cc: m.hulme@uea.ac.uk
date: Wed, 16 Sep 1998 15:02:55 +0100
from: Dave Slade <DASL@WPO.NERC.AC.UK>
subject: RE: CONFIDENTIAL REFEREE FEEDBACK FOR STANDARD GRANT GR3/11950
to: nwal@soton.ac.uk

Dear Dr Arnell,

This e-mail contains comments provided by the referees of your standard
grant (grant number in subject line) which you may wish to respond to.
Please note that you are only being sent those parts of the report that
contain queries or criticisms of your proposal and these may not reflect
the overall tone of the report or the funding recommendation. 

Should you wish to respond to the points raised by the referees, please
do so by the end of 29th September 1998 to allow us to include your
response in the papers sent to the committee reviewing your grant. We
realise that this is a short period of time and would encourage you to
submit your response by e-mail or fax where possible. Please make sure
that you include your grant number. Where appropriate (and possible) this
e-mail has been copied to the other PI's on the grant in case the lead PI is
out of contact. However we can only accept one response to these
comments (preferably from the lead PI) - not one from each PI contacted. 

Please note that further referee comments that have arrived late may be
forwarded to you as they come in.

Please acknowledge receipt of this e-mail.

Yours sincerely

Dave Slade
-----------------------------------------------------------------------------------------
REFEREE C

The proposal seeks to simulate time series of monthly streamflow using
regional climatic time series or patterns of rainfall, temperature and other
data that can be given to the hydrological model which in turn are
calculated to be correlated to large scale atmospheric or oceanic indices. 
The latter might be the North Atlantic Oscillation, El Nino or spatial patterns
of sea surface temperature or atmospheric variability. The way this is to
be done is described very little. In fact some fairly concrete examples of
possible approaches to one or two specific problems could have been
given. I have a feeling that the hydrological modelling part of this proposal
has been thought out more than the climatic inputs.

There has been much work relating ENSO to rainfall deficiency or excess
over the globe, and also to temperature anomalies. Presumably this work 
(key papers might have been referenced) can be readily adapted to
calculate large scale streamflow using the CRU climatic data sets and the
large scale model.  The links between NAO and rainfall and temperature in
the winter half year have also been investigated by several authors and
are very strong in some parts of Europe and perhaps North West Africa.
Evidence is also starting to emerge from UKMO research that the winter
NAO might have a predictable component and that decadal modulation of
the interannual variability is important. In addition there is evidence of
strong non-ENSO influences of sea surface temperature patterns (e.g.
in the Atlantic) on some South American and African rainfall, and
presumably streamflow. Note that work is going on in Australia and South
Africa, with  links to the UKMO, on this problem. This background favours
the proposal.
   
The proposal does not note that a complicating factor in ENSO-related
rainfall predictability is the time-varying nature of some of the links. An
example is Australia where this problem has long been known, if so far
largely ignored in, for instance, the Bureau of Meteorology RAINMAN
service to farmers and hydrologists. Recent research seems to implicate
decadal to multidecadal modulation of sea surface temperatures in the
Pacific. One might suspect this problem is important in Southern Africa
where large quasi-bidecadal variations of rainfall are well known. In both
cases active research is now being done to investigate decadal
modulation of interannual rainfall teleconnections. The misbehaviour of
some 1997-8 El Nino teleconnections might be related in part to this.
Decadal/multidecadal modulation of interannual variability is starting to 
emerge more generally as an important paradigm, stimulated by recent
published evidence for mechanisms in the Pacific and the Atlantic. Since
the proposal involves hydrological simulations over the whole twentieth
century, this aspect of the problem should be explicitly considered.

The proposal needs to be revised to demonstrate in more detail that a
strong start can be made in two or three key large regions (including
Europe perhaps), using current or emerging knowledge about climatic
teleconnections. It would help to reformulate the project to  consult with
other selected projects that have related aims, and do a better review of
teleconnection and climate mechanisms literature. 


REFEREE E

The overall goal of the proposal is to test the hypothesis that there is
an influence of large scale phenomena (such as ENSO) in hydrological
records. This is a laudable and meritorious goal and one that should
be addressed. However, the test should be made using real
hydrological records not with synthetic hydrological records created
by introducing a synthetic daily pattern into monthly precipitation
records and then calculating synthetic runoff records. The daily
pattern of precipitation may itself depend on the global phenomenon
and there will always be argument about the realism of the
hydrological model (despite the intention to test it against data). The
net effect of introducing this artificiality will be to denigrate rather
than enhance the credibility of any relationship found. The fact that
the actual hydrological records may not be globally available or may
be unreliable in certain parts of the world (if this is in fact what
motivates suggesting the approach described in the proposal) does not
justify replacing true records with artificial records to carry out the
test. Rather, the testing should be restricted to regions where the data
is robust enough to allow it.
:::::::::::::::::::::::::::::::::::::::::::::::::::
Mr Dave Slade
Awards & Training
NERC
Tel: 01793 411783
Fax: 01793 411655
e-mail:david.a.slade@nerc.ac.uk
