 |
| Titel |
The UPSCALE project: a large simulation campaign |
| VerfasserIn |
Matthew Mizielinski, Malcolm Roberts, Pier Luigi Vidale, Reinhard Schiemann, Marie-Estelle Demory, Jane Strachan |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250089430
|
| Publikation (Nr.) |
 EGU/EGU2014-3633.pdf |
|
|
|
|
|
| Zusammenfassung |
| The development of a traceable hierarchy of HadGEM3 global climate models, based upon
the Met Office Unified Model, at resolutions from 135 km to 25 km, now allows the impact of
resolution on the mean state, variability and extremes of climate to be studied in a robust
fashion.
In 2011 we successfully obtained a single-year grant of 144 million core hours of
supercomputing time from the PRACE organization to run ensembles of 27 year
atmosphere-only (HadGEM3-A GA3.0) climate simulations at 25km resolution, as used in
present global weather forecasting, on HERMIT at HLRS. Through 2012 the UPSCALE
project (UK on PRACE: weather-resolving Simulations of Climate for globAL
Environmental risk) ran over 650 years of simulation at resolutions of 25 km (N512), 60 km
(N216) and 135 km (N96) to look at the value of high resolution climate models in
the study of both present climate and a potential future climate scenario based on
RCP8.5.
Over 400 TB of data was produced using HERMIT, with additional simulations run on
HECToR (UK supercomputer) and MONSooN (Met Office NERC Supercomputing
Node). The data generated was transferred to the JASMIN super-data cluster, hosted
by STFC CEDA in the UK, where analysis facilities are allowing rapid scientific
exploitation of the data set. Many groups across the UK and Europe are already taking
advantage of these facilities and we welcome approaches from other interested
scientists.
This presentation will briefly cover the following points;
Purpose and requirements of the UPSCALE project and facilities used.
Technical implementation and hurdles (model porting and optimisation,
automation, numerical failures, data transfer).
Ensemble specification.
Current analysis projects and access to the data set.
A full description of UPSCALE and the data set generated has been submitted
to Geoscientific Model development, with overview information available from
http://proj.badc.rl.ac.uk/upscale . |
|
|
|
|
|
|