"Our greatest responsibility is to be good ancestors."

-Jonas Salk

Tuesday, August 18, 2009

Klotzbach in the Blogosphere

I completely concur with James' summary of the state of play. We haven't yet gotten down to whether the Klotzbach mechanism is or isn't realistic or supported by the evidence. But we do have a misleading headline from R Pielke  Jr,
Well, one thing it does not mean is the main title of RPJnr's post, that the surface temperature trend is overstated. There is nothing in the paper (despite three gratuitous plugs of Watt's photography site) that actually argues for the measured temperature trend being wrong in any way.
Yet the conclusion of a temperature bias is the one prominently displayed as the headline of a blog posting, and echoed on several websites including Watts, which RP Jr is as yet not willing to amend or retract.

I am less than favorably impressed by this trend of writing papers of whatever quality, and then prominently but informally misrepresenting their results in ways to give substantively wrong comfort to one (or the other) side in a political debate. For a man who talks of separating science and politics, I am forced to conclude that RP Jr. sets a very peculiar example.

Note also that RP Sr is piling on at Watts, Pielke Sr. on warm bias in the surface temperature trend "Provides evidence of the significant error in the global surface temperature trend analyses of NCDC" and using Klotzbach as evidence toward the point. Now if that isn't telling the Watts crowd watt they want to hear, especially posted right after Roger Jr's thing, I guess I don't know watt is. Perhaps some judicious hairsplitting can save the latter headline, (like, for instance, "oh, I just meant the analysis is erroneous, not that the trend is actually wrong" and "significant means memorable or important, everybody knows that") but that is not straight talk by any means.

Whether Koltzbach and Pielke Sr. have actually resolved the temperature trend differential or not is something I am keeping an open mind about for now. But if they have, it means that both temperature records are substantially correct. I can't imagine why any scientist would be dissatisfied with a result like that, or what scientific motivation there could be to paint it as something else, and then trumpet the unsupported assertion as a new result on a contentious website like Watts'.


Paul said...

Gavin Schmidt has weighed with some important sounding stuff on the surface to altitude amplification ratios alluded to by James in his blog post.


Paul Middents

Michael Tobis said...

Informal translation of Gavin's comment. Somebody please correct me if this is wrong.

"One wouldn't expect tropospheric amplification over land and it doesn't appear in the models. Therefore the problem you propose to solve does not exist.

"The reason to not expect amplification over land is simple and requires no appeal to subtle boundary layer processes. Most tropical land is not far from the ocean, and the ocean dominates the patterns in the tropical belt. Accordingly, the tropospheric profile in the tropics will be close to the oceanic one. On the other hand the land surface heats much faster than the ocean. (This is well known. It's basically because the ocean transports heat downward in currents and eddies so you have to warm a larger mass to make a difference.) So the usual amplification over the ocean doesn't appear over land in the models or in the data, there really isn't much to explain, and there certainly is no need to appeal to the effectiveness of boundary layer mixing."

Informal translation of K et al reply, as usual, is beyond me, but it seems to say "thank you for confirming our results"!

Deep Climate said...

Something funny going on - the link to the post doesn't work, but it's still up at the front page.

I have a feeling they jumped the gun.

Anyway. If you look at the recalculated amplification metrics, Hadley shows better agreement with satellites over land than over ocean. Doesn't that completely negate the essential point about land "bias"?

And I still think the significance thresholds don't look right (as I commented at Janes Annan's).

The response from K et al is hilarious, unintentionally of course.

I'd like to see a comment, or even better, a paper from Schmidt and others (wonder if Santer has time), with a proper analysis of land vs ocean amplification in models and observations. It's not as if a proper analysis hasn't been done before (e.g. for the tropics).

Poor old JGR - that's two problematic papers in a month.

Anonymous said...

Never mind - the Paul's link to RP jr seems OK again.

Martin Vermeer said...

Hmm... so there is no boundary layer effect due to nighttime out-radiation to speak of. It's all the thermal inertia of the ocean...

I's like to be a fly on the wall when the time comes to agree on the title of the new paper.

Michael Tobis said...

It appears I've gone a little too far above. James summarizes here.

Paul said...


Where do think you went a little to far?

Paul Middents

Anonymous said...


Where do think you went a little to far?

Not to speak for MT, but if James/Gavin are correct, it's looking less and less likely that the Klotzbach paper actually does reconcile the suspected discrepancy between surface and topospheric trends, which MT had stipulated for the purpose of argument and thought was significant enough of a finding as to make all of the sidelong winking at the Watts crowd bizarre.

Michael Tobis said...

In the comments, I paraphrsaed Gavin as saying "Therefore the problem you propose to solve does not exist."

James says "This correction immediately knocks off half of the missing amplification effect that Klotzbach et al was explaining."

In fact I don't entirely understand what James is saying, but James is one of those people who, when when they disagree with me on a matter of substance in their expertise, I am immediately inclined to abandon my position.

Maybe I retreated too far now that I think about it.

I think what James is saying is that there remains a small difference between what the models say should be the difference between the surface and the tropospheric trends vs what the observations say
about it.

Is that what we are supposed to be losing sleep over? A small difference between the small differences of small trends between the predictions of an imperfect model and an imperfect measuring system? Not my idea of a very robust statistic, even if you can paint the latitude-height map of it as a bright pink bullseye.

Deep Climate said...

I don't think you and James are that far apart.

Klotzbach et al claim that various surface/satellite discrepancies provide evidence of a "warm bias" in the surface land temperature series.

That claim simply doesn't hold up, due to various problems, including use of the wrong amplification factor and incomplete statistical analysis.

I've expanded on this at James's blog, so I won't repeat it all now, although I'm interested in your reaction (and his).

But, to me, one telling point is that incorporating GISS amplification factors results in better agreement between HadCRU and the sat observations over land than over ocean (or globally). So it can no longer be argued that there is "warm bias", even a reduced one, in the HadCRU land temperature series.

Or am I missing something?

For more see:

Deep Climate said...

Faster link to the comments at James's for anyone interested.


Paul said...

Gavin again addresses the so-called "bias" in surface trend: