Supplementary memorandum submitted by
Professor Donald Knight (FL 85a)
Letter from Professor Donald W Knight
to Reg Purnell, Chief Engineer, Flood Management Division, Department
for Environment, Food and Rural Affairs, 27 July 2007
INDEPENDENT REVIEW
OF DEFRA/EA
RESEARCH
It was good to talk last week. I wanted to await
the discussions on Monday last, before writing to you concerning
the above, and also specifically concerning the Theme Advisory
Group (TAG) on Fluvial, Estuarine and Coastal Processes (FCEP),
whose demise the Report recommends. I have enjoyed serving on
that Group for the last six years and also engaging in offering
advice to you earlier, notably through the "Learning to live
with rivers" ICE independent Commission & Report. I want
to make just five points.
1. I welcome your proposals concerning a
high-level board and key managers. The statement in 1.5 (page
3) that "the overall direction of the programme is weak"
is a reasonable assessment in my experience, and your proposals
concerning a high-level board should correct this. This appeared
to be supported by most people on Monday, although there remains
some fuzziness in the chain of command/consultation and management.
However, it necessitates sufficiently technically competent personnel
to be found to fill these positions. As you are aware, I am still
somewhat concerned about the capability of some of the senior
management within the Environment Agency in this respect, particularly
when dealing with technical fluid flow issues. You have to look
no further than the Jubilee channel as a monument to the "skills
shortage" within the EA. A new £90 million channel takes
only 2/3 of its design flow—how basic can you get? Especially
when the EA has a new "Conveyance Estimation System",
developed through a £0.5 million R&D managed programme,
and then "launched" in June 2004. But it gets even worse,
as, in another criticism of the EA, I have to say that the CES
software is still not available to anybody within the Agency or
by consultants who want it, due to EA "procedures"!
They had two years advance warning of this and it is now three
years since that project started. I could say more but I will
no doubt bore you. Mervyn drove this very well, but there appear
to very few champions like him within the EA. Thus staffing and
skills issues within the EA are vital.
2. With respect to the "lack of focus",
you may need to ask why the R&D programme (and TAGS) lacked
"focus" in the past. Part of the reason is I believe
because the previous R&D report (1999) was itself such a step-change
that it probably got the division of subject material wrong, and
hence the focus has been blurred. The joint Defra/EA approach
was very sound and commendable, but it had six TAGs and three
of them—Broad Scale Modelling (BSM), Engineering (ENG) and
Fluvial, Estuarine and Coastal Processes (FCEP) should in my view
have been incorporated into a single group as the basic core "engineering"
or "technical" side of river engineering. Albeit a large
group, it is important to keep these topics together. Experience
of working in these three TAGs has demonstrated this. Part of
the problem therefore has been the lack of appreciation concerning
how "fluid flow" aspects (or call this old-fashioned
hydraulics and hydrology "engineering" aspects if you
like) drive the whole. My third point follows on from this.
3. I fear that the present report (June
version) makes the same mistake again, particularly if you analyse
the breakdown of topics in the Project Area Groups in Table 3.2.
My concern is the way in which the topics are dealt with. This
Table really makes little sense, and I was a surprised to see
the muddle and confusion in their scientific thinking. For example,
"core" modelling topics are seemingly distributed in
a haphazard manner. Why are all those topics on "data",
"uncertainty" and "models" placed under the
"Risk and social dimensions" Theme? What has "benchmarking
of 2-D models" got to do with this—surely this should
be in an "Engineering" Theme? And why just 2-D models,
what about the various types of 2-D and numerous 3-D models? I
have just completed, together with another university, some research
for the EPSRC on this very topic. My worry list could continue
. . . Why are "data" and "sensitivity" placed
in the "Flood incident" Theme? Why is "improved
hydrometric practices" place solely in the same flood incident
theme? Hydrometry is the often neglected topic, but germane to
all. If ever there was a need to "measure more", "think
more" and "model less", it is now. Clearly modelling
is going to play a vital role in nearly all flood risk management,
but there is a real need to understand the basic processes. Most
river processes are very complex, and increasingly models are
being used to deal with these problems. Hence there is crying
need to understand the underlying basic processes better and hence
to use the "tool" better. The user experience is vital.
Running "models" is not just a question of button pushing.
Two decades ago, the non-thinking automation of structural analysis
led to a marked decline in the understanding of structural behaviour,
and as a result the IStrucE had to re-educate their "tool"
users. The same is happening in our profession, as young modellers
get ever divorced from the reality of water in motion.
And so to be positive, I think it is important
that you do not attach "data" to "risk" as
in your modified draft project project area grouping, as proposed
on Monday. There is certainly a role for "risk" as a
separate theme, maybe as a cross-cutting theme. I believe this
was confirmed by the reaction of people on Monday in the discussions
about "subject division". So please keep "data"
with "processes" in an "engineering" theme.
I was glad to see that the defra response was indeed to set up
a more comprehensive "engineering" theme. I would add
"data" to it. I concur with your view to increase the
themes to five, on the proviso that due recognition (including
resources) be given the larger engineering theme. I know that
group is potentially unwieldy, with coasts, estuaries, rivers,
data, processes and modelling as its core, but they fit together.
4. While I welcome any attempt to link some
R&D work more closely with the Research Councils (recommendation
2.7, page 46), I totally disagree with the conclusions reached
in section 1.5, No 3 (page 4), Section 2.3.2 (page 12), Section
2.6.1 (weaknesses, No 2, page 16), Themes (page 27) and conclusions
2.3.2 (page 46). It is wishful thinking that the EPSRC or NERC
will undertake all such basic process work. From my long experience
on EPSRC Panels and also having been an SERC co-ordinator, I know
that much river work is unattractive to reviewers and there are
other fluid flow processes that are often perceived to be more
"attractive" and hence brought to the fore. In the last
year I have seen at least three flood related bids regarding as
"outstanding" (the highest grade within EPSRC) be turned
down for various reasons, not always the most rational. It is
therefore essential that defra and EA maintain their own core
process element where it is required. Also, having been party
to setting up the FRMRC, I have to say that it is still ignoring
many of the basic research topics outlined in our Flood Network
report of 2001.
5. I will not make much comment about the
demise of the FCEP TAG, except to say that the reasons given in
3.2.2 (page 11) are unjustified and flawed. The logic on page
12 smacks of intellectual laziness ("difficult to evaluate",
page 11) and ignorance "no clear customers", page 12).
I reject both of these statements. With regard to the first what
about improvements to FEH, High Flow data, Extension of rating
curves and the CES work? With respect to the second the 34 page
"Vision concerning the hydraulics knowledge base required
in river engineering and associated research needs" sets
out very clearly the issues and elaborates at least 10 "User
needs". Did the authors of the report look at this? It is
on your website. Further, as argued above, processes form a vital
part of engineering, and many topics will never be researched
by the Research Councils in the way required for Flood Risk Management.
Finally, as you know, I will retire in a few
years time, but I remain passionately concerned and committed
to maintaining the science of fluid mechanics, as applied to rivers.
We have an excellent MSc course at Birmingham and many PhD students
(I have seven at the moment) all working in the area of interest
to you. I am therefore ready to assist in whatever way I can,
and trust that these comments will not just be taken as criticisms,
but positive pointers. There is much that is good about UK Flood
Risk Expertise, Hydrodynamic modelling, etc, and your R&D
programme has many excellent aspects. I wish you well with your
task.
Donald W Knight
Professor of Water Engineering
July 2007
|
