World Climate Report - Global Warming and Whoppers #1 and #2
"Show Me the Evidence":A Tale of Two Whoppers
Very soon, U.S. citizens may get a request from their duly elected representatives: to sacrifice income for the sake of protecting the environment from theoretically catastrophic warming. The few folks with full confidence in their government will happily mail in their checks, no questions asked. But the rest of us will say, "Show me the evidence."
At which point a taxpayer-financed federal scientist will pull out a stack of maps of future global temperatures so covered by red and yellow that they look like a Whopper has oozed ketchup and mustard onto the pages. The "expert" will then proclaim with proper authority that "the world's top scientists have run their data through complex climate models, and the resulting evidence all points to a dramatic and significant global warming that will have serious impacts on the world's climate, health, and resources"
And that, friends, is Whopper No. 2: General circulation models (GCMs) are not evidence.
In scientific parlance, the word that best characterizes these models is hypothesis: a tentative assumption set forth to be proved or disproved. Climate model results represent only one possible outcome that, to be scientifically accepted, must be verified by comparison with other, confirmed information, known as an "independent data set." Simply put, model forecasts must be able to duplicate past or current climate conditions and variations. And we're talking about actual measurements--you can't verify a model by comparing it to the output from another model.
So, say you devise a climate model that factors in various atmospheric conditions. You input this actual data--carbon dioxide levels, solar activity, and so on--set the model in motion, then see what it says the future might be. You check your work by seeing whether your model accurately "predicted" past temperatures and variability. If your model passes this test, it's likely to be accurate for future predictions too.
To date, the models have been consistently too warm. What's more, they've been unable to match the historic global temperature curve, unless they are deliberately tinkered with.
Indeed, current models face many problems, the most serious of which is their inability to properly account for clouds (in what are known as "cloud parameterization schemes"). This shortcoming is highlighted in a recent Journal of Climate article by W. Lee et al. of Scripps Institution of Oceanography.
After their thorough comparison of parameterization methods, the authors "find significant differences in the magnitudes and even the signs of cloud radiation feedbacks, depending on [a modeler's] choice of parameterization schemes." In other words, a change in how clouds are handled will cause global cooling in some models where it causes global warming in others!
"It is somewhat unsettling," they say, "that the results of a modern atmospheric GCM, which may contain several tens of thousands of lines of code, can be affected so much by making what seem to be reasonable changes in only a few of those lines."
It's also "somewhat unsettling" that our political leaders are willing to stake billions of our dollars on the results of climate models that can't even tell us if future cloud-cover conditions will make the planet warmer or colder.
Research grants--huge ones--are available for climate impact studies of every stripe--projects that deal with questions like, "How will global warming influence honey production in Provo, Utah?" And millions are spent using so-called "evidence" that is at best, faulty, and at worst, not really evidence at all.
Our bee-production researcher, for instance, might take GCM data for the grid cell closest to Provo, Utah--probably someplace like Denver or Sacramento, which actually have climates bearing little resemblance to Provo's. Our researcher will fully realize this, of course. And taking climate modelers' advice not to use the forecast for any single grid point, he or she will "adjust" the model output so it looks more reasonable. (Some people call this the "fudge factor.")
In a recent series in the Journal of Climate, J. Palutikof and colleagues effectively dismiss this approach. Comparing two test sites in Spain and Michigan, they make it clear that GCM output should not be used for specific locations or moments in time. The results they came up with are absolutely damning to the utility of current GCMs. Say the authors: "Our results, although limited to only two locations, suggest that it is possible to use empirical techniques to produce downscaled temperature scenarios for a perturbed climate that are more plausible than the GCM values themselves." Which is to say, our bee scientist would be better off just making up his own scenarios than using any of the GCM's results.
We'll ask again: Could someone out there please show us even a shred of actual evidence that a catastrophic climate change is now taking or will ever take place as a result of increasing greenhouse gases? References:
Lee, W.-H., et al., 1997, Cloud radiation forcings and feedbacks: General circulation model tests and observational validation, Journal of Climate, 10, 24792496.
Palutikof, J.P., et al., 1997, The simulation of daily temperature time series from GCM output. Part I: Comparison of model data with observations, Journal of Climate, 10, 24972513.
Winkler, J.A., et al., 1997, The simulation of daily temperature time series from GCM output. Part II: Sensitivity analysis of an empirical transfer function methodology, Journal of Climate, 10, 25142532.
HotTopics Online Preprocessor
Copyright © 1996, 1997 Waveshift Inc.