date: Fri, 7 Dec 2001 10:04:28 -0000
from: "Declan Conway" <d.conway@uea.ac.uk>
subject: RE: 6-12-01-2001_Hulme
to: "Centre for Water Level Fluctuations" <caspianlevel@kaznet.kz>, "Hulme Mike" <m.hulme@uea.ac.uk>

   Hi Terry

   I guess that for their report Mike and Tim use their pattern scaling method which involves
   'taking the difference between a future period (in thier case centred on hte 2080s) and a
   control period and then standardising the spatial pattern by dividing by hte global-mean
   warming for the particular model experiment and for the appropriate time slice (2080s) the
   resulting standardised climate change pattern is thus expressed per degree C of global
   warming. These change fields can then be re-scaled for a range of global mean temperature
   increases to represent the uncertainty in hte emissions and the climate sensitivity.' (this
   is why the time horizon for predicted changes may therefore be altered - because we don't
   actually know when certain CO2 concentrations will be reached and how sensitive the climate
   system is to a given change in conc.).

   This approach is followed for both T and P change.

   I think this also answers your second question - for the report the changes will have been
   calculated from the difference between the control and the future run from the respective
   GCMs - NOT using the baseline.

   Okay I hope this is useful - I have received leaflets from Geof Jenkins on hte RCM - I'll
   forward these to you - best wishes

   Declan.



   -----Original Message-----
   From: Centre for Water Level Fluctuations [mailto:caspianlevel@kaznet.kz]
   Sent: 06 December 2001 04:01
   To: Hulme Mike
   Cc: Declan Conway
   Subject: 6-12-01-2001_Hulme




   Dear Mike,


   It's a while since we were in contact and this is just a very small query following up your
   Caspian Sea Report of many months ago. It is less of a query but more a question of
   educating my ignorance.

            When you adjust the model to a standard sensitivity value I assume that the
   outputs of, for example precipitation and temperature, also change (close to a pro-rata
   amount?). However, would not this  automatically change the time horzon of the output. I
   find the concept a little perplexing in that if one changes the sensitivity of the model
   then one implicitly changes the time of the prediction for that model. Obviously I don't
   under the process involved.

            We are finalising the inputs and outputs for the CEH hydrological model and are
   having to iron out some anomalies. A comparison of Declan Conway's work on GCM inputs to
   the model have been compared with your estimates for the Volga catchment and also similar
   work on a hydrological model by the State Hydrological Institute in St Petersburgh. CEH
   have suggested that maybe you have calculated your outputs from the GCM control runs rather
   than from recorded data such as your CRU Baseline data.and this may account for some
   differences.


   Sorry to trouble you, the Stage 2 of our project ends today so if a very short and very
   quick response were possible it would be appreciated.


   Many thanks and Merry Christmas,



   Terry
