Tuesday, March 19, 2013

Another WUWT Fail! What is Willis Missing?

Sou | 8:56 PM Go to the first of 4 comments. Add a comment
Willis Eschenbach has written an article about a recent PNAS paper, Homogeneous record of Atlantic hurricane surge threat since 1923 by Grinsted et al, which finds:

that warm years in general were more active in all cyclone size ranges than cold years. The largest cyclones are most affected by warmer conditions and we detect a statistically significant trend in the frequency of large surge events (roughly corresponding to tropical storm size) since 1923. In particular, we estimate that Katrina-magnitude events have been twice as frequent in warm years compared with cold years (P < 0.02).
Click here to download the paper and supporting info.

What is Willis missing?

In a strong surge in overconfidence, Willis makes up two charts of the surge indices listed Table S2 in the supporting info (he didn't get the paper itself).  He says he sees no trend, and asks:
What am I missing?
The Emperor has no clothes

The emperor has no clothes...

I wonder how long it will take before someone tells him what he's missing - like he should have seen already if he'd bother to read the supporting info that he got his data from.  Or even if anyone goes to the paper and supplementary info that the fourth WUWT commenter kindly links to, and reads it. (So far ignored by the next 35 47 commenters.)

Here is what Willis was missing, staring him right in the face in the middle of page 1 of the supporting paper he downloaded, under the heading S2. Events with the Largest Surge Index: (my bold and italics)
...In Table S2 we have also calculated accumulated cyclone energy (ACE) and USACE over the week centered on the date shown. We caution against comparing the relative rank of individual events. The surge index ranking reflects the impact at the specific tide-gauge locations and therefore should not be interpreted as a storm ranking. The purpose of this list is to demonstrate that the surge index truly captures cyclone activity, rather than providing a storm severity ranking.

Still, Willis got what he was looking for...

No worries, Willis got his pseudo-analysis wrong, but he is generating the comments he's looking for - including some completely unrelated to storm surges in the Atlantic: "Arctic ice recovery continues" (huh?), "grant money", "they just make it up", "How can they claim the opposite??", "where's the hockey stick".  

Mike M is particularly hot and bothered, demanding the Governments of China, Finland, Sweden and the UK appear before the US Congress and refund his tax dollars, writing "If they are lying on our tax dollars then haul the responsible gov agency in front of Congress and demand they either back up their claim or forfeit anymore funding for 10 years."

Willis shows just how easy it is to fool fake skeptics.  (Willis has always suffered from Surges in Overconfidence.)

Update: Apparently Pielke Jr has been taking potshots as usual.


William Holder said...

Willis simply plotted the data supplied against hotter and colder years and can discern no trend. I look forward to your analysis of the paper where you explain the adjustments they made to come to the conclusion they reached.
The fact is, that as of now and given the available information, the conclusion offered in the abstract does not seem justified.

Sou said...

Thank you for your comment, William. Helps put your other comments into perspective.

I take it you're saying that, just like Willis and his hordes of fake skeptics, you didn't read either the paper (link above) or the supplement (link above) either. Nor did you read my blog article with the relevant excerpt (as above).

Nor do you care that Willis 'plotted' the wrong data. Emperor Willis can wear no clothes but none of the clueless fake skeptics will tell him so.

A DuKE of deniers (TM Lotharsson)! (Fake skeptics is too good a term for them.)

faustusnotes said...

William, Willis plotted an incomplete data set. He plotted the 50 highest rank storm surges, but there were more than 50 surges - they just aren't shown in the Supplemental Index. If they are plotted on Willis's chart, they would all be very small numbers. The actual data range is from at least 80 years and some years have more than one storm surge (e.g. 1933) so there should be >80 points on Willis's chart. He has only plotted the 50 from the table.

Once you include the full data set used by teh authors, you can then get the p<0.02 that they provide in the abstract.

This rookie error - only plotting some of the data, and biasing your analysis by leaving out a subset with no events - is a teaching moment in first year stats. It was taught to me in the context of the safety experiments for the first shuttle that blew up. That Willis doesn't know about this error tells me all I need to know about his critical skills in this field.

The fact that there is more data than he plotted is obvious from the link he provides, where charts are shown as thumbnails. It's clear there are more than 50 data points in those charts.

Incidentally, sou, have you noticed that these deniers (McIntyre, Watts, Willis, etal) never have access to the original articles? Even though this is basically both their business and their mission (to "save the world" from energy poverty in Willis's case) they won't fork out the $25 to buy the original article they are critiquing. So they really aren't willing to put their money where hteir mouth is. I'm guessing it would cost Watts about $100 a week to get these fundamental tools. Surely his site rakes in that much a week?

This also tells me another thing: despite their claims that there is no scientific consensus on warming, they have no contacts in the scientific world willing to download the article and send it to them. They are very isolated ...

Sou said...

(Correct my previous reply - Willis did plot his data according to temp anomaly, but didn't take any notice of the caution.)

Thanks, faustusnotes.

You are too kind to Watts and Willis. Often the papers are available and easily found for free (as in this case), but they have a tendency to jump in feet first with their opinion of a 'pretty picture' without bothering to refer to the paper. They know their readers don't bother to check and that most wouldn't understand what they read even if they did. (Most WUWT-ers wouldn't disagree with WUWT articles in any case, because science isn't the main purpose of his blog, it's political ideology / 'mob emotions'.)