13 September 2012

The IPCC Sinks to a New Low

Back in May, Chris Field, co-chair of IPCC Working Group II emailed me with a request:
As per your request, the IPCC is prepared to take another look at the AR4 text on disaster trends. The spirit of these "second looks" is to evaluate whether the assessment should have said something else, based on the literature cited and on the information that was available within the window for AR4 literature.  The error protocol does not allow a new assessment based on literature published since the AR4 literature cutoff, and it is not intended as a broad reinterpretation of the information assessed by the authors.

To clarify your request, can you send a specific statement of the alleged error or errors that you would like to see addressed?
As my request to Field was informal, and not one I ever expected to see action on, this initiation of contact with me seemed to me like the IPCC was turning a corner, and taking seriously scientific accuracy on disasters and climate change. So I prepared a concise and specific reply to Field's request. Today I heard back from the IPCC. The response is laughable, and indicates that the IPCC is more interested in playing games than in scientific accuracy. Nothing below is complicated or nuanced.

Here are the details from the response that the IPCC sent to me today, annotated with my comments. In the material below the four passages under "Text from Roger Pielke, Jr." is that which I provided to Chris Field in May in response to his email request. Under each of those I have blockqouted the IPCC response to my claims, which is titled "CLA Finding." Below that I highlight my comments today in response to each of the four responses.

With that, let's have a look . . .

Alleged errors in the treatment of disaster trends in Chapter 1, WGII, AR4
CLA response from Cynthia Rosenzweig and Gino Casassa
August 23, 2012


Alleged Error #1

Text from Roger Pielke, Jr.

Error #1: IPCC p. 110: “These previous national U.S. assessments, as well as those for normalised Cuban hurricane losses (Pielke et al., 2003), did not show any significant upward trend in losses over time, but this was before the remarkable hurricane losses of 2004 and 2005.”

FACTUALLY INCORRECT: Figure 5 in the following paper, in press prior to the IPCC AR4 WGII publication deadline, clearly shows that the addition of 2004 and 2005 losses do not alter the long-term trend in hurricane losses:

Pielke, Jr., R. A. (2006), Disasters, Death, and Destruction: Making Sense of Recent
Calamities. Oceanography 19 138-147.
http://sciencepolicy.colorado.edu/admin/publication_files/resource-2449-2006.02.pdf

This same information was also in the report of the 2006 Hohenkammer Workshop on Climate Change and Disaster Losses, which was cited by the AR4 WGII: http://cstpr.colorado.edu/sparc/research/projects/extreme_events/munich_workshop/pielke.pdf

RECOMMENDED CORRECTION: ““These previous national U.S. assessments, as well as those for normalised Cuban hurricane losses (Pielke et al., 2003), did not show any significant upward trend in losses over time, and this remains the case following the remarkable hurricane losses of 2004 and 2005.”
CLA Finding

There is no error in the statement. No correction is needed and the text can stand as is.

Rationale
The clause about the published analyses being before the 2004 and 2005 hurricane seasons is a statement of fact about the time line, and it is not a statement that the results were different after including 2004 and 2005. The statement does not infer that the overall pattern of losses would be different; instead it suggests that 2004 and 2005 were remarkable years in terms of hurricane losses, which they were.
PIELKE RESPONSE SEPTEMBER 13:  This boggles the mind. The time line was such that published analyses (I provided 2!) that were available to the IPCC when drafting the AR4 included 2004 and 2005. The IPCC is say that up is down, and with a straight face. Did they not even read what I wrote?

UPDATE: Here is how the IPCC handled this issue in the 2007 AR4 review process.

Alleged Error #2

Text from Roger Pielke, Jr.

Error #2: IPCC pp. 110-111: “Global losses reveal rapidly rising costs due to extreme weather-related events since the 1970s. One study has found that while the dominant signal remains that of the significant increases in the values of exposure at risk, once losses are normalised for exposure, there still remains an underlying rising trend.”

That “one study” is Muir-Wood et al. (2006), a white paper prepared as input to a workshop that I organized.

FACTUALLY INCORRECT: (a) The first sentence should say “1950s” not “1970s,” which is the starting point of the Munich Re dataset being referred to. (b) Several normalization studies (not “one study”) available at the time of the AR4 had noted that a dataset that begins in 1970 and ends in 2005 will show an annual rate of increase, including papers parallel to the Muir-Wood et al. (2006) presented at the Hohenkammer workshop as well as the IPCC TAR (2001) and Munich Re (2000). All such studies find no evidence of “an underlying rising trend” over longer time periods.

RECOMMENDED CORRECTION: ““Global losses reveal rapidly rising costs due to extreme weather-related events since at least the 1950s. Multiple analyses have found an increase in normalized losses since 1970, due entirely to US hurricanes, but such studies are in agreement that no such trend can be found over longer time periods (dating to 1950 globally and 1900 for US hurricanes).” In addition, the full text of this section should be clarified along these lines.
CLA Finding

There is no error in the statement. No correction is needed and the text can stand as is.

Rationale
The year 1970 was used in the analysis of Muir-Wood 2006. They used a data set with information going back to 1950, but they decided to omit the period prior to 1970 because they are limited in completeness. The “correction” proposed by Pielke acknowledges that analyses for the period starting in 1970 observe an increasing trend, which is all that is reported in the text. The qualifier about “one study” documenting a trend indicates the limited foundation for the conclusion, in the context of other studies showing the large importance of trends in exposure.
PIELKE RESPONSE SEPTEMBER 13: Again, the IPCC is saying that up is down. Multiple studies showed disaster losses increasing from 1970 (I point them to several, including IPCC TAR). The reason for this was simple -- low hurricane damages in the 1970s and 1980s. Not new or interesting. To highlight one study and then now suggest that this was done to suggest that the conclusion is tenuous makes the IPCC look utterly clueless. The IPCC cherry picked a date to make a suggestive claim, and ignored contrary data. Not good. The cover-up is worse than the crime.

Alleged Error #3

Text from Roger Pielke, Jr.

Error #3: IPCC p. SM.1-4: Figure SM-1.1 and caption: “An example from the literature of one study analysing rising costs of normalised weather-related catastrophes compared with global temperatures. Data smoothed over ±4 years = 9 years until 2001 (Muir Wood et al., 2006).”

FACTUALLY INCORRECT: Neither the figure nor the underlying data appear in the scientific literature (peer reviewed or grey) at any time. The figure was created by Robert Muir-Wood and included in the report with an intentional mis-citation to circumvent the IPCC publication deadline, according to Muir-Wood himself (an audio recording of his admission is available on the website of the Royal Institution in London, and linked from my blog here:
http://rogerpielkejr.blogspot.com/2010/02/ipcc-mystery-graph-solved.html ).
He apparently believed that the figure would appear in a future paper – it did not, and that future paper, eventually published in 2008, found no relationship between temperatures and disaster losses.

RECOMMENDED CORRECTION: The figure and caption should be removed as well
as reference to it in the text.
CLA Finding

Post erratum to the caption of figure SM-1.1 clarifying the source of the data in the figure.

Rationale

The figure is a replotting of the data used for the analysis of Muir Wood et al. (2006). These data combined temperature data from CRU (Climatic Research Unit 2006),(University of East Anglia, Norwich) with the disaster loss database compiled by and described in Muir Wood et al. (2006). Since Muir Wood et al. (2006) did not cite CRU, the figure caption would be more accurate if modified to read:
Figure SM-1.1 Costs over time of normalized weather-related catastrophes compared with global temperatures. Data smoothed over ±4 years = 9 years until 2001. Based on the dataset used in Muir Wood et al.(2006) and temperature data from CRU, 2006*.
*ADD CRU TO THE REFERENCES FOR THE SUPPLEMENTARY MATERIAL.
(Climatic Research Unit (CRU) (2006).University of East Anglia, Norwich.
www.cru.uea.ac.uk/cru/data/temperature/#sciref .)
Pielke Response September 13: This one takes the cake. Utterly remarkable. The graph was conjured up by Robert Muir-Wood based on data in his possession, he intentionally miscited the analysis and there is no scientific basis for plotting damages against temperature. None. The IPCC response on this issue is that their falsified graph was not fully cited, but otherwise OK is an insult to scientists everywhere and a mockery of the IPCC process. I cannot express this strongly enough. The IPCC has demonstrated that it is utterly incapable of correcting even the most egregious violation of its standards.

Alleged Error #4

Text from Roger Pielke, Jr.

Error #4: Erroneous IPCC Press release of 25 January 2010 (a) “one study detected an increase in economic losses, corrected for values at risk, but that other studies have not detected such a trend,” and (b) “In writing, reviewing, and editing this section, IPCC procedures were carefully followed to produce the policy-relevant assessment that is the IPCC mandate.”

FACTUALLY INCORRECT: (a) as documented above in Error #2, multiple studies had identified an increase in economic losses since 1970 (but not from earlier starting points), and (b) as documented above in Error #3, IPCC procedures were not carefully followed, but violated.

RECOMMENDED CORRECTION: The IPCC should withdraw its 25 January 2010 press release and issue a press release noting the inaccuracies in both the report and the release.
CLA Finding and Rationale

The January 25, 2010 IPCC statement is not part of an IPCC report, and the error correction protocol is therefore not relevant
PIELKE RESPONSE SEPTEMBER 13: Not that the IPCC has shown a commitment to accuracy, but here it relies on a bureaucratic dodge to ignore the false information it put out via press reslease. Not good.