The Environmental statistics section of the Royal Statistical Society
is holding a two day meeting on the 11/12th January at the National
Oceanography Centre, Southampton (http://www.noc.soton.ac.uk) on the
subject of 'Statistics and Computer Models'. Both days will start at
11 am and finish at 4 pm.
Day one (11 January) will consist of an introduction and a 'hands-on'
tutorial with two software packages in this area. These are the R
package BACCO and the GEM-SA package. The tutors will be Robin Hankin
(NOCS) and John-Paul Gosling (Sheffield).
Day two (12 January) will consist of a number of research talks in
this area. These are
11:00 Jim Gattiker (NOCS) Calibrating Climate Models: a case study
and comparison of DACE approaches.
An ensemble of intermediate complexity climate simulation runs,
parameterized by three system parameters, is used as a case study for
demonstrating calibration to observed data of atmosphere and ocean
temperature. Three independent analyses of the data have now been
performed, showing broadly consistent results. The approaches and
implementations of these analyses differ in some key ways, and some
aspects of the alternatives will be discussed in the context of this
problem of multivariate response modeling.
11:45 Nadja Leith (UCL) Quantifying uncertainty in sequences of
climate model output
13:30 Neil Crout (Nottingham) Title TBD
14:15 Jonathan Rougier (Bristol) Emulator diagnostics and emulator
construction
All emulators make drastic structural simplifications for
tractability. We need diagnostics to evaluate the propriety of these
simplifications, and also, in practice, to sell our emulator to the
modeller. Predictive diagnostics, such as leave-one-out and its
generalisations, are powerful but expensive. But the need for good
diagnostics is so compelling that 'lightweight' emulators, which are
very cheap to construct, and for which the diagnostic information is
abundant, are justifiable, and not only as a first step.
15:15 Tony O'Hagan (Sheffield) The Estimation of the Net CO2 Flux for
England & Wales and its Uncertainty using Emulation
If you are planing to attend day 1 could you please let us know as we
would like to have enough computers for participants. Also please
tell us if you are planning to bring a laptop.
Any queries please contact me ([log in to unmask]) or Robin
Hankin ([log in to unmask])
Peter Challenor
Room 254/39, James Rennell Division,
National Oceanography Centre, Southampton, SO14 3ZH, UK
Phone (+44) 23 80596413 Fax (+44) 23 80596400
|