date: Fri Aug  5 16:12:50 2005
from: Phil Jones <p.jones@uea.ac.uk>
subject: Re: DTR trends
to: "Russell Vose" <Russell.Vose@noaa.gov>, Adrian.Simmons@ecmwf.int

    Adrian,
       Russ and me have been having some email exchanges about which of your
    two datasets we need. We have maps of DTR from Russ (the one he emailed
    you an hour or so ago) and from Aiguo Dai and Lisa Alexander.
       I've been concentrating on Australia (as we have maps of trends from BoM
    for 1950-03 and 1979-03 (attached). Maybe Australia isn't the best place to
    look but we're confident BoM is likely to be right.
       Of the time series you sent, Australia shows the largest difference between the
    0-6h forecast and the 6 hrly analyses, particularly from about 1990. This divergence
    since 1990 also occurs in other regions although not over North America nor much
    over Europe, which suggests it is something to do with the data. This divergence
    isn't an issue in the mean temp from the 2004 paper, so any thoughts.
         Russ thinks the 6 hrly analyses look better compared to his plot, but I reckon
    that the 0-6hr forecasts are better in some regions.
       So, is it possible to send both datasets. Russ can cope with the format.
    Cheers
    Phil
   At 14:31 05/08/2005, Russell Vose wrote:

     Adrian:
     Many thanks for these figures.  At first glance it appears that my analysis for
     1979-2004 and yours for 1979-2001  (6-hrly analysis) exhibit some general agreement.
     It's not a perfect match, but both suggest that many areas (western U.S., Europe,
     western Asia, Australia) experienced an increase in DTR since 1979.
     Do you have a global (land surface) trend for 1979-2001 for the 6-hrly analysis?
     Adrian Simmons wrote:

     Phil
     Attached are a set of results based on DTR computed both from the max/min temperatures
     in the four daily 0-6h forecasts and from the max/min of the analysed temperatures for
     00, 06, 12 and 18UTC. No mask based on CRUTEM2V has been applied, but I've masked out
     sea points, based on selecting only 5deg boxes that are more than 50% land.
     The first two attachments show maps of the least-square linear trends in DTR computed
     for 1958-2001 and 1979-2001. The third attachment shows time series of some area
     averages. These are plotted as deviations from the 44-year mean, and smoothed by a
     12-month running average. Clearly there are some things here that must be treated with
     much suspicion - the drop from the 60s to the mid-70s in the tropics (and southern
     hemisphere?) for example is suggestive of a shift in the assimilation associated with
     change in data coverage. perhaps, though, there is something useful in the information
     for Europe (at least from 1967 onwards after the SYNOP coverage improves) and North
     America.
     Best regards
     Adrian
     Phil Jones wrote:

      Adrian,
         Is it a trivial exercise for you to work a trend from land regions of DTR (annual)
      for 1979-2001?  I'm just after a number. Does ERA-40 have DTR still going down?
      CRU mask not essential. Omit Antarctica.
         If it isn't trivial then don't bother.
      Cheers
      Phil
     At 17:04 03/08/2005, Adrian Simmons wrote:

     Phil, Russell
     As promised earlier, attached is the first of the datasets. It is the set of
     monhtly-averages of the daily maximum two-metre temperature, computed from the four 0-6h
     forecasts per day carried out from the ERA-40 analyses.
     Once you have run gunzip on the dataset, you should be left with a simple character
     file, suitable for a formatted read into a FORTRAN program. Looking at the file with any
     editor, you should see a sequence comprising a month identifier (format is I4,I2)
     showing YYYYMM and then the field itself. Each row (format 72f7.2) contains the
     temperature in Kelvin for grid boxes running east from the dateline, ie for longitudes
     180W-175W, 175W-170W, 170W-165W,..., 175E-180E, as in the CRUTEM2V dataset downloadable
     from the CRU website. There are 36 rows running from north to south, ie for latitudes
     90N-85N, 85N-80N,80N-75N,...,85S-90S. The months run from 1958 1 to 200112.
     If you can read the data without problem, I'll send the corresponding monthly-average
     minimum temperatures computed the same way, and the alternative monthly averages of the
     max/min temperatures computed each day from the 00, 06, 12 and 18UTC analyses. I'll send
     one field per email to avoid hitting possible mailbox limits.
     Best regards
     Adrian
     --
     --------------------------------------------------
     Adrian Simmons
     Head of Data Division
     European Centre for Medium-Range Weather Forecasts
     Shinfield Park, Reading, RG2 9AX, UK
     Phone: +44 118 949 9700
     Fax:   +44 118 986 9450
     --------------------------------------------------

     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
     ----------------------------------------------------------------------------

     --
     Russell S. Vose, Chief
     Climate Analysis Branch
     National Climatic Data Center
     151 Patton Avenue
     Asheville, North Carolina 28801
     Phone: (828) 271-4311
     Fax: (828) 271-4328
     E-mail: Russell.Vose@noaa.gov

   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
   ----------------------------------------------------------------------------
