Monday, February 21, 2011

Metadata

The recently published paper by Legras, Mestre, Bard, and Yiou, which started out as "On misleading solar-climate relationship" and was published as "A critical look at solar-climate relationships from long temperature series" is being discussed over at Open Mind and Real Climate, and Eli guesses a lot of other places. Go to the links for the science, the Rabett wants to discuss the metadata, aka the Interactive Discussion at Climate of the Past.

In the hurly burly talking about the open model for publishing, Eli remarked that these interactive discussions are great for bringing home what the manuscripts, and the eventually published papers are and are not about, where the strong and where the weak points are. This metadata enables non-experts to follow the bouncing ball, and indeed the Legras, et al. discussion is a great example. To start at the top, all of the referees HATED the original title, and the authors quickly saw their point, slowly. Reviewer #1 put it this way:

The title of the study (On misleading solar-climate relationship) does not quite relate to the arguments the authors make. The authors argue that a solar-climate relationship does not really exist - at least that it cannot be convincingly shown that it differs from randomness. The point I like to make is that a non-existing relationship cannot be misleading. My suggestion would be to change the title to something like A critical look at solar-climate relationship (although I realize that this might remind readers to the work of Pittock (1978)).
A couple of the referees wondered why this was submitted to Climate of the Past and not the journal that the article was criticizing, to which Legras, et al. replied
We have chosen to submit our manuscript to Climate of the Past and not in the original journal where LMKC and KLMC have appeared for two main reasons: (1) by redoing all the calculations and using new homogenized data not included in LMKC, we go beyond a simple comment and (2) we considered that the matter requires an open review process which is a unique feature of Climate of the Past.
Which certainly warmed the hearts of the Climate of the Past editors, but also show why this model has a bright future. LMKC and KLMC are
Kossobokov, V., Le Mouël, J.-L., and Courtillot, V.: A statistically significant signature of multi10decadal solar activity changes in atmospheric temperatures at three European stations, J. Atmos. Solar-Terr. Phys., 72, 595–606, doi:10.1016/j.jastp.210 2010.02.016, 2010a
and
Le Mouël, J.-L., Kossobokov, V., and Courtillot, V.: A solar pattern in the longest temperature series from three stations in Europe, J. Atmos. Solar-Terr. Phys., 72, 62–76, doi:10.1016/j. jastp.2009.10.009, 2010.
The scientific issues boil down to
a) is sun spot number a good proxy for solar insolation, better put is it the best currently available model for solar insolation going back a few centuries.

b) can you look at raw data from a few stations without knowing or doing anything about possible problems with the data.
c) given multiple independent variables, especially if there is some fortuitous correlation between them, can you ignore all but the one you are interested in when you look for a correlation with an observed quantity.
LeMouel and friends say sure, not a problem. Legras, et al. demur but their explanations in the replies to the referee reports are informative. On point a)
In our paper, we criticized the oversimplified use by Le Mouël et al. (2010) 5 and Kossobokov et al. (2010a), hereafter LMKC and KLMCa, of the sunspot number (SSN) as a proxy for solar forcing. In response, Kossobokov et al. (2010b), hereafter, KLMCb state that "the international sunspot numbers are the only and therefore, the best available proxy of solar activity in the last 250 years". This statement is wrong because other solar proxies cover this time period such as records of aurorae (e.g. Silverman 1992), cosmogenic isotopes (e.g. Delaygue & Bard 2010) and the aa geomagnetic index (over 240 yr, e.g. Lockwood & Stamper 1999).

In addition, it has been shown by many authors that past observations on sunspots can be used to derive better solar forcing records than the raw SSN used by LMKC & KLMCa. For example, previous authors used the envelope of the SSN 11-year cycle, the length and decay rate of the solar cycle, the structure and decay rate of individual sunspots, the mean level of SSN, the solar rotation and the solar diameter (see review by Bard et al. 2000 and references therein).

Lean et al. (1995) proposed that the solar forcing record could be divided into two superimposed components: an 11-year cycle based on the parameterization of sunspot darkening and facular brightening and a slowly-varying background. Solanki and Fligge (1998) reconstructed a solar irradiance record back to 1874 AD by using different relation active regions (quadratic calibration between the SSN and the spacecraft TSI record over a decade) and to the long-term component (linear calibrations between brightness and chromospheric emission or length of the activity cycle). LMKC & KLMCa could also refer to Solanki (2002) who published solar forcing curves over the past 150 years.
and
Kossobokov et al. refer to "ongoing controversies, such as that between Scafetta and Willson (2009) and Krivova et al. (2009)". This statement is misleading because there is no ongoing debate between these two particular studies: Scafetta and Willson (2009) used the model developed by Solanki et al. (2005) and Krivova et al. (2007). These authors (Krivova et al., 2009) simply discovered that Scafetta & Willson had made fatal mistakes in using Krivova et al.s model, but that its correct use leads to a stable irradiance baseline. Hence, the study by Krivova et al. (2009) constitutes a clear-cut and definitive refutation of the previous claim by Scafetta and Willson (2009), a paper that should be no longer cited as a valid reference (as done by LMKC and KLMC). Further demonstrations of the errors made by Scafetta & Willson are developed in Gray et al. (2010).
So not having paid much attention, Eli now goes and reads Krivova et al. 2009. Tamino, pretty much shows the problems with b), that in at least one of the three series that LMKC use, Bologna, there is a huge artifact that, as luck would have it falls right where it "supports" LMKC's argument. Referee #4 points out that such problems should average out over time. Legas et al respond
Homogeneity errors do not need to project always on solar forcing but we have a nice example with the spurious bump in the Bologna TX series between 1865 and 1880. This period coincides with an isolated high solar cycle according to LMKC and contributes significantly to the solar shift as shown in our supplementary file. An other example is the jump near 1950 in the Maastricht temperature series. Such jump correlates with both the the initiation of anthropogenic forcing and a sequence of active solar cycles
c) you can find in the paper as well as the replies
These minima occurred during the intervals 1795–1830 (Dalton Minimum) and 1880–1920 (Modern Minimum) as evidenced with various solar indicators: sunspots (Hoyt and Schatten, 1998), aurorae (Silverman, 1992), aa geomagnetic index (Lockwood and Stamper, 1999; Lockwood et al., 1999), cosmogenic nuclides (Delaygue and Bard, 2010). However, these two time periods also include some of the largest volcanic eruptions ever recorded in history. The first period comprises the cold decade linked to the Tambora (1815) and the 1809 stratospheric eruption (Cole-Dai et al., 2009), whereas the second phase includes a series of major eruptions starting with the Krakatoa in 1883 and ending with Mt Katmai in 1912
(Robock, 2000).

Climate modelling allows to quantify the collective impact of these forcings in order to explain the temperature historical record of the past few centuries (e.g., see the model data compilation in IPCC AR4 Sect. 6.6.3.4 with Fig. 6.13 and 6.14 in Solomon et al. (2007), (http://www.ipcc.ch/ publications and data/ar4/wg1/en/figure-6-14.html), or the more recent paper by Gao et al. (2008) and the study by Wagner and Zorita, 2005). The Northern Hemisphere temperature drops corresponding to the Dalton (0.2–0.3 C) and Modern solar minima (0.1–0.2 C) are partly linked to an enhanced volcanic forcing (see Hegerl et al., 2007b, and references herein). This implies that the attempt by LMKC and KLMC at studying the Sun-climate relationship cannot be performed with a simple approach that omits the influence of volcanic eruptions.
Another advantage of the interaction review is how a brief comment in the paper which only the consignetti would understand
The ECA&D dataset and metadata are freely available through ECA&D web interface [http://eca.knmi.nl]. The temperatures used by LMKC are three daily series of maximum (TX) and minimum (TN) temperatures collected in Praha since 1775, Bologna since 1814 and Uccle since 1833. Owing to policy changes, the Uccle data were not available during the writing of this paper.
is explained in the replies
Referee #1page 774, line 12: consulting the ECA&D staff made clear that the policy changes were at the side of the Belgian Met. Office (KMI) rather than ECA&D. This point also relates the remark made at page 778, line 8.
It is true that the Uccle daily temperature series has been removed from ECA&D dataset upon request of the Belgian Met. Office who considers that ECA&D free distribution scheme is interfering with its commercial policy. Although the data are plausibly still available for research purposes, we have decided not to require them since it would not have been possible to redistribute them with our supplementary material. We have added a few words to make clear that the policy change is not due to the ECA&D team who, on the contrary, strongly supports open access to data. In any case, we do not think that our demonstration suffers from the lack of analysis of Uccle data.

7 comments:

Pinko Punko said...

I did not realize that the Main Actors on our current stage have a history of terrible.

This previous nugget is not short of words that indicate expectations for LMKC should be set on the low side.

Anonymous said...

@Pinko Punto :

About their previous works, they were smacked down by two sessions of the French National Science Academy (even though some manoeuvers were tried behind to mitigate the final report). After a last attempt not to loose the face, they went silent.
Courtillot went back to the CCS lab and is courting institutes dealing with that ; Allegre didn't get his wished position in the government and is now at a loose end in his ancient lab. He tries to engage discussion with people here.
The debunked article was their last straw of credibility in the climatology field. No wonder they replied the way they did.

Bratisla

David B. Benson said...

Bratisla --- Aha! Thanks for that comment.

Steve Bloom said...

OT: In case you don't see it, Eli, I left the following over at CP re the recent snippiness:

'I think Eli's "as usual" (in #6 above) referred to a common problem with these sorts of analyses, not with this blog, although Rabbettese being what it is it was easy to conclude otherwise.

'On the substance, I agree with Eli that a post that seeks to discuss Egypt's food supply problems in the context of "the midterm and long-term" and advocates for "a strong foundation for lasting and durable economic growth" really is missing the boat if it *entirely* neglects (as the post did, not even mentioning it as a subject for a follow-up) the complex of factors related to sea level rise, groundwater depletion and the dam. As things stand, in the long term these factors will more than cancel out any progress that can be made in the areas described in the post. Noting that neither of the other two linked CAP articles touched on them, perhaps a little internal consultation and education is in order.

'I look forward to seeing the follow-up post.'

Kloor is, of course, as useless as ever.

Robert Way said...

Eli,
note the Cryosphere is another great open journal. It is for matters relating to snow and ice.

Robert Way said...

Another thing eli should consider is that the authors responded with a total of 35 pages to the reviewers. Isn't it interesting how Odonnell et al complained about having responded with 80 pages? Methinks that Odonnell et al don't realize how rigorous peer review actually is!

Georg Hoffmann said...

I found in particular the near perfect recycling rate of KLMC remarkable.
Have a look
http://www.scienceblogs.de/primaklima/2010/05/klimaskeptiker-sind-fur-umweltschutzendes-recycling.php

4 out of 8 figures are identical.
Georg