date: Tue Feb  1 10:45:41 2005
from: Keith Briffa <k.briffa@uea.ac.uk>
subject: Re: Dai's PDSI dataset
to: "Femmie en Gerard van der Schrier" <fem-gerard@zonnet.nl>

   OK- just talked to Phil and he says the calibration period is probably 1951-80 - hence the
   changing range (and means at zero) but it is still apparent that there is great asymmetry
   in the extreme values, much wetter than drier, and perhaps this is a regional bias (poor
   data in some ares for calibration?) .
   These extreme values may all lie in ares of poor data coverage .The comparisons with our
   European (and the US) grid will probably be much better behaved.
   Apparently Dai (and Trenberth) are making much of the declining trend in recent years, but
   this Figure shows that this is the result of less wet rather than more dry - probably an
   important point in terms of impact.
   Keith
   Gerard
   I agree that this looks very much like a post-analysis truncation . What geographic area is
   represented by these numbers? What seems even more strange though, is the inhomogeneity at
   around 1956 , and perhaps between 1956 an 1980. This surely is not the result of a natural
   circulation related phenomenon. Have you tried comparing any of his previuos extreme maps
   against latest ones (and yours in overlap region)? You could plot PDF for particular dry
   and wet years to see if there is a sharp truncation at pdsi values of +or _ 15.
   Give me a ring (and I wil call you back)
   Keith
   At 07:21 01/02/2005, you wrote:

     Keith,

     It seems that Dai et al. have managed to get rid of the 'spikes' in their latest PDSI
     dataset (the one published in J. Hydrometeorology, 2004). At a cost though. Attached is
     a figure of the mean, the minimum value and the maximum value of the PDSI for each of
     their maps (horizontal in the figure is time in months, starting in January 1870).

     It seems to me that the max. and min. values are artificially bounded between +15 and
     -15. What do you think?

     My guess is that they, either during the PDSI calculations or as a post-processing
     routine, reset the largest PDSI values to be within +/-15.

     Cheers, Gerard

   --
   Professor Keith Briffa,
   Climatic Research Unit
   University of East Anglia
   Norwich, NR4 7TJ, U.K.

   Phone: +44-1603-593909
   Fax: +44-1603-507784
   [1]http://www.cru.uea.ac.uk/cru/people/briffa/

