Are you interested in environment issues? Are you worried about
climate change? Would you like to make a contribution to our
developing understanding of future climate prediction? If so, the
Department of Mathematical Sciences at the University of Durham (UK)
has an EPSRC-funded Studentship entitled "Statistical Issues for
Climate Prediction: Public Resource Distributed Computing and the
Climateprediction.net Project" which might be right up your street.
Public Resource Distributed Computing (PRDC) is the exploitation of
`spare' CPU cycles on personal computers using the Internet. It
became well-known through the SETI@home project, and has now been
taken up much more widely; a standard platform for PRDC now exists,
known as the Berkeley Open Infrastructure for Network Computing
(BOINC). The climateprediction.net experiment uses PRDC and the BOINC
to improve our understanding of future climate, and in particular to
quantify our uncertainty about critical factors such as the climate's
sensitivity to increased levels of atmospheric carbon dioxide. The
challenge is to improve the effectiveness of climateprediction.net by
helping to identify carefully-designed experiments making the best
possible use of existing knowledge about climate and climate
simulators. This is a generic issue in PRDC experiments, so that
improvements in technique can be rapidly rolled out to a wide variety
of scientific experiments through the BOINC.
This is primarily a PhD in statistical design for computer
experiments, for which there exists a literature going back over 20
years. The challenge is to turn the broad guidelines in the
literature into specific recommendations for actual experiments. The
reward will be to see those experiments in action and to help analyse
the results, leading to better predictions about future climate. The
Statistics Group at the University of Durham are world-leaders in the
design and analysis of computer experiments, and work closely with
several of the UK's key climate research groups. This is a very good
opportunity for a student who would like to make a contribution to an
important issue, and a very good starting point for a career as a
statistician or an environmental scientist.
More details. The University of Durham and the Maths Department:
"http://www.dur.ac.uk/" and "http://www.maths.dur.ac.uk/". The
climateprediction.net experiment and the BOINC:
"http://www.climateprediction.net/" and "http://boinc.berkeley.edu/".
An introduction to practical climate science:
"http://www.realClimate.org/".
Funding. For a UK National: three years full fees, plus a stipend of
£13,000 pa, and the possibility of an extension of up to one year.
For an EU National: full fees but no stipend.
Requirements. The candidate must have strong mathematical skills
(EPSRC specifies a minimum of 2.i or equivalent in
Mathematics/Statistics or a related subject); statistical and/or
computational skills would be an advantage. Some familiarity with
issues relating to climate change and climate prediction would be
helpful but are not a prerequisite.
Enquiries and deadline. Informal enquiries should be directed to Dr
Jonathan Rougier, Department of Mathematical Sciences, University of
Durham, Science Laboratories, South Road, Durham DH1 3LE, UK,
([log in to unmask]). For applications, a CV and covering
letter should be submitted by Friday 29 April 2005, to the same.
Please include details of courses taken and grades achieved so far.
|