Global Warming Data Sets Reconciled
Ronald Bailey | May 3, 2006, 10:44am
The National Oceanic and Atmospheric Administration has just released a report that looks at the various global temperature data sets and finds that they are now all "consistent" with man-made global warming. The chief cause is the emission of greenhouse gases such as carbon dioxide released from the burning of fossil fuels.
Global warming skeptics (and I was definitely one of them) have cited the findings of John Christy and Roy Spencer at the University of Alabama at Huntsville who have produced a temperature series based on satellite measurements since 1979. Throughout the 1980s and 1990s, the UAH data series saw little or no warming and its findings were bolstered by separate weather balloon data that also found little warming. In the past few years corrections made to the data sets have boosted average global temperatures in both.
NOAA's new report takes a look at all of the data sets and finds that they all point toward a trend of increasing average warmth:
Global-average temperature increased at a rate of about 0.12 degrees C per decade since 1958, and about 0.16 degrees C per decade since 1979. In the tropics, temperature increased at about 0.11 degrees C per decade since 1958, and about 0.13 degrees C per decade since 1979.
However, the question of how high temperatures are likely to go in the future is still open. Christy told the Washington Post that he has "a minimalist interpretation" of the report because Earth is not heating up rapidly at this point. And questions about what policies should be adopted, e.g., cutting emissions, fostering a technological revolution, adapting, or some combination, are now clearly on the table. The next couple of decades are going to be interesting and very contentious.
thoreau | May 3, 2006, 12:07pm | #
New data sets are great, but it still seems odd to go about identifying and correcting errors in old data.
If the old data is different from the new data, the logical thing to do is examine both data sets and see what sort of errors might be found in each data set. You search for errors not by discarding stuff that you disagree with, but rather by examining the limitations of your instrument, check for dubious assumption in the data analysis, etc.
I'm only through the abstract so far and this is throwing up red flags. Perhaps it was legit, but why no effort was made to modify the models in question when data didn't reconcile is curious.
*sigh*
As I understand it, climate modelers have continued to do calculations. It not like they folded up shop after the first round. The whole point of scientific modeling is not to simply generate a computer program that will spit out a plot that looks just like the experimental data. You've already got the data.
Rather, the point of modeling is to make a minimal set of assumptions about the system in question, assumptions grounded in well-documented experimental facts (e.g. basic equations of heat transport, light transport, etc.). Then you derive equations from those assumptions and solve the equations (e.g. time-evolve a set of partial differential equations). If the outcome of solving those equations matches the experimental data then you conclude that your assumptions capture the basic phenomena and processes governing the behavior of the system.
If there's a mismatch between data and models you have to question both sets of results. You certainly go back to your model and reconsider all of your assumptions and parameters. However, you usually only modify those assumptions and parameters in ways that can be carefully justified from well-supported experimental evidence (e.g. basic laws of physics, basic phenomena of transport, etc.). You don't just introduce any old assumption that would give a result that looks like the experiment.
Why not? Well, who says that the experiments were right? Experimenters can make just as many mistakes as modelers. So they also go back to the drawing board and do more careful measurements, just as the modelers use faster computers and hence finer computational grids, and tweak their assumptions within certain reasonable bounds.
The only time when modelers would fundamentally change their assumptions is if a data set has been reproduced a few times and checked extremely carefully. Why? Because the assumptions that go into models are generally very minimal and well-grounded, or at least they're believed to be so based on the best available information at the time. To drastically alter the assumptions would be to venture into territory with little experimental support. Modeling is a fundamentally conservative endeavor.
As to chewy's 11:48 am post:
The reason why one explanation is favored rather than another usually requires more analysis than is put into an executive summary. If you read the full report and you still aren't satisfied then come back and we'll talk some more, and I'll consider the possibility that the scientists are dishonest and sloppy.
chewy | May 3, 2006, 3:38pm | #
thoreau (and others),
You've made some valid points, though have failed to address others. As reading this report was a task I had already begun, I thought it best to wait to reply. Some of my questions have been addressed in the report, while others simply have not.
It is first important to note two things. First, the authorship of the press release, abstract/front matter/preface, executive summary, and each of six chapters differs, with some overlap. Second, this is the first of 21 Synthesis and Assessment products. The others remain unpublished, while the conclusions of this one have been published and offered to the public through the media.
The main issue is why, as the group states, when the discrepancies between the model and observed data are noted to be the result of problems with the model, or the data, or a combination, the group states that they favor problems with the data alone (the second option). The reason for this favoring is not resolved in this report, though perhaps it will be addressed in subsequent reports.
What the report does, as the questions posed as chapter titles indicates, is to look at issues with the observable data and adjust to account for discrepancies between old and new data. This is done adequately. However, the question of why focus on the data alone, rather than the models and data both, isn't resolved. Perhaps they were simply searching the best lit parts of the alley first, but this should then be stated. Nonetheless, data was adjusted (in a disciplined fashion) and now more findings fit the model. Some parts still do not, and the possibility of problems with the models are still not raised.
The report itself is satisfactory, but doesn't support the overall conclusion drawn (problems were with the data alone, not the models). Perhaps this will be addressed in later reports. But it was dishonest and misleading to release this report alone with the conclusions that were drawn. I'm not claiming that the scientists that did the data collection or correction were dishonest. I'm not claiming that the scientists that developed and ran the models were dishonest. I'm claiming that the scientists making the decision to produce this report alone, and make the claims they have, on a subject as politicized as this (by both sides of the aisle along with all the third party hecklers, green or libertarian) were both dishonest and unethical.
Urinated State of America | May 3, 2006, 6:00pm | #
"IIRC, a couple of moderate examples are the Medieval period (very warm) and the Little Ice Age (1700's??)."
Problem with raising those issues is that this was a localized phenomena. If you look at tropical corals for the same timeframe, you see opposite temperature trends.
"Both periods were very extreme in comparison to what we currently experience. Barely 30 years ago scientists were sounding the alarm of a new Ice Age and/or the population bubble bursting dramatically."
Err, not true. Newsweek is not a scientific journal, folks.
"Reliable temperature records and detailed climatological observations only go back to 1880 or so in the US."
But back to the 1600s in the UK.
"What I rarely hear factored into the global warming debate are the very macro things that can impact climate in a huge way: solar cycles;"
This has been settled. See the coverage of the Scripps/Lawrence Livermore paper last year (http://scrippsnews.ucsd.edu/article_detail.cfm?article_num=666). [I guess those nuclear weapons labs are hangouts for pinko tree-hugging scientists too.]
The next stage of the argument is start cut'n'pasting Idso and arguing that increased CO2 will bring a new golden age of increased plant growth, prosperity & fluffy bunnies.
Next is to accuse those seeking reduction of CO2 emissions of driving us to a Marxist stone age, despite the fact that emissions-trading is the most proposed mechanism for CO2 reduction, and estimates of the cost to mitigate CO2 emissions run aroudn 1-3% of GDP.
Sorry for the sarcasm folks, but some haven't been paying attention.
Paul | May 3, 2006, 8:14pm | #
The point has always been: mass media and politics are no way to make science progress. If I don't need more proof to believe what physicists say about wave-behavior of electrons, it is because it does come from them.
Couldn't agree more. It was very hard to hear the real scientists through the din of the environmentalist whackos, predicting global cooling, global warming, overpopulation, mass-starvation, mass-obesity, acid rain, soil erosion, rain forests gone by the early nineties, increasing rainfall, decreasing rainfall, greater temperature highs, lower temperature lows... need I go on?
Global Warming became, and still is for many people, a religion. There was no evidence that didn't, in some way, support the theory.
Science, when in the hands of any activist, is almost universally a bad thing.
In the nineties, when I cared about this subject and debated it extensively, I read scientific paper after scientific paper regarding Global Warming(tm), and the evidence presented, once read in its painfully boring technical speak was... alas, underwhelming. The real scientific reports could be more accurately summarized as thus:
o Yes the planet is getting warmer, but climate isn't static- planet is getting warmer or getting cooler at any moment in time.
o Yes, the planet may be getting warmer due to increased C02 and climate forcing.
o No, we don't know how much it'll warm, but we *think* it will warm by this much, unless it doesn't.
o Sea ice is decreasing over here, but increasing over there.
o These temperatures are what we predict based upon models, but they don't all ways coincide with real observations.
o Some storm ferocity may increase, but frequency may decrease.
o Historic records of high atmospheric C02 concentrations don't always PREcede a temperature increase, but sometimes appear to PROceed a temperature increase.
o We're coming out of an ice age, so the extent to which C02 is warming the planet may be exaggerated.
o Disease vectors may change, or may not-- malaria, afterall was very common near the arctic circle before the widespread use of DDT- so temperature may not play a direct role.
o C02 is a
minor greenhouse gas, and as of yet, we don't fully understand the feedback mechanisms of the major greenhouse gas:
wator vapor
Any clear thinking individual, after reading these things in the actual papers would hardly conclude that we're CAREENING towards disaster. And as such, I'm not convinced yet, either.
uncle sam | May 4, 2006, 9:51am | #
Sorry, I haven't moved the goal posts. I just want them to show:
1. that they understand the mechanisms of climate change well to describe how they KNOW that
anthropogenic CO2 has contributed significantly to historic warming trends.
2. how they have diferentiated between anthropogenic contributions and natural causes.
So far they have given evidence that there has been warming. The have even managed to improv the computer models to more closely reflect the observed increase in warming.
They have not shown (proven their case) regarding the points I set out at 1:55 am.
As near as I can tell, the most they have been able to give us is: likelys, probablys, possiblys, maybes, etc.
They have not shown, for instance, that ocean warming is not responsible for the increase in atmoshperic CO2, or how they came up with the proportions they attribute to anthropogenic and natural contributions to GW.
Yes, I understand geometric proofs, but proof also can be shown in the form of evidence to support theories. Given that warming has been observed for some time, and inferred from historical evidence, they haven't really needed to "prove" that, as it is not really a theory so much as an observation. I am willing to accept observations as a form of proof, but I am still waiting for climate scientists to demonstrate:
1. that their computer models can accurately (reasonably) predict the future trends any.
2. that said trends will vary from straightline projections of current/historic trends.
3. that observed trends indicate an impending environmental disaster.
4. that popular proposals will significantly alter climate trends. This is one area that apears to be based soley on assertion.
