from: Tim Osborn <t.osbornatXYZxyz.ac.uk>
subject: Fwd: Salvage plan for severe droughts project
to: "Phil Jones" <p.jonesatXYZxyz.ac.uk>
from the message below and one from Gerry which you were also cc'd on, and also from
speaking to Rob on the phone, I have to say that they're not too chuffed about my
revelation of errors in the scenarios. They (Rob) did, however, very much appreciate our
honesty and openness in announcing the problems. At first I thought that we could just
tell them about ARPEGE errors, but in fact HIRHAM errors are just as large. The problems
are indeed large and couldn't be glossed over if we ever hoped to publish them.
Rob has proposed the plan below. The crux of it is items 3 and 4 below, in which Steven
Wade could re-run the old (wrong) flows with a simpler yield model (item 4) and if it gave
similar results to the water companies' own results, then he could also use his simpler
model with the new (correct) flows (item 3) and thus obtain the necessary correct results.
Apparently he can do this quicker and cheaper than the water companies can. Obviously (4)
should be done before (3), because if he can't replicate their results with the old flows,
there's little point him using the new flows.
The problem for us is that Steven would require more money. He is talking about �3k
(including overheads I think). I've looked at the PMA accounts for this project and there
is apparently >�5k (*plus* overheads) of staff costs that is currently unspent. I'm not
sure if you have plans/commitments for this. I think it was originally partly for Adam (I
never understood why Janice only gave him a 5 month contract rather than a 6 month one -
that extra month would have been very useful with hindsight!) and partly for us.
If possible can we talk on Monday AM before the Christmas lunch about this plan and whether
we are happy to go along with it, including switching �3k of money from UEA to
HR-Wallingford? I know you'll have many things to catch up on, but Steven would need to
get going before Christmas if he were to have some results in time for the project meeting
in mid Jan.
Date: Fri, 16 Dec 2005 16:29:20 +0000
From: "Rob Wilby" <rob.wilbyatXYZxyzironment-agency.gov.uk>
To: <gSpraggsatXYZxyzlianwater.co.uk>, <s.wadeatXYZxyzallingford.co.uk>,
<p.jonesatXYZxyz.ac.uk>, <t.osbornatXYZxyz.ac.uk>, <Janet.MakinatXYZxyzlc.co.uk>
Subject: Salvage plan for severe droughts project
** High Priority **
Having discussed the matter with each of you I'm proposing the
following six-point salvage plan:
1. Rebrand the old runs as "sensitivity" tests of the OSAY and MOSPA
models. Tim will perform some additional analyses and provide additonal
commentary on how representative our envelope of three runs are of the
broader family of GCM/RCM combinations.
2. Evaluate differences between the old runs and new runs for the 2050s
and 2080s. Tim will also undertake this quick assessment and circulate
plots as per 2020s.
3. CRU and HR will negotiate reprofiling/contracting of resources to
enable Steven to perform runs using latest scenarios and a generic yield
assessment tool. This could potentially bring added-value to the project
by enabling a consistent modelling approach to be performed for both
4. Steven will apply the same generic tool to the old runs to assess
its ability to replicate MOSPA and OSAY runs.
5. The new runs will be presented by Steven to the PSG on 12 Jan 2006,
by powerpoint presentation.
6. Steven to receive general comments on the draft interim report as
circulated, and to weave in the old (sensivity) and new (generic)
results in the final report.
None of the above would, of course, preclude Gerry from re-running OSAY
with the new inflow sequences once he can summon the
time/energy/enthusiasm. Likewise for Janet. However, one potentially
useful outcome of this fix maybe that we find the Steven's 'magic'
spreadsheet model closely replicates the behaviour of MOSPA and/or OSAY.
In which case we will all have learnt something!
Unless I hear any strong objections, I hope that everyone is content
with this pragmatic line. Thanks for your constructive input along the