IPCC use of non-peer reviewed material?

This is my first contribution for “Ask Stoat“, and I’m doing it because it is low hanging fruit :-). I was going to do the even lower-hanging “airbourne fraction” but that will come. This is for Brian.

So, the issue is in the news because of the 2350 / 2035 kerfuffle, and links to Brian’s other question, “What do you think of WG II?” I’ll answer that one first, because I can think of a cutting answer, which is “I don’t”. Oh, cruel. But true: when I was in the game, I was interested in WG I stuff, which is to say, the physical basis. Someone has to be interested in impacts and adaption, of course: but not me.

As I wrote off in Planet 3.0 recently,

Everyone knows that the WGII and WGIII reports are nowhere near as good as WGI. In fact, taking this further, everyone knows that releasing the WGI, II and III reports at the same time is silly. WGI is supposed to provide the physical science, which should be an input into the other reports. But II and III don’t want to miss the limelight and get released a year later, as they should be.

That is a touch over-harsh, but only a touch. WG I would never have made the mistake WG II made over this 2350 / 2035 stuff, for two reasons. Firstly, they are subject to line-by-line scrutiny because people actually *care*. And second they just do a better job with better people. The only even vaguely comparable issue I can think of is fig 7.1.c in IPCC ’90, and the skeptics rather dislike drawing the obvious moral over that. Before you mistake me, I’m not saying that WG II is rubbish, or valueless: not at all. It’s just not as good as WG I.

[RG in comment 8 has valuable things to say about the variation in quality within WG II; I think he has said enough to convince me that I’m being unfair tarring the whole report with the same brush -W]

So what about the use of non-P-R material? This seems to have been one of those things that everyone knows that turns out to be false. The IPCC *is* allowed to use non-P-R literature. perhaps it shouldn’t be; I don’t much care, as long as the literature is of good quality. But the WWF report should not have been used. The (fairly full) version of the story as I understand it is at [[Criticism of the IPCC AR4#Projected date of melting of Himalayan glaciers.3B use of 2035 in place of 2350]] (that is an oldid so won’t change; I’ve done that in case some wazzock breaks it, you never know with wiki; be sure to check the current version too). However, the reason the WWF report should not have been used is because its not very good, not because it isn’t P-R. The “original” source, which has the 2350 claim in it, is an ICSI report. that would probably count as good enough by the std “P-R rules”, and no-one would have raised an eyebrow. However, what no-one else has pointed out (as far as I can see) is that the 2350 claim in there is trash – it is based on wildly unreliable extrapolation and has no value at all.

Let’s have a look at that, shall we?

The degradation of the extrapolar glaciation of the Earth will be apparent in rising ocean level already by the year 2050, and there will be a drastic rise of the ocean thereafter caused by the deglaciation-derived runoff (see Table 11 ). This period will last from 200 to 300 years. The extrapolar glaciation of the Earth will be decaying at rapid, catastrophic rates– its total area will shrink from 500,000 to 100,000 km² by the year 2350. Glaciers will survive only in the mountains of inner Alaska, on some Arctic archipelagos, within Patagonian ice sheets, in the Karakoram Mountains, in the Himalayas, in some regions of Tibet and on the highest mountain peaks in the temperature latitudes.

But that is from the summary about GW impacts at the end of the ICSI report. The source for 2350 there appears to be earlier at:

Taking air temperature records of the Tien Shari Weather Station and assuming that the same linear trend is to keep up (unfortunately, linear extrapolation is inevitable here), we may find that the mean annual temperature in Central Asia may go up 1.5° C by the year 2350. Proceeding from these rough estimates, in Tables 10, and 11 we have derived figures for glaciation shrinkage DS, changes in the specific glacier melt runoff Rd and the volume of this runoff QRd, as well as for the overall rise of the ocean level Z SL. These data apply to Central Asian glaciers and all of the extrapolar glaciation of the Earth.

Which is why I say it is trash: extrapolating trend form one weather station to 2350 is clearly an utter waste of time and completely invalid. Incidentally, I’ve no idea where table 11 is.

However, whilst use of non-P-R stuff is clearly within the rules, this seems to have surprised lots of people, including IPCC authors. So clearly the culture is to use P-R. Clearly the IPCC is embarassed:

It has, however, recently come to our attention that a paragraph in the 938-page Working Group II contribution to the underlying assessment2 refers to poorly substantiated estimates of rate of recession and date for the disappearance of Himalayan glaciers. In drafting the paragraph in question, the clear and well-established standards of evidence, required by the IPCC procedures, were not applied properly.

Notice that they aren’t, formally speaking, apologising for using non-P-R and I think that is correct: that isn’t the problem: the problem is that they used a low-quality source, and they shouldn’t have.

Tim Lambert has some more on the issue, including a link to John Nielsen-Gammon who says, correctly:

The article calls the statement that glaciers would disappear from the Himalayas by 2035 a “central claim”. As I noted in my analysis of the error, it was not a central claim of the IPCC report; it was not mentioned in the chapter summary, let alone the summary for policymakers. However, it has gained much more emphasis since the IPCC report was published.

Other people may also have said useful things. But I am far far behind on my blog reading.

But I’m rambling now. Brian pushes his luck and asks, “why does the stratosphere cool under global warming?” This is an issue that confuses many many people. If you want to watch a climate person squirm, try asking them to explain this. but not me: as it happens, I answered that years ago (tsk, Brian, where is your due diligence :-). However, I am uneasily aware that Gavin doesn’t believe my (well, actually it is Howard Roscoe’s) explanation, and I’m also aware that though my theory and code does produce a cooling when run through my model, the cooling doesn’t look big enough. You might be better off at RealClimate. Though now I look it seems to have caused *them* (what do I mean, them? I was one of them, then) some trouble too: their first go carries the disclaimer “This post is obsolete and wrong in many respects”. I’m not sure the explanation that they end up linking to is any better than mine. So I’m sticking with mine.

[Update: on the strat stuff, Eli points me to his Stratospheric cooling rears its ugly head…. which sort-of agrees I’m right, but also contains some info on the strongly related question of “was it the ozone or the GHG’s?”. CIP also weighted in but not, I think, usefully -W]