Precision and accuracy

There is a bit of a brouhaha going on about the top ten warmest years. One scientist noticed something odd in NASA data and a result was that the NASA list of the warmest years was revised. The NASA data is only one of several data sources and the others do not seem affected by this adjustment.

Of course, there are those who assert that the repaired analysis error shows a massive flaw in the man caused global warming advocacy. The advocates say it is no big deal.

There are two big issues here. One is about the argument and the other about the data. The significant point of concern about the argument is that NASA was (and still is) reluctant to disclose its data and its methods in interpreting surface weather records. One of the central NASA scientists involved is also the one who has been very vocal in political allegations. Rather than debate the need for the ability of scientists to inspect raw data and methods used to obtain conclusions, it is about trivial matters of differences in outcomes. This, like labeling skeptics “deniers,” is an indication that the debate is not on very ‘scientific’ grounds of truthfulness and integrity.

The problem with the data is one that confuses precision and accuracy. The debate is about temperatures where the difference in on the order of a tenth of a degree or so. The proper question is whether this precision is matched by the accuracy of the measure.

One effort raising questions about accuracy is an examination of the surface stations use to collect the raw data. That effort has revealed that many of these stations do not meet established standards. Making corrections to historical records to adjust for changes in location or environment or other influences require an ‘interpretation’ of the data. It is this sort of interpretation that was behind the NASA correction. This is why the methods need to be available for inspection so that those who use the data can determine its quality.

Another problem some may have about accuracy is in trying to match up the assertions that the climate is warming by a degree or so over decades with the observations that some think are caused by the warming. A degree or so isn’t going to melt glaciers or entire continental ice masses yet that seems to be the claim. What that means is that global warming by itself must be matched with a chain of causality. Each link in that chain has its own accuracy. It then becomes less of a certainty at the end than at the beginning of the chain and this uncertainty is not reflected in the certitude of pronouncements about the dire effects of the theoretical outcomes.

It is one thing to disagree, to have differing opinions. The certitude of a conclusion must match an appropriate assessment of the precision and accuracy of measure to be a matter of science. The discussion about global warming has a lopsided certitude that does not speak towards it being a matter of science but rather of something else.

Update: Hitting the same points, and also demonstrating an appropriate Limbaugh criticism to contrast with that illustrated in the entry Ignorance on parade is Michael Fumento on James Hansen’s Hacks:

Rush Limbaugh was incorrect in saying the new figures are “just more evidence” that “this whole global warming thing is a scientific hoax.” Conversely, global warming hotheads are also wrong in insisting the revelation deserves no more mention than the back of a Trivial Pursuit card. The GISS, which is directed by global warming guru James Hansen, is saying likewise. He’s wrong. Part of the importance is in the data and part is in how Hansen’s agency behaved, which might be labeled a cover-up.

Criticism on what was actually said that is sourced is shown here. Then there is the notice that the temperature record problem has two main parts centering data problems and behavior problems.

Comments are closed.