Objective
Objectives :
To provide new and improved tools for the dynamical
simulation and prediction of climate on a human/
biospheric scale. To further understand error sources in
regional climate models (RCMs) and to validate these
models using new, high resolution, observations.
Brief Description of the Research Project :
Analysis of current simulations indicates some model
deficiencies are due to errors in representing physical
processes and a lack of spatial resolution. For the
former, physical packages imported from GCMs have behaved
differently when used at the higher RCM resolutions.
These deficiencies will be overcome by a concerted effort
to understand and reduce the errors, by modifying the
physical parametrizations, and to develop and test new
high resolution RCMs. For boundary conditions the
simulations will use the 15 year ECMWF reanalyses (ERA)
(from 1979 to 1993, coinciding with the Atmospheric Model
Intercomparison Project (AMIP) II period).
An important area where RCMs behave differently than GCMs
and provide much greater realism is in the simulation of
daily precipitation, especially heavy events. Detailed
validation of daily precipitation distribution
simulations will be performed on a series of models at
different resolutions. Other elements of the hydrological
cycle will also be validated, both to establish their
realism and as diagnostics to help understand errors.
An understanding and reduction of errors in GCM
simulations over Europe is vital for using nested RCMs
for climate (change) simulations. The project will
further understanding of the main sources and effects of
these errors and liaise with GCM modellers regarding
their reduction. It will assess AMIP II GCM simulations
and RCM simulations nested therein. Also, it will
analyse, over Europe, the results of climate GCMs
assimilating ERA data for information on the forcing
which causes these errors. This technique will also be
used in a high resolution RCM.
Decadal or longer timescale fluctuations in climate
simulations will not be captured using the proposed RCMs
under current computing limitations. To help overcome
this problem, existing RCM integrations will be used to
train a statistical dynamical downscaling technique to
provide links between the large scale flow of the driving
data and the corresponding RCM simulations. In addition
to being able to extend multi annual RCM simulations to
multi decadal time scales, this method will also be used
to extend short test integrations to longer periods.
The project will result in three European RCMs and a
strectched-grid GCM of about 50km resolution with reduced
errors. Four new high-resolution (about 20km) RCMs will
also be produced, for Europe, north-west Europe, the
greater Alpine area and south-east Europe (including the
Alps). Validation of these models will be extended to
include comparisons with daily precipitation and
hydrological cycle observations. Detailed analysis of GCM
large-scale errors over Europe and resultant errors
induced in nested RCMs will be produced and disseminated
to the relevant GCM modellers. A new technique for
generating RCM output from large scale driving data
statistically will be developed using existing RCM
simulations and validated with independent data.
Call for proposal
Data not availableFunding Scheme
CSC - Cost-sharing contractsCoordinator
EX1 3PB EXETER
United Kingdom