Guest Post By John R. Christy of the University of Alabama at Huntsville
via Dr. Roger Pielke Sr’s blog: Climate Science
The three warm-color time series are taken from Hansen’s published testimony in June 1988 in which global surface air temperatures were projected under three scenarios by his global climate model.
The red curve follows a scenario (A) of continued emissions growth based on the previous 20 years before 1988 (which turned out to be an underestimate of actual emissions growth.) The orange represents a scenario (B) of fixed emissions at the rate achieved in the 1980s. The yellow curve portrays a scenario (C) in which “a drastic reduction” in GHG emissions is assumed for 1990-2000. The observations are global tropospheric temperatures adjusted to mimic the magnitude of surface temperature variability and trends according to published climate model simulations (i.e. a reduction in satellite anomalies by 0.83.)
After tying all time series to a 1979-83 reference mean, one can see the significant divergence in the results. (Notes: 1. observed 2010 is Jan-Jul only; 2.) tropospheric temperatures are used as the comparison metric due to many uncertainties and biases in the surface temperature record, i.e. Klotzbach et al. 2009, 2010 ; 3.) both models and observations included the 1982 eruption of El Chichon while B and C scenarios included a volcano in the mid 1990s – not too different from Mt. Pinatubo.)
The result suggests the old NASA GCM was considerably more sensitive to GHGs than is the real atmosphere since (a) the model was forced with lower GHG concentrations than actually occurred and (b) still gave a result that was significantly warmer than observations.
Dr. Christy,
Thanks for your reply in which you addressed some of my comments. Here are some further comments on your responses.
Okay…I know you have to line things up somehow but I am just wondering if that relative shift could be somewhat of a problem.
Well, this assumes that the satellite data is perfect…I.e., the remaining discrepancy in trends between the satellite data and the surface temperatures is all the fault of the surface temperatures (and that the 1.2 factor is correct). I think that is a questionable assumption, especially given the history of the satellite data analysis corrections over the years. [There is also apparently some issue of whether the trend that should be compared to is the land meteorological stations trend, the only one available at the time of Hansen’s prediction, or the Land + Ocean trend.]
I don’t think that is a completely fair accounting of the history. There were also several corrections that needed to be made to the satellite data analysis…and these corrections were mainly in one direction, i.e., they made the temperature trend more positive than what was originally reported. A while ago, in response to a thread here, I actually went back and computed the trend with the current version of your UAH analysis over various past time intervals and compared it with the trends you had reported in older papers (actually focussing mainly your 1998 paper http://ams.allenpress.com/archive/1520-0442/11/8/pdf/i1520-0442-11-8-2016.pdf ) in order to see what part of the difference in trend was due to the longer data series and what part was due to changes in the analysis. My conclusion at the time was that the two effects where approximately equal. I.e., that about half the change in the trend between that 1998 paper and now was due to a longer time series and half was due to fixing the errors in the analysis. Here is what I said at the time in detail based on my trend calculations from your data that I did in Excel:
(My calculation was performed in January 2009. As I understand it, you have made some further small changes to your analysis since then too, although I thought you concluded that they don’t affect the long-term trends very much?)
I noticed that my link to the 1998 is no longer working. Here is one that seems to: http://journals.ametsoc.org/doi/full/10.1175/1520-0442%281998%29011%3C2016%3AAOTMPF%3E2.0.CO%3B2 Just to be clear, that 1998 paper reports a new trend of -0.046 C / decade for the January 1979–April 1997 period but notes that this is a change from -0.076 C / decade over the same period with the previous analysis procedure. As of January 2009 when I looked at the data, the latest version of the UAH analysis gives a trend of +0.029 C / decade for the January 1979–April 1997 period.