Monday, August 16, 2010

Rolling A Very Big And Many-Sided Die

There have been a number of stories implying that the scale and severity of the Russian drought and accompanying wildfires were in and of themselves so bad that they could only be ascribed to global warming. I have sort of winced at this idea, because as nasty as the situation is there, the whole point of the climate is to look for a long-term consistent pattern. And one, or even a few, data points, no matter how startling they may be, are not a very convincing demonstration of a pattern. So it was with some relief I read this article from CSM:
Mann said to think of the occurrence of these extreme events as rolling a loaded die. Rolling a six is like having a record-setting high temperature. With global warming, the die is loaded so that sixes come up increasingly more often — as if the numbers one, two and three were slowly being replaced with sixes.

Rolling a six, or having an extreme weather event, will become more common as the climate changes, Mann said. But rolling back-to-back sixes by chance alone will always be possible, regardless of global warming.

These double sixes, however, will come up far more often than would be expected in the absence of human-caused climate change — a trend that scientists are already seeing.
Now the nice thing about an article like that is that it clarifies the issue, shows that there is disagreement in the community and helps clarify what that disagreement is and why it exists. The problem with the former stories is that they're clear, and they're based on an understandable exaggeration. When that exaggeration is inevitably caught, it's blamed on science and scientists generally, not the specific scientist(s) and journalist(s) responsible. And as I expected, writers with denialist tendencies are all over this and other badly done reports.
The story states: “Worldwide temperature readings show that this January-June was the hottest first half of a year since record keeping began in the mid-19th century.” This implies that we are experiencing something unprecedented. Isn’t it strange that 60% of the U.S. had cooler than normal temperatures during this period. The article statement is untrue and due partially to NOAA computer programs actually manufacturing temperature readings where none exist.
Did you catch what he did there? “Worldwide temperature readings show that this January-June was the hottest first half of a year since record keeping began in the mid-19th century,” is untrue because according to this source, 60% of the US from January-July was cooler than normal.I have no reason to question the veracity of the High Plains Regional Climate Center, and I plan to look over their site when I have more time. But I find it odd that the opinionator would link to the image and not credit and link the site; it's as if in his ranting about NOAA "manufacturing" data, he didn't want anyone to notice HPRCC is part of NOAA's regional climate centers. Oh, and about that "manufacturing:" there are two maps in the National Climate Data Center's (NCDC) monthly report. Here's the first:And here's the second:Now his complaint is that NCDC/NOAA are "manufacturing" the data points in the second not present in the first, and avoids telling the reader that the first is in fact the first one shown, or even exists. "Manufacturing" implies to me "making up," and I did find this sort of startling, though I had some suspicions right away.

Sure enough, in its report, NCDC not only cites the means they're using to get the second chart, but links to the full paper. (~1 MB PDF) Yay, publicly available online peer-reviewed literature! To the extent I've been able to plow through it, it seems to confirm my guess: it adds in satellite data and interpolation. One passage says "The SST [sea surface temperatures] estimates from satellite, ships, and buoys are merged using a weighted sum of the different inputs, with weights inversely proportional to the noise estimate for each type." Now I'll grant that interpolation isn't as desirable as direct measurement, and I'm sure the algorithm isn't perfect. But it certainly isn't what came into my mind when I saw repeated uses of the phrase "manufacturing data," nor was the latter word choice accidental.

In conclusion, the author says
It didn’t take me too long to do some fact checking of this AP story. I wonder why AP or the Arizona Daily Star didn’t bother checking the facts. Could there be some political agenda in running such stories?
Nope, it doesn't take too long at all to bring up your contrarian, denialist bookmarks, and slap in a few that seem to fit, not link the NCDC you criticize, and instead use AP wire reports and press release stenographers as your straw man. Look, we know MSM science reporting is a disaster, and I appreciate that there is some value in criticism and skepticism. But if you want to go there, don't use AP as your proxy- go to the source. Fully link and credit your own sources. Be willing to engage with the ideas at hand- don't just dogmatically try to find contradictory evidence, and when you find one bit, assume that devastates the credibility of an overwhelming pile on the other side. The whole point of skepticism is not to be convinced, but to be willing to be swayed one direction or the other based on the evidence.

We are rolling a very big and many-sided die here, about eight thousand miles in diameter. I'd like to be confident we've got some kind of shot at winning this gamble.

1 comment:

Professor Chaos said...

If the US is cooler, can't we just ignore the data from all the less important countries and just extrapolate from American temps that the world is cooling down? Also, the Braves won last night, so I'm pretty sure that means they are undefeated and will never lose another game.