Science Gets The Stratosphere Wrong
Time and again the proponents of catastrophic climate change use the mantra of “settled science” to shout down their critics. This is nothing less than blind faith that science actually knows what is going on in the complex environment that regulates this planet's climate. Imagine a part of that system that is literally only 10km from anywhere on Earth, a component of our environment that science thought it understood quite well. Now imagine the embarrassment when a major review in a noted journal finds that previous datasets associated with this component are wrong and have been wrong for more than a quarter of a century. Yet that is precisely what has happened. The area in question is Earth's stratosphere and the impact of this report is devastating for climate scientists and atmospheric modelers everywhere.
Scientists have been launching instrument packages into the upper portions of Earth's atmosphere for a long time. Instruments used for such research were standardized decades ago and programs to collect such data on a world wide basis put into place. If any part of atmospheric science was considered well in hand, if not actually “settled” (a phrase seldom used by real scientists) it would be the long term monitoring of global stratospheric temperatures. However, a report in the 29 November 2012 issue of Nature, “The mystery of recent stratospheric temperature trends,” says that things are not so.
The perspective article by David Thompson, et al., reports that what we thought we knew well we hardly knew at all. A new data set of middle- and upper-stratospheric temperatures indicates that our view of stratospheric climate change during the period 1979–2005 is strikingly wrong. Furthermore, “[t]he new data call into question our understanding of observed stratospheric temperature trends and our ability to test simulations of the stratospheric response to emissions of greenhouse gases and ozone-depleting substances.”
What is particularly troublesome about this report is the scope of the damage done. The problem involves two different sets of historical data from two respected agencies: the UK Met Office and America's National Oceanic and Atmospheric Administration (NOAA). How significant the error and the puzzlement over what to do about it is shown in the article's title, where it is referred to as a mystery. The background of the problem is stated by the authors this way:
The surface temperature record extends for over a century and is derived from multiple data sources. In contrast, the stratospheric temperature record spans only a few decades and is derived from a handful of data sources. Radiosonde (weather balloon) measurements are available in the lower stratosphere but do not extend to the middle and upper stratosphere. Lidar (light detection and ranging) measurements extend to the middle and upper stratosphere but have very limited spatial and temporal sampling. By far the most abundant observations of long-term stratospheric temperatures are derived from satellite measurements of long-wave radiation emitted by Earth’s atmosphere.
The longest-running records of remotely sensed stratospheric temperatures are provided by the Microwave Sounding Unit (MSU), the Advanced Microwave Sounding Unit (AMSU), and the Stratospheric Sounding Unit (SSU). The SSU and MSU instruments were flown onboard a consecutive series of seven NOAA polar-orbiting satellites that partially overlap in time from late 1978 to 2006; the AMSU instruments have been flown onboard NOAA satellites from mid-1998 to the present day.
The widely accepted, continuous record of temperatures in the middle and upper stratosphere going back to 1979 was based exclusively on SSU data. The SSU data were originally processed for climate analysis by scientists at the UK Met Office in the 1980s and further revised as newer satellite data became available in 2008. Here is were things begin to get a bit dodgy.
There are rules that scientists must follow in order for their work to be judged valid. The work must be done openly, transparently—there can be no secret steps or hidden incantations. This is because the work must be reproducible, not just by those who originated it but by outsiders as well. Things began going off the rails when NOAA recently reprocessed the SSU temperatures and published the full processing methodology and the resulting data in the peer-reviewed literature. This is as it should be, NOAA followed the rules. But it soon became obvious that there were grave discrepancies between the new NOAA data and the older Met Office data.
Time series of monthly mean, global-mean stratospheric temperature anomalies.
The global-mean cooling in the middle stratosphere, around 25–45 km in altitude, is nearly twice as large in the NOAA data set as it is in the Met Office data set (see the figure) The differences between the NOAA and Met Office global-mean time series do not occur in a single discrete period of time, but begin around 1985 to increase until the end of the record. According to the Nature article: “The differences between the NOAA and Met Office global-mean time series shown in Fig. 1 are so large they call into question our fundamental understanding of observed temperature trends in the middle and upper stratosphere.”
How did the Met Office get their data so wrong? Well there's the rub. You see, the methodology used to develop the Met Office SSU product was never published in the peer-reviewed literature, and certain aspects of the original processing “remain unknown.” Evidently the boffins at the Met didn't bother to write down exactly how they were massaging the raw data to get the results they reported. Indeed, those who did the data manipulation seem to have mostly retired.
“The methodology used to generate the original Met Office SSU data remains undocumented and so the climate community are unable to explain the large discrepancies between the original Met Office and NOAA SSU products highlighted here,” Thompson et al. summarize. And the damage doesn't stop there.
The data from the erroneous dataset has been used widely to help drive and define computer climate models, the same models used to prop-up alarmist claims of impending catastrophic climate change. According to the report: “Two classes of climate models commonly used in simulations of past climate are coupled chemistry–climate models (CCMs) and coupled atmosphere–ocean global climate models (AOGCMs). By definition, the CCMs explicitly simulate stratospheric chemical processes, whereas the AOGCMs explicitly simulate coupled atmosphere–ocean interactions... A key distinction between the model classes that is pertinent to this discussion is that in general the CCMs resolve the stratosphere more fully than do the AOGCMs.”
One of the predictions made by climate models is that as surface temperatures rise temperatures in the stratosphere should drop. Precisely why this should be so is complex and not important to the point being made here. Suffice it to say, the Met Office version of the SSU data suggests that the models overestimate the observed stratospheric cooling, whereas the NOAA SSU data suggest that the models underestimate it. As the authors put it:
If the new NOAA SSU data are correct, they suggest that the stratospheric mass circulation is accelerating at a rate considerably higher than that predicted by the CCMs, at least in the middle and upper stratosphere (that is, at the altitudes sampled by the SSU instrument). Again, it is possible that the models are correct and that the SSU data are in error. But the fact that the discrepancies between the magnitudes of the simulated and observed cooling in the tropical stratosphere extend to MSU channel 4, which samples the lower stratosphere and exhibits trends that are fairly reproducible from one data set to the next suggest that model uncertainties should not be discounted.
The bottom line here is that models based on this almost universally accepted data are wrong. “If the NOAA SSU data are correct, then both the CCMVal2 and CMIP5 models are presumably missing key changes in stratospheric composition,” the report plainly states. The article goes on to suggest corrective actions to prevent such a travesty being repeated in the future. Alas, the damage has already been done.
What is documented here is simply astounding. That which was thought to be understood is found to be misunderstood. Readings thought to be accurate are shown to be inaccurate. How the data were derived is found to be a secret now lost. The impact of the bogus data ripples through past results and, in particular, climate models, rendering old assumptions invalid. What was that line again about “settled science?”
This is an egregious example of sloppy science, slipshod science, bad science. How other climate scientists blindly accepted the Met Office's manufactured data, even when their models could not be reconciled with nature, leads one to question the scientific integrity of many of those in the field. This is not acceptable behavior in any realm of scientific endeavor, and when the results of research are used to inflame the public and drive questionable socioeconomic programs the malfeasance could be considered criminal. This is what happens when the race for fame, government funding and political advantage collide with science—the validity of the science is destroyed.
Be safe, enjoy the interglacial and stay skeptical.