From: Ben Santer To: "Thomas R. Karl" Subject: Re: [Fwd: Re: [Fwd: concerns about the Southeast chapter]] Date: Thu, 30 Jul 2009 18:41:44 -0700 Reply-to: santer1@llnl.gov Cc: Virginia Burkett , Thomas C Peterson , Michael Wehner , Karl Taylor , peter gleckler , "Thorne, Peter" , Leopold Haimberger , Tom Wigley , John Lanzante , Susan Solomon , "'Philip D. Jones'" , carl mears , Gavin Schmidt , Steven Sherwood , Frank Wentz Dear Tom, Thanks for forwarding the message from John Christy. Excuse me for being so blunt, but John's message is just a load of utter garbage. I got a laugh out of John's claim that Santer et al. (2008) was "poorly done". This was kind of ironic coming from a co-author of the Douglass et al. (2007) paper, which used a fundamentally flawed statistical test to compare modeled and observed tropospheric temperature trends. To my knowledge, John has NEVER acknowledged that Douglass et al. used a flawed statistical test to reach incorrect conclusions - despite unequivocal evidence from the "synthetic data" experiments in Santer et al. (2008) that the Douglass et al. "robust consistency" test was simply wrong. Unbelievably, Christy continues to assert that the results of Douglass et al. (2007) "still stand". I can only shake my head in amazement at such intellectual dishonesty. I guess the best form of defense is a "robust" attack. So how does John support his contention that Santer et al. (2008) was "poorly done"? He begins by stating that: "Santer et al. 2008 used ERSST data which I understand has now been changed in a way that discredits the conclusion there". Maybe you or Tom Peterson or Dick Reynolds can enlighten me on this one. How exactly have NOAA ERSST surface data changed? Recall that Santer et al. (2008) actually used two different versions of the ERSST data (version 2 and version 3). We also used HadISST sea-surface temperature data, and combined SSTs and land 2m temperature data from HadCRUT3v. In other words, we used four different observational estimates of surface temperature changes. Our bottom-line conclusion (no significant discrepancy between modeled and observed lower-tropospheric lapse-rate trends) was not sensitive to our choice of observed surface temperature dataset. John next assets that: "Haimberger's v1.2-1.4 (of the radiosonde data) are clearly spurious due to the error in ECMWF as published many places". I'll let Leo Haimberger respond to that one. And if v1.2 of Leo's data is "clearly spurious", why did John Christy agree to be a co-author on the Douglass et al. paper which uses upper-air data from v1.2? Santer et al. (2008) comprehensively examined structural uncertainties in the observed upper-air datasets. They looked at two different satellite and seven different radiosonde-based estimates of tropospheric temperature change. As in the case of the surface temperature data, getting the statistical test right was much more important (in terms of the bottom-line conclusions) than the choice of observational upper-air dataset. Christy's next criticism of our IJoC paper is even more absurd. He states that: "Santer et al. 2008 asked a very different question...than we did. Our question was "Does the IPCC BEST ESTIMATE agree with the Best Data (including RSS)?" Answer - No. Santer et al. asked, "Does ANY IPCC model agree with ANY data set?" ... I think you can see the difference. Actually, we asked and answered BOTH of these questions. "Tests with individual model realizations" are described in Section 4.1 of Santer et al. (2008), while Section 4.2 covers "Tests with multi-model ensemble-mean trend". As should be obvious - even to John Christy - we did NOT just compare observations with results from individual models. For both types of test ("individual model" and "multi-model average"), we found that, if one applied appropriate statistical tests (which Douglass et al. failed to do), there was no longer a serious discrepancy between modeled and observed trends in tropical lapse rates or in tropical tropospheric temperatures. Again, I find myself shaking my head in amazement. How can John make such patently false claims about our paper? The kindest interpretation is that he is a complete idiot, and has not even bothered to read Santer et al. (2008) before making erroneous criticisms of it. The less kind interpretation is that he is deliberately lying. A good scientist is willing to acknowledge the errors he or she commits (such as applying an inappropriate statistical test). John Christy is not a good scientist. I'm not a religious man, but I'm sure willing to thank some higher authority that Dr. John Christy is not the "gatekeeper" of what constitutes sound science. I hope you don't mind, Tom, but I'm copying this email to some of the other co-authors of the Santer et al. (2008) IJoC paper. They deserve to know about the kind of disinformation Christy is spreading. With best regards, Ben Thomas R. Karl wrote: > FYI > > -------- Original Message -------- > Subject: Re: [Fwd: concerns about the Southeast chapter] > Date: Mon, 27 Jul 2009 09:54:22 -0500 > From: John Christy > To: Thomas C Peterson > CC: Thomas R Karl > References: <4A534CF9.9080700@noaa.gov> > > > > Tom: > > I've been on a heavy travel schedule and just now getting to emails I've > delayed. I was in Asheville briefly Thursday for a taping for the CDMP > project at the Biltmore estates (don't know why that was the backdrop) > while traveling between meetings in Chapel Hill, Atlanta and here. > > We disagree on the use of available climate information regarding the > many things related to climate/climate change as I see by your responses > below - that is not unexpected as climate is an ugly, ambiguous, and > complex system studied by a bunch of prima donnas (me included) and > which defies authoritative declarations. I base my views on hard-core, > published literature (some of it mine, but most of it not), so saying > otherwise is not helpful or true. The simple fact is that the opinions > expressed in the CCSP report do not represent the real range of > scientific literature (the IPCC fell into the same trap - so running to > the IPCC's corner doesn't move things forward). > > I think I can boil my objections to the CCSP Impacts report to this one > idea for the SE (and US): The changes in weather variables (measured in > a systematic settings) of the past 30 years are within the range of > natural variability. That's the statement that should have been front > and center of this whole document because it is > mathematically/scientifically defensible. And, it carries more weight > with planners so you can say to them, "If it happened before, it will > happen again - so get ready now." By the way, my State Climatologist > response to the CCSP was well-received by legislators and stakeholders > (including many in the federal government) and still gets hits at > http://*vortex.nsstc.uah.edu/aosc/. > > There also was a page or so on the tropical troposphere-surface issue > that I didn't talk about on my response. It was wrong because it did > not include all the latest research (i.e. since 2006) on the continuing > and significant difference between the two trends. Someone was acting > as a fierce gatekeeper on that one - citing only things that agreed with > the opinion shown even if poorly done (e.g. Santer et al. 2008 used > ERSST data which I understand has now been changed in a way that > discredits the conclusion there, and Haimberger's v1.2-1.4 are clearly > spurious due to the error in ECMWF as published many places, but > analyzed in detail in Sakamoto and Christy 2009). The results of > Douglass et al. 2007 (not cited by CCSP) still stand since Santer et al. > 2008 asked a very different question (and used bad data to boot) than we > did. Our question was "Does the IPCC BEST ESTIMATE agree with the Best > Data (including RSS)?" Answer - No. Santer et al. asked, "Does ANY IPCC > model agree with ANY data set?" ... I think you can see the difference. > The fact my 2007 tropical paper (the follow-on papers in 2009 were > probably too late, but they substantiate the 2007 paper) was not cited > indicates how biased this section was. Christy et al. 2007 assessed the > accuracy of the datasets (Santer et al. did not - they assumed all > datasets were equal without looking at the published problems) and we > came up with a result that defied the "consensus" of the CCSP report - > so, it was doomed to not be mentioned since it would disrupt the > storyline. (And, as soon as RSS fixes their spurious jump in 1992, our > MSU datasets will be almost indistinguishable.) > > This gets to the issue that the "consensus" reports now are just the > consensus of those who agree with the consensus. The > government-selected authors have become gatekeepers rather than honest > brokers of information. That is a real tragedy, because when someone > becomes a gatekeeper, they don't know they've become a gatekeeper - and > begin to (sincerely) think the non-consensus scientists are just nuts > (... it's more comfortable that way rather than giving them credit for > being skeptical in the face of a paradigm). > > Take care. > > John C. > > p.s. a few quick notes are interspersed below. > > > Thomas C Peterson wrote: >> Hi, John, >> I didn't want this to catch you by surprise. >> Tom >> >> -------- Original Message -------- >> Subject: concerns about the Southeast chapter >> Date: Tue, 07 Jul 2009 09:25:45 -0400 >> From: Thomas C Peterson >> To: jim.obrien@coaps.fsu.edu >> CC: Tom Karl >> >> >> >> Dear Jim, >> >> >> First off and most importantly, congratulations on your recent >> marriage. Anthony said it was the most touching wedding he has ever >> been to. I wish you and your bride all the best. >> >> Thank you for your comments and for passing on John Christy's detailed >> concerns about the Southeast chapter of our report, /Global Climate >> Change Impacts in the United States/. Please let me respond to the key >> points he raised. >> >> In Dr. John Christy's June 23, 2009 document "Alabama climatologist >> responds to U.S. government report on regional impacts of global >> climate change", he primarily focused on 4 prime concerns: >> >> 1. Assessing changes since 1970. >> >> 2. Statements on hurricanes. >> >> 3. Electrical grid disturbances (from the Energy section). >> >> 4. Using models to assess the future. >> >> >> >> /1. Assessing changes since 1970./ >> >> The Southeast section has 5 figures and one table. One figure is on >> changes in precipitation patterns from 1901-2007. The next figure is >> on patterns of days per year over 90F with two maps, one 1961-1979, >> the other 2080-2099. One figure is on the change in freezing days per >> year, 1976-2007. The next figure is on changes to a barrier island >> land from 2002 to 2005. And the last figure was on Sea Surface >> Temperature from 1900 to the present. The table indicates trends in >> temperature and precipitation over two periods, 1901-2008 and >> 1970-2008. As Dr. Christy indicates in his paper, the full period and >> the period since 1970 are behaving differently. To help explain this, >> the table shows them both. Of the 5 figures, only one shows the >> changes over this shorter period. >> >> Since, as the IPCC has indicated, the human impact on climate isn't >> distinguishable from natural variability until about 1950, describing >> the changes experienced in the majority of the time since 1950 would >> be a more logical link to future anthropogenic climate change. In >> most of the report, maps have shown the changes over the last 50 >> years. Because of the distinct behavior of time series of >> precipitation and temperature in the Southeast, discussing the period >> since 1970 seemed more appropriate. Though as the figures and table >> indicate, this shorter period is not the sole or even major focus. > > See crux of the matter in email above - looking at the whole time series > is demanded by science. Any 30 or 50-year period will give changes - > blaming the most recent on humans ignores the similar (or even more > rapid) changes that occurred before industrialization (e.g. western > drought in 12th century). The period since 1970 WAS the major focus in > the SE section (mentioned 6 times in two pages). And, OF COURSE any > 30-year sub-period will have different characteristics than the 100-year > population from which it is extracted ... that doesn't prove anything. >> >> >> >> /2. Statements on hurricanes./ >> >> Dr. Christy takes issue with the report's statements about hurricanes >> and quotes a line from the report and quotes an individual hurricane >> expert who says that he disagrees with the conclusions. The line in >> the report that Dr. Christy quotes comes almost word for word out of >> CCSP SAP 3.3. While individual scientists may disagree with the >> report's conclusions, this conclusion came directly out of the >> peer-reviewed literature and assessments. Dr. Christy also complains >> that "the report did not include a plot of the actual hurricane >> landfalls". However, the section in the Southeast chapter discussing >> landfalling hurricanes states "see /National Climate Change/ section >> for a discussion of past trends and future projections" and sure >> enough on page 35 there is a figure showing land falling hurricanes >> along with a more in depth discussion of hurricanes. >> > You didn't read my State Climatologist response carefully - I mentioned > page 35 and noted again it talked about the most recent decades (and > even then, the graph still didn't go back to 1850). This hurricane > storyline was hit hard by many scientists - hence is further evidence > the report was generated by a gatekeeper mentality. >> >> >> /3. Electrical grid disturbances (from the Energy section)./ >> >> Moving out of the Southeast, Dr. Christy complains about one figure in >> the Energy Chapter. Citing a climate skeptic's blog which cites an >> individual described as the keeper of the data for the Energy >> Information Administration (EIA), John writes that the rise in weather >> related outages is largely a function of better reporting. Yet the >> insert of weather versus non-weather-related outages shows a much >> greater increase in weather-related outages than non-weather-related >> outages. If all the increases were solely due to better reporting, >> the differences between weather- and non-weather-related outages would >> indicate a dramatic decrease over this time period in non-weather >> related problems such as transmission equipment failures, earthquakes, >> faults in line, faults at substations, relaying malfunctions, and >> vandalism. >> >> Thanks to the efforts of EIA, after they took over the responsibility >> of running the Department of Energy (DOE) data-collection process >> around 1997, data collection became more effective. Efforts were made >> in subsequent years to increase the response rate and upgrade the >> reporting form. It was not until EIA's improvement of the data >> collection that the important decoupling of weather- and >> non-weather-related events (and a corresponding increase in the >> proportion of all events due to weather extremes) became visible. >> >> To adjust for potential response-rate biases, we have separated >> weather- and non-weather-related trends into indices and found an >> upward trend only in the weather-related time series. >> >> As confirmed by EIA, *if there were a systematic bias one would expect >> it to be reflected in both data series (especially since any given >> reporting site would report both types of events).* >> >> As an additional precaution, we focused on trends in the number of >> events (rather than customers affected) to avoid fortuitous >> differences caused by the population density where events occur. This, >> however, has the effect of understating the weather impacts because of >> EIA definitions (see survey methodology notes below). >> >> More details are available at: >> http://*eetd.lbl.gov/emills/pubs/grid-disruptions.html > > The data were not systematically taken and should not have been shown > .. basic rule of climate. >> >> >> >> /4. Using models to assess the future./ >> >> Can anyone say anything about the future of the Southeast's climate? >> Evidently according to John Christy, the answer is no. The basic >> physics of the greenhouse effect and why increasing greenhouse gases >> are warming and should be expected to continue to warm the planet are >> well known and explained in the /Global Climate Change/ section of the >> report. Climate models are used around the world to both diagnose the >> observed changes in climate and to provide projections for the >> future. There is a huge body of peer-reviewed literature, including a >> large number of peer-reviewed climate change assessments, supporting >> this use. But in Dr. Christy's "view," models should not be used for >> projections of the future, especially for the Southeast. The report >> based, and indeed must base, its results on the huge body of >> peer-reviewed scientific literature rather than the view of one >> individual scientist. > > No one has proven models are capable of long-range forecasting. > Modelers write and review their own literature - there are millions of > dollars going into these enterprises, so what would you expect? > Publication volume shouldn't impress anyone. The simple fact is we > demonstrated in a straightforward and reproducible way that the actual > trends over the past 30, 20, and 10 years are outside of the envelop of > model predictions ... no one has disputed that finding with an > alternative analysis - even when presented before congressional hearings > where the opportunity for disagreement was openly available. >> >> I hope this helps relieve some of your concerns. >> >> Regards, >> >> Tom Peterson >> >> >> > > > -- > ************************************************************ > John R. Christy > Director, Earth System Science Center voice: 256-961-7763 > Professor, Atmospheric Science fax: 256-961-7751 > Alabama State Climatologist > University of Alabama in Huntsville > http://*www.*nsstc.uah.edu/atmos/christy.html > > Mail: ESSC-Cramer Hall/University of Alabama in Huntsville, Huntsville AL 35899 > Express: Cramer Hall/ESSC, 320 Sparkman Dr., Huntsville AL 35805 > > > > -- > > *Thomas R. Karl, L.H.D.* > > Director, NOAA’s National Climatic Data Center > > Lead, NOAA Climate Services > > Veach-Baley Federal Building > > 151 Patton Avenue > > Asheville, NC 28801-5001 > > Tel: (828) 271-4476 > > Fax: (828) 271-4246 > > Thomas.R.Karl@noaa.gov > > > -- ---------------------------------------------------------------------------- Benjamin D. Santer Program for Climate Model Diagnosis and Intercomparison Lawrence Livermore National Laboratory P.O. Box 808, Mail Stop L-103 Livermore, CA 94550, U.S.A. Tel: (925) 422-3840 FAX: (925) 422-7675 email: santer1@llnl.gov ----------------------------------------------------------------------------