Update (Jul 28, 2008): On Jan 18, 2008, two days after this article was posted, RSS issued a revised version of their data set. The graphics below are based on RSS versions as of Jan 16, 2008, the date of this article, and, contrary to some allegations on the internet, I did not “erroneously” use an obsolete data set. I used a then current data set, which was later adjusted, slightly reducing the downtick in observations. On Jan 23, 2008, I updated the graphic comparing Hansen projections using the revised RSS version. Today I re-visited this data, posting a further update of this data including the most recent months. While some commentators have criticized this post because the RSS adjustment reduced the downtick slightly, the downtick based on the most recent data as of July 28, 2008 is larger than the RSS adjustment as of Jan 2008.)
In 1988, Hansen made a famous presentation to Congress, including predictions from then current Hansen et al (JGR 1988) online here . This presentation has provoked a small industry of commentary. Lucia has recently re-visited the topic in an interesting post ; Willis discussed it in 2006 on CA here .
Such discussions have a long pedigree. In 1998, it came up in a debate between Hansen and Pat Michaels (here); Hansen purported to rebut Crichton here, NASA employee Gavin Schmidt on his “private time” supported his NASA supervisor, Jim Hansen here , NASA apologist Eli Rabett believed to be NASA contractor Josh Halpern here . Doubtless others.
It seems like every step of the calculation is disputed – which scenario was the “main” scenario? whether Hansen’s projections were a success or a failure? even how to set reference periods for the results. I thought it would be worthwhile collating some of the data, doing chores like actually constructing collated versions of Hansen’s A, B and C forcings so that others can check things – all the little things that are the typical gauntlet in climate science.
I’ll compare this graphic with some other versions. On another occasion, I’ll discuss the forcings in Hansen et al 1988. First, I’m going to review the prior history of this and related images.
Hansen et al 1988
Hansen et al 1988 defined 3 scenarios (A,B, C), illustrated in the two graphics below taken from Figure 3 of the original article. Each scenario described forcing projections for CO2, CH4, N2O, CFC11, CFC12 and the other Montreal Protocol traces gases as a group. In subsequent controversy, there has been some dispute over which scenario was Hansen’s “primary” scenario. In the right panel, only Scenario A is taken through to 2050 and in both panels, Scenario A is plotted as a solid line, which could be taken as according at least graphic precedence to Scenario A. [following sentence revised at Jan 17, 2008 about 9 am] Despite the graphic precedence to Scenario A in the right panel graph, Hansen mentioned in the running text (9345):
Scenario A, since it is exponential must eventually be on the high side of reality in view of finite resource constraints, even though the growth of emissions (`1.5% per year) is less than the rate typical of the past century (~4% per year).
and, then inconsistently with the graphic shown on the right side only showing Scenario A out to 2050, said (p 9345) that Scenario B was “more plausible”, an aside that subsequently assumed considerable significance.
Hansen Debate 1998
In 1998, 10 years after the original article, in testimony to the U.S. Congress and later in the debate with Hansen, Pat Michaels compared observed temperatures to Scenario A, arguing that this contradicted Hansen’s projections, without showing Scenarios B or C.
[Update: Jan 17 6 pm] To clarify, I do not agree that it was appropriate for Michaels not to have illustrated Scenarios B or C, nor did I say that in this post. These scenarios should have been shown, as I’ve done in all my posts here. It was open to Michaels to take Scenario A as his base case provided that he justified this and analysed the differences to other scenarios as I’m doing. Contrary to Tim Lambert’s accusation, I do not “defend” the exclusion of Scenarios B and C from the Michaels’ graphic. This exclusion is yet another example of poor practice in climate science by someone who was then Michael Mann’s colleague at the University of Virginia. Unlike Mann’s withholding of adverse verification results and censored results, Michaels’ failure to show Scenarios B (and even the obviously unrealistic Scenario C) was widely criticized by climate scientists and others, with Klugman even calling it “fraud”. So sometimes climate scientists think that not showing relevant adverse results is a very bad thing. I wonder what the basis is for climate scientists taking exception to Michaels, while failing to criticize Mann, or, in the case of IPCC itself, withholding the deleted Briffa data. [end update]
In any event, in the debate, Hansen responded irately, arguing that Scenario B, not shown by Michaels, was his preferred scenario, that this scenario was more consistent with the forcing history and that the temperature history supported these projections.
Pat [Michaels] has raised many issues, a few of which are valid, many of which are misleading, or half truths, some of which are just plain wrong. I don’t really intend to try to respond to all of those. I hope you caught some of them yourself. For example, he started out showing the results of our scenario A, even though the scenario that I used in my testimony was scenario B…
I don’t have a copy of his testimony and can’t say at this point whether Scenario B was or was not used in the 1988 testimony.
[Update: The testimony is now available and Hansen’s statement that Scenario B was “used” in his 1988 testimony is very misleading: Hansen’s oral testimony called Scenario A the “Business as Usual” scenario and mentioned Scenario B only in maps purportedly showing extraordinary projected warming in the SE USA
On the other hand, Hansen also testified in November 1987 online here and, in that testimony (but not in the 1988 testimony), he did say that Scenario B was the “most plausible”, though in a context that was over a longer run than the 10-20 year periods being discussed here. At present, I do not understand how the trivial differences between Scenario A and B forcings over the 1987-2007 period can account for the difference in reported Scenario A and B results, that’s something I’m looking at – end update]
In any event, Hansen argued in 1998 that real world forcings tracked Scenario B more closely and that warming was even more rapid than Scenario B:
There were three scenarios for greenhouse gases, A, B and C, with B and C being nearly the same until the year 2000, when greenhouse gases stopped increasing in scenario C. Real-world forcings have followed the B and C greenhouse scenario almost exactly. … and the facts show that the world has warmed up more rapidly than scenario B, which was the main one I used.
Hansen and Schmidt
In one of the first realclimate posts (when they were taking a temporary respite from their active Hockey Stick defence of Dec 2004-Feb 2005), NASA employee Gavin Schmidt here tried to defend the 1988 projections of his boss (Hansen) against criticism from Michael Crichton. Hansen issued his own defence here , covering similar ground, citing the realclimate defence by his employee (Schmidt), done in Schmidt’s “spare time”. On the left is the version from realclimate updating the history to 2003 or so; on the right is the image from Hansen’s article, updating the instrumental history to 2005 or so.
NASA employee Schmidt said of his boss’ work (not mentioning in this post that both were NASA employees, although Schmidt’s online profile in the About section says that he is a NASA employee):
The scenario that ended up being closest to the real path of forcings growth was scenario B, with the difference that Mt. Pinatubo erupted in 1991, not 1995. The temperature change for the decade under this scenario was very close to the actual 0.11 C/decade observed (as can be seen in the figure). So given a good estimate of the forcings, the model did a reasonable job.
Hansen re-visited the predictions in Hansen et al (PNAS 2006) – this is his “warmest in a millllll-yun years” article – where he updated the temperature history to 2005, introducing a couple of interesting graphical changes. In this version, the color of Scenario A was changed from red (which is visually the strongest and most attention-grabbing color) to a softer green color. He also plotted two instrumental variants – the Land only and the Land+Ocean histories. Hansen argued that Scenario B was supported by the data and, continuing his feud with Crichton, asserted that his results were not “300% wrong”, footnoting State of Fear. NASA employee Schmidt loyally continued to support his boss on his”spare time” at realclimate, once more visiting the dispute at realclimate in May 2007, re-issued the graphic in substantially the same format as Hansen’s 2006 article, with further changes in coloring. It seems likely that Schmidt, as a NASA employee, had access to the digital version that Hansen used in his 2006 paper and, unlike us at CA, did not have to digitize the graphics in order to get Hansen’s results for the three scenarios. Schmidt stated:
My assessment is that the model results were as consistent with the real world over this period as could possibly be expected
Re-constructing the Graphic
While NASA employee Schmidt, in his “spare time”, has access to NASA digital data, we at CA do not. Willis Eschenbach digitized Hansen scenarios A,B and C to 2005 and I extended this to 2010. The three scenarios to 2010 are online here in ASCII format: http://data.climateaudit.org/data/hansen/hansen_1988.projections.dat
In the graphic shown above, I’ve compared the NASA scenarios to two temperature reconstructions: the NASA global series and the RSS satellite series (not showing CRU and UAH versions to reduce the clutter in the graphic a little.) The GISS temperature history has been used in the prior graphics and the RSS version has been preferred by IPCC and the US CCSP.
In order to plot the series, several centering decisions have had to be made. The GISS GLB (and other series) is centered on 1951-80, while the three scenarios were centered on the control run mean. The 1951-1980 means of the three scenarios, which included forcings for the period, were thus higher than the 1951-80 zero for the target temperature series by 0.1 deg C for Scenario A and 0.07 deg C for Scenarios B and C. The scenarios were only available in digital form for 1958-1980. For the GISS GLB series, there was negligible (less than 0.01 deg C) difference between the means for 1958-1980; 1958-1967 and 1951-1980.
In order to put the three Scenarios apples and apples to the GISS GLB temperature series (basis 1951-1980), I re-centered the three scenarios slightly so that they were also zero over 1958-1967. This lowered the projections very slightly relative to the instrumental temperatures. (Lucia recognized the problem as well and dealt with it a little differently). I applied a similar strategy with respect to the satellite series which did not commence until 1979. In this case, I re-centered it so that its 1979-1988 zero matched the 1979-1988 values of the GISS GLB series. This yielded the diagram shown above:
I’ll talk some more about the forcings in a day or two.
I’ve shown 1987, the last year of data for Hansen et al 1988. The 2007 RSS satellite temperature was 0.04 deg C higher than the 1987 RSS temperature and there was substantial divergence between Scenario B in 2007 and the RSS satellite temperature (and even the GISS temperature surface temperature series). Strong increases in the GISS Scenarios start to bite in the next few years. To keep pace, one must really start to see increases in the RSS troposphere temperatures of about 0.5 deg C. sustained over the next few years.
The separation between observations and Scenario C is quite intriguing: Scenario C supposes that CO2 are stabilized at 368 ppm in 2000 – a level already surpassed. So Scenario C should presumably be below observed temperatures, but this is obviously not the case.
It should be noted that Hansen et al 1988 considers other GHGs (CH4, N2O, CFC11 and CFC12). Methane is a curious situation as methane concentrations may have stabilized – making the SRES methane projections in even the A1B case possibly very problematic.