cc: Ben Santer , "Karl E.Taylor" , Keith Briffa , Jerry Meehl , Jerry Mahlman , Steve Smith , rrichels , Richard Moss , Steve Schneider , Bob Harriss , Bob Watson , Bill Gutowski , Malcolm Hughes , Alan Robock , Caspar M Ammann , Chick Keller , covey1@llnl.gov, DANNY HARVEY , Dave Schimel , Haroon Khehsgi , "James A. (Jae) Edmonds" , Jane Leggett , Joel Smith , Linda Mearns , Martin Manning , Martin Parry , Kevin Trenberth , "Michael E. Mann" , Michael Oppenheimer , Michael Schlesinger , Mike Hulme , Mike MacCracken , Peter Backlund , Phil Jones , Raymond Bradley , Tim Osborn , "tim.carter" , Tim Carter , Naki Nakicenovic , Hugh M Pitcher , Warren Washington , Ron Stouffer , Steve Fetter , "simon.shackley" , Tom Wigley , grubler@yale.edu, riahi@amazon.iiasa.ac.at date: Sun, 16 Nov 2003 11:16:04 -0500 from: Marty Hoffert subject: Re: brief question to: Nebojsa Nakicenovic , Tom Wigley Naki & Tom et al: It may not be useful to express uncertainties of future climate change in terms of combined uncertainties of atmospheric physics and SRES emissions projections based on forecasts of social-economic-technology evolution in the 21st century. The SRES authors were right in my opinion not to assign probabilities to their 40 scenarios. The mere existence of a possible emission path shouldn't effect climate change uncertainty estimates if it's probability can't be estimated by a rational and tested methodology. Sorry, Steve Schneider, I usually agree with you, but not on this. Issac Asimov, in his classic SF Foundation novels -- space operas about a future human galactic civilization roughly modeled on "The Decline and Fall of the Roman Empire" -- invented the hypothetical science of "psychohistory " with which statistical probabilities of different futures were calculable. We'd probably call these these probability distribution functions (pdfs) now. Asimov's idea was plausible in the 50s and 60s, before we knew about nonlinear dynamics and chaos theory. Paleontologist Steven Jay Gould, in Wonderful Life: The Burgess Shale and the Nature of History, reminds us that history is contingent on apparently small random events with large future consequences. Rewind the tape of life and replay it, and evolution might not lead to Home sapiens sapiens, and we wouldn't be here asking these questions. A counter-argument is that weather is unpredictable beyond a short time horizon; but climate (perhaps an "attractor" for weather states) is predictable. That's why we're in this business. So might the probability of our future carbon emissions be predictable, in principle. Proponents of the "Many-Worlds" interpretation of quantum mechanics (see, e.g., J. Richard Gott's Time Travel in Einstein's Universe) would argue that all 40 SRES storylines, along with countless others, actually exist in a spacetime multiverse of parallel universes, into which reality is constantly splitting. (In the classical quantum mechanics experiment, and according to "Many Worlds," a single photon goes through two separate slits by splitting into parallel universes and reemerging in ours). The wavefunctions of the SRES parallel universes are roughly analogous to their pdfs. I think you will all agree that we're a long way from being able to compute these, though one can't rule it out in the fullness of cosmic time. It does pose some interesting problems relating to free will, the role of humans in the cosmos, etc. It seems more productive to focus on atmospheric science uncertainties -- like climate sensitivity -- which we can estimate not only from climate models but from paleoclimate records (see, e.g., Hoffert & Covey attached). These haven't changed for decades from the 1.5 to 4.5 degrees Celsius steady state warming for a CO2 doubling (this uncertainty range resulting mainly from cloud radiative feedback uncertainties). These can be linked to allowable carbon emissions and CO2-emitting energy production and energy demand implied by projected growth of GDP and declining energy intensity (E/GDP). The shortfall between the allowable energy from fossil fuels necessary to keep global warming below some specified level and total energy demand has to be made up by new emission-free-energy technologies, including the effect of climate sensitivity uncertainties. An analysis of this problem by Ken Caldeira, Atul Jain, and me was published last March in Science (attached). Comments most welcome. Cheers, Marty Hoffert Professor of Physics Andre and Bella Meyer Hall of Physics Room 525, Mail Code 1026 4 Washington Place New York University New York, NY 10003-6621 NYU Phone: 212-998-3747 NYU Fax: 212-995-4016 Home Phone: 516-466-9418 Home Fax: 516-487-0734 Cellphone: 516-972-4779 Email: marty.hoffert@nyu.edu Web page: http://www.physics.nyu.edu/people/hoffert.martin.html At 9:34 PM +0100 11/14/03, Nebojsa Nakicenovic wrote: Dear Tom and Colleagues, It is not easy to respond to your request. At face value, my reaction to Q1 is YES and to Q2 and undecided MAYBE. However, the original statement at which the questions are directed is at best misleading. In my view it is simply wrong. First, storylines have little to do with the actual SRES emissions. The primary determinant of the ranges of emissions (the full range and not only the upper values) was primarily the body of emissions scenarios literature. Storylines were used as a tool for framing the driving forces and their relationships for six integrated assessment models that developed the scenarios. In TAR, nine different integrated models were used. There are probably two dozen or so multi-regional integrated models in the world. The six SRES (and nine TAR) models were representative of different modeling approaches including bottom-up and top-down. Thus, jointly they do not have any obvious upward or downward bias with respects to future emissions. It is quite curious to see this research effort reduced to "storylines" versus "science" which appears to imply that all the demographic, economic, engineering, etc. approaches that converge in integrated models, emissions scenarios and the underlying literature are labelled as something else than "science". I challenge anyone to produce a literature assessment that results in reduced range of future emissions compared to SRES. In fact, the opposite would be the case, the full range in the literature is broader than SRES. The 40 SRES scenarios jointly cover about 5th to 95th percentile of the frequency distributions of driving forces and emissions in the literature. This unequivocally means that the highest emissions scenarios are substantially higher than SRES and that there are some mitigation scenarios that are substantially lower than the SRES range. The question (Q3) that I would like to ask is what is the "new science" that is so much different from SRES and other scenarios in the literature? As we all know, it was not the SRES scenarios that determined the range of temperature change, but rather indeed GCMs and simple climate models. My understanding is that only about half of this uncertainty is due to scenarios and the other half is due to the climate uncertainties. Curiously, the difference is that the emissions uncertainties can be reduced through mitigation. In other words, about half of the TAR temperature range could be reduced through mitigation measures and policies. Best regards, Naki At 09:17 AM 11/14/2003 -0700, Tom Wigley wrote: Dear all, I have had a disagreement with someone about a statement they made in which I was mentioned. When I read this, I thought it implied that I was endorsing their view. The statement is given below together with two questions. For each question all I want is a YES, NO or MAYBE answer ..... >For the record, while we think TAR erred in allowing new storylines >rather than new science (as Tom Wigley has pointed out) to drive a >new upper limit to the temperature range ... Q1: Do you think this implies that I endorse the claim that the TAR (i.e., IPCC) erred? Q2: Do you think this amounts to an accusation that the TAR (IPCC) used the SRES scenarios because they produced a higher upper-bound warming than previously? In question 2, I am not asking about the truth of the 'accusation', but whether or not the statement could be construed as an accusation. The key word in the statement is 'allowing'. Thanks for your response, Tom. Attachment Converted: "c:\eudora\attach\Caldeira_et alScience .pdf" Attachment Converted: "c:\eudora\attach\Hoffert&Covey_Nature-92.pdf"