Monday, October 19, 2009

Means, Extremes, and Hype-ocracy

October 21 Update:
If the term "weather weenie forum" seems a bit harsh, consider the responses when it was pointed out that the original claim was incomplete, misleading, and therefore possibly subject to misuse. First, the comment posted in the forum:
Not strictly correct as stated:
Means, Extremes, and Hype-ocracy [link to this original post]
Some of the responses:
Nobody implied anything about global warming with the stats....Paranoid much, Steve?

He'll get to it in a minute as soon as his dial-up starts working again

Who the Hell are you?

I vote capital punishment for capitol climate. :bike:

He's not a big fan when you question him about his beliefs.

Original post:
Weather geeks love their numbers, and to paraphrase Dr. Freud, "Sometimes a stat is just a stat." With the political polarization of the global warming issue, however, things can sometimes get a little out of hand. As noted here previously, a recent remarkably dank spell has led to several record low maximum temperatures here in the National Capital region, including 2 successive records on Friday and Saturday (October 16 and 17).

Quoting a weather weeenie forum, the earnest young lads at the WaPo's local weather blog have overstated the case: "This was the first time DCA has had back-to-back record low highs in October since 1891." Undoubtedly, this has already made it to the climate deniosphere in some such form as, "WaPo reports global cooling back to 1891 levels." Unfortunately, as is usually the case, the more accurate statement is a lot less sexy: "Of the currently existing record low maximums at Washington, DC for October, the only others on consecutive days were in 1891."

The significant point about extreme records is that they are cumulative. It's a lot easier to set records in the early years because there are fewer previous values to exceed. Therefore, records should become less frequent as time goes by. That means that the 2 consecutive records this year are significant, but not as much as the quote would indicate. A couple of minutes with Excel™, the Swiss Army Knife™ of Data Analysis shows at least one contradictory example nearly 50 years more recently. Consider the record low max of 42° for tomorrow, October 20. It was set in 1940. The current record for today is 45° in 1972. What was the second lowest? It was 46° in 1940, so that was the previous record, and back-to-back records were set in 1940 on the 19th and 20th. There was also at least a third record set in 1940, on the 16th (the one that was broken on Friday).

On the other hand, the October average to date (through the 18th) in 1940 was 59.9°, 1.8° above the average so far this year of 58.1°. Even with declining averages in the next couple of weeks, that's still a long way, though, from the record cold October of 50.7° in 1876 or the 52° tie in 1907, 1917, and 1925.

Besides being tied for the second coldest October, 1925 is also the currently reigning champ for record low maximums; it holds 4 daily records on the 10th, 22nd, 29th, and 31st. Although the daily data are not readily available that far back, those records on the 29th and 31st suggest strongly that 1925 set a third consecutive record on the 30th as well. Interestingly, that month was much more extreme on the high side than the low side, since no currently existing record daily low minimums were set in October that year.

No comments:

Seasonal Outlook

Latest seasonal forecast: Click here.

Latest 3-month temperature outlook from Climate Prediction Center/NWS/NOAA.