date: Mon Jul 18 09:44:11 2005 from: Phil Jones subject: Re: Text and CQ stuff to: Kevin Trenberth , Kevin Trenberth , Peter Ambenje , Roxana Bojariu , David Easterling , Fatemeh Rahimzadeh , Jim Renwick , Matilde Rusticucci , Brian Soden , Panmao Zhai , Albert Klein Tank , David Parker , "Solomon, Susan" Kevin, David has all the numbers. I'll be down in Exeter tomorrow and Wednesday. I'll try and find a few minutes to talk to him about it. Cheers Phil At 22:59 16/07/2005, Kevin Trenberth wrote: Hi all I have started going thru the text a bit more thoroughly. At present the description of the global mean temperature record is for a warming of 0.6C during the 20th Century. That is the linear reprducible value. But it is not a useful value as the trend is not linear. In the recent paper by Raper et al on SST they make a point to give values for both the linear trend and the change from the low pass filtered record. The latter is quite a bit bigger. I would like to see us adopt something similar. The question then is how to characterise the record. Here is my attempt: words However, the record is best characterized as level prior to about 1920, a warming to 1940 or so, leveling out or even slightly decreasing until 1970, and a fairly linear trend since then. Going by the low pass filtered data, the overall warming through 2005 is 0.75ºC, with 0.5ºC increase occurring after 1970. To illustrate this I tried to capture the sense of this in the accompanying ppt. There are two slides. Make sure you go into slide show mode to view them. You will see the first has a smoothed trend the second has linear segments that join. The idea is to also capture the overall error bars to a reasonable degree, as you can see. In fact this could be linked to the modeling and attribution chapter to say that the warming in the first part of the 20th century was partly due to solar, the cooling from 1940 to 1970 to increased aerosol, and the warming after 1970 to the increasing GHGs. This could work very well as part of the CQ2. Ideally the background global mean values should not have the red bars on it but should just be a time series with error bars. The curves which I fitted by eyeball using power point should be done more rigorously, perhaps using a cubic spline fit with strong tension., or a series of segments with divides at 1940 and 1970. Then a linear value with the given starting point could be determined for both the mean and both end of the error bars. I am seeking feedback on this idea. 1) Is it a good idea and has your support? 2) Any comments or suggestions? 3) Any volunteers to do it more rigorously? Any such person would need the mean and error bars to do this from David or Phil? 4) Do you prefer the straight lines or smoothed values? Thanks Kevin -- **************** Kevin E. Trenberth e-mail: [1]trenbert@ucar.edu Climate Analysis Section, NCAR [2]www.cgd.ucar.edu/cas/ P. O. Box 3000, (303) 497 1318 Boulder, CO 80307 (303) 497 1333 (fax) Street address: 1850 Table Mesa Drive, Boulder, CO 80303 Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ----------------------------------------------------------------------------