors of various sorts by the onboard MSU equipment (not the least of which is the need to correct for gray body satellite equipment temperatures themselves), data reduction methods (which have improved considerably in the last 7 or 8 years), the methods by which vertical mixing, poleward advection of heat in lower latitudes, and ozone depletion are accounted for in the models. Other factors like aerosol cooling, volcanic activity, and short term decadal variations not related to long term climate change (e.g. El Nino Southern Oscillation) also play a role.

Another factor that has emerged as a significant contributor to the discrepancy is the fact that MSU devices do not measure the troposphere only - they receive a small, but important contribution from the stratosphere. This layer is known to have cooled over the period when satellite measurements have been available, and for known reasons that do not conflict with model predictions (the most important being another human induced problem - ozone depletion). In the last 5 to 10 years metadata and analysis techniques improved considerably and the gap between observation and expectation decreased. General circulation models improved and began yielding much better characterizations of climate behavior. As they improved, predictions generally moved toward rather than away from observations and continue to do so to this day.

Adding to the confusion was the fact that several analyses of the data from these satellites yielded differing results. Of these only one, from a team at the University of Alabama, Huntsville (UAH) obtained results that disagreed with climate model predictions. Not surprisingly, these results were seized upon by industry and Far-Right interests as refuting global warming to the complete exclusion of all other datasets. The UAH products were thorough and carefully prepared—no one doubted the quality of that team’s work, then or now. But in summer of 2005 the journal Science published research demonstrating that the UAH analyses contained a math error in their corrections for satellite orbital drift that resulted in an underestimation of their temperature trends. It was a minor error in sign in one data correction—the sort that can easily be made by anyone and overlooked by everyone—but that had far-reaching impacts. When the error was corrected their results agreed with climate model predictions to within observational accuracy.

The same issue of Science also contained results from other research which found a systematic error in most radiosonde (weather balloon) datasets that led to their being overcorrected for the effects of direct sunlight (as opposed to shade temperatures). As a result, these datasets also underestimated troposphere temperature trends, and their agreement with the satellite based datasets was thereby shown to be spurious. This effectively demolished skeptic claims that upper atmosphere temperatures disproved global warming, and skeptic forums have since avoided that argument. A few Far-Right commentators still point to the satellite record as “Proof” that global warming is a myth (e.g. Fred Singer of the Science & Environmental Policy Project, and Steven Milloy, commentator for Fox News), but only among the most vociferous and least credible zealots.