date: Mon Sep 22 14:12:56 2008
from: Phil Jones <p.jones@uea.ac.uk>
subject: Re: Downward trend in relative humidity over land?
to: Adrian.Simmons@ecmwf.int, Kate Willett <kate.willett@metoffice.gov.uk>

    Adrian,
       Thanks for the plots.
    Kate didn't see much happening with RH, but the 1-1.5% reduction is
    probably too small to be that clear. It may also not be statistically
    significant either.  Kate's series finished in 2003, so the last 3 years
    may be a factor as you say.
     The increase in q wasn't quite as large as expected from temperature,
    but the thought at the time was that this was due to the air over land
    not being saturated as it would be over the ocean. Thinking aloud, we
    could do a simple back of the envelope calculation and see if the slight
    reduction in RH explains the difference between the expected from T
    changes in q and what Kate sees with the real q obs.
      These are probably worth showing, to see if anybody next week has any
    more thoughts. I guess you can't produce similar series for q?
       Kate has received the proofs. I'll attach these. The password you need
    to open this is JCLI2274.   JCLI and then 2274.  Kate is also
    putting together another paper.
       Glad to hear you'll have some time later in the year.
    Cheers
    Phil
   At 11:58 22/09/2008, Adrian Simmons wrote:

     Phil, Kate
     I'm afraid I've been hopelessly busy with my day job and standard GCOS
     matters of late, so the surface humidity stuff was put to one side until
     yesterday evening.
     I'm in Boulder next week, and will be giving a talk on reanalysis at
     NOAA/ESRL on the Friday. I thought I would include the surface humidity
     comparisons between HADCRUH and ERA, and decided last night to set off
     the data retrieval needed to plot relative humidity time series from ERA
     (-40 and -Interim versions).
     I've plotted the results this morning (for all land points, not just for
     the grid boxes where there are HADCRUH data) and attach them for
     selected areas. I hope the plots are self-explanatory. Time series are
     adjusted to give zero mean value for the overlap period between ERA-40
     and ERA-Interim (1989-2001).
     These plots give a clear impression of a downward trend in relative
     humidity (cf AR4, where it is stated "trends are uncertain, but suggest
     that it [relative humidity] has remained the same overall" - perhaps you
     wrote this Phil). This may not be that inconsistent with earlier
     evidence - Kate's thesis summary does indicate a trend in specific
     humidity that is slightly less than that expected for constant relative
     humidity, although you state that trends in RH are statistically
     indistinguishable from zero, and the indication of a downward trend in
     the ERA results is strengthened by the low values analysed for recent years.
     What do you make of this? I've used a quite standard set of programs to
     do the calculations, so it's unlikely (but certainly not impossible)
     that I've made a mistake. I would be wary about the quality of the
     reanalysis for the tropics, but reasonably confident about it for much
     of the extratropics.
     I'll not have much time to work on this for the next five or six weeks,
     but things may get easier in Nov/Dec, and by then ERA-Interim will be
     close to the present day.
     Best regards
     Adrian
     --
     --------------------------------------------------
     Adrian Simmons
     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
   ----------------------------------------------------------------------------
