date: Wed Feb 25 13:09:59 2004
from: Phil Jones <p.jones@uea.ac.uk>
subject: Re: Trends in DTR from ERA-40
to: Adrian.Simmons@ecmwf.int

    Adrian,
        A couple of things have come up that need rapid attention, so I should get to your
    draft on Friday and the weekend. If I go through and put my comments on my copy and
    then email this back - will that be OK ? I say this as I've got into the habit of adding
    comments to a word version. You might be using a different system as you've sent a
    pdf file.
        I think hard copy comments should work.
    Cheers
    Phil
   At 09:26 25/02/2004 +0000, you wrote:

     Phil
     The SSTs in the simulation were the same analyses as used in ERA-40. So the simulation
     can be regarded as an ERA-40 analysis without atmospheric, land-surface and ocean-wave
     observations, but still, effectively, using observations of SST. Obviously that provides
     quite a control.
     The analyses were from the Hadley Centre for the period up to November 1981 and from
     NCEP thereafter. The switch was made because NCEP provided higher time frequency. There
     is a report on this in an advanced stage of preparation.
     Best regards
     Adrian
     Phil Jones wrote:

      Adrian,
         I've been away, but I've printed off the draft paper and the recent set of figures.
     I will
      go through these when I've cleared a few pressing things in the next couple of days.
      A quick look at the figures seems to indicate that ERA-40 is much better than NCEP
      and a paper would be a useful addition to the literature.
         I will contact the people at NCDC about the max/min series and possible updates to
      nearer the present as well as a gridded dataset.
         I have a couple of quick questions:
       - the specified SST run, is this the same as ERA-40 but with SSTs just having the same
      annual cycle year on year? The proper ERA-40 run used SSTs as they were measured -
      a dataset developed at the HC.
      Cheers
      Phil
     At 13:08 24/02/2004 +0000, Adrian Simmons wrote:

     Phil
     I've now computed a set of monthly means of daily max and min temperatures from ERA-40.
     They are not based directly on any observations of Tmax and Tmin. Instead they are
     derived from maximum and minimum values computed during each of the six-hour background
     forecasts of the data assimilation system.
     The first attachment shows trends in the daily temperature range for the periods
     1958-2001 and 1979-2001. This shows a reduction over time for North America (1958-2001
     at least) and for eastern Europe and much of Asia. Values in the tropics must be treated
     with most caution as we know we have spurious trends in some aspects of the hydrological
     cycle which could well impact the daily temperature range. The trend over the US may be
     biased a bit by a too-dry troposphere from 58-63 caused by wrongly encoded humidity
     observations in one of the radiosonde datasets supplied to us - we picked this up too
     late to allow correction in ERA-40 production.
     The other plots show the trends in maximum and minimum temperatures for 1958-2001 and
     1979-2001. Overall the short-range forecast trends are quite consistent with the trends
     for the daily mean analyses (Fig 7 in the document attached to yesterday's mail). The
     increase in daily temperature range over northern latitudes is seen to be associated
     with a larger rate of increase of minimum than maxiumum temperatures - at least we seem
     to have got the most basic feature right.
     The Tmax and Tmin trends for 1958 to 2001 also show that the almost certainly erroneous
     cooling trends seen in the daily mean analyses for Australia and much of tropical South
     America arise much more from a decrease over time in maximum temperatures than from a
     decrease in minimum temperatures. This suggests that the ERA-40 background forecasts
     (and "analyses") of daytime (rather than nightime) temperatures are the ones that are
     biased warm in the earlier years in which we assimilated little or no SYNOP data over
     these areas.
     Western Europe seems to show a slightly larger trend in maxiumum than minimum
     temperatures. Hotter summers and drier soils?
     It will be interesting to make more quantitative comparisons with the new data from
     NCDC.
     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
     ----------------------------------------------------------------------------

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