date: Fri Aug  5 15:34:58 2005
from: Phil Jones <p.jones@uea.ac.uk>
subject: Re: Correlation values (fwd)
to: Anders Moberg <anders@misu.su.se>, David Lister <D.Lister@uea.ac.uk>

    Anders,
       David's shown me the plot and forwarded the email. The plot is showing what
    we thought it would. Before you do others though I thought I should comment
    upon it.
       As the plotted points (which get joined up) are correlations between the
    average Tx series for winter and the 2/5/10 and 90/95/98 extreme series, I was
    wondering whether you should be joining up the 10 and 90 points. If we had a
    50th percentile, then that should be correlated with the average to almost 1.
    Obviously it would be silly to plot 50, but maybe a break in the lines between
    10 and 90 is preferable.
       Just a thought when I first saw it.
     It clearly shows that the cold tail for all three points has slightly higher
    correlation than the warm tail, for the average value (black) and most of
    the grey individual stations. There are a few places though where this
    isn't the case. It would be good to say if these were in one part of Europe,
    if they are not randomly located.  This situation might differ for other
    variables and seasons.
    Cheers
    Phil
   At 15:20 05/08/2005, you wrote:

     ---------- Forwarded message ----------
     Date: Fri, 05 Aug 2005 15:39:18 +0200
     From: Anders Moberg <anders@misu.su.se>
     To: David Lister <D.Lister@uea.ac.uk>
     Subject: Re: Correlation values
     Hi David,
     The attached figure shows the kind of plots I intend to make from your
     correlation tables. So, once I have the rest of the tables (I conclude that
     the file you sent were for winter?), I can finish the figures. This,
     however, will not happen during next week which is my final summer holiday
     week. In other words, it is no hurry with sending me any more files now. I
     will be back at work on Monday, August 15. I will not read emails before
     then, so if you have any new plots/tables for me you can send them all on
     Friday, August 12, if you like.
     Just in case, I repeat which kinds of correlation tables I need:
     1. correlation between MEANTX and TX2P, TX5P, ..., TX98P (i.e. 6
     correlations per station)
     2. the same for TN
     3. correlation between PRECTOT and PREC90P, ..., PREC98P (3 correlations
     per station)
     4. correlation between SDII and PREC90P, ..., PREC98P (3 correlations per
     station)
     All these, of course, both for winter and summer. For the paper, I guess I
     will have to make a representative selection of the correlation results.
     Have a good time until we hear from each other again (I guess/fear there
     will be lots of of regional graphs to look at on August 15...)
     Cheers,
     Anders
     At 12:02 2005-08-05 +0100, you wrote:

     Okay Anders,
     More to follow after lunch .......
     Cheers
     David
     On Fri, 5 Aug 2005, Anders Moberg wrote:
     > David,
     >
     > Thanks, the table looks good. I don't need to know about the significances.
     > I think these are anyway not very useful numbers, as all correlations
     > calculated here are clearly made between dependent variables. I rather
     > regard the correlations as objective measures of the degree of similarity
     > between the variables. You are right about the lats/lons.
     >
     > Cheers,
     > Anders
     >
     >
     > At 11:47 2005-08-05 +0100, you wrote:
     > >Anders,
     > >
     > >I have attached a file that gives the correlation values for MEANTX with
     > >the other TX extreme variables.  Please have a look at it and let me know
     > >if this is okay ......... whilst most/all correlations are significant
     > >(5%), do you need to know which are and which are not?  I guess that you
     > >do not need station lat./lons.
     > >
     > >Cheers
     > >
     > >David
     > >
     > >
     > >On Fri, 5 Aug 2005, Anders Moberg wrote:
     > >
     > > > David,
     > > >
     > > > When you do the multi-plot map for winter precip - could you omit
     the brown
     > > > circle in the legend as there are no <-80 trend  in winter? (just
     to avoid
     > > > annoying someone who likes spotting silly details...)
     > > >
     > > > Anders
     > > >
     > > >
     > >***********************************************************************
     ****
     > >David Lister
     > >Climatic Research Unit
     > >University of East Anglia
     > >NORWICH   NR1 7TJ
     > >United Kingdom
     > >
     > >Telephone +44 (0)1603 593818
     > >Fax       +44 (0)1603 507784
     > >
     > >CRU web site  [1]http://www.cru.uea.ac.uk/
     >
     >

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

