InfraStructure for the European Network for Earth System modelling - - PowerPoint PPT Presentation

InfraStructure for the European Network for Earth System modelling IS-ENES Sylvie Joussaume, CNRS/IPSL, France

Combine GA, Exeter, 26/05/2011

Scientific Board : S. Joussaume, J.C. André, J. Mitchell, T. Palmer,

• J. Marotzke, R. Budich, A. Navarra, P. Kabat, B. Lawrence

ENES

http://www.enes.org

• A network of European groups in climate models of the Earth system

Launched in 2001 by Guy Brasseur (MOU)

• More than 40 groups from academic, public and industrial world
• Main focus : discuss strategy to accelerate progress in

climate/Earth system modeling and understanding

• Several EU projects

PRISM, ENSEMBLES, METAFOR, COMBINE, IS-ENES, EUCLIPSE EMBRACE (under negotiation)

IS-ENES : Infrastructure for ENES FP7 project - 2009-2012 18 partners 7.6 M€ Infrastructure : Models and their environment Model data Interface with HPC ecosystem Users : The ENES community Regional climate models Impact studies Europe : 6 global climate models CMCC, COSMOS, EC-Earth, Hadley, IPSL, Meteo-France Network activities / Service activities / Joint Research activities

Objectives:

• Integration of the European ESM community

ENES strategy, exchange of expertise/training, portal

• Development of ESMs

service on common tools and model components, common evaluation international standards

• Prepare for high-end simulations

interface with PRACE, improve model performance

• Foster application of ESM simulations for climate change

impacts dissemination of model results, interface with decision makers and users

http://is.enes.org/

http://is.enes.org/

High-performance computing ecosystem PRACE/DEISA Web service interface for the impact community PROTOTYPE Use cases / methodologies Model data distributed network

IPCC AR4 & AR5

Climate Models information & software ESM evaluation

Virt rtual Earth System Resourc rce Center

European Unified access to http://is.enes.org/

A unified access to European ESM information & software Model documentation & A network of experts Evaluation toolkits for models

Support to users OASIS 4 Standard configurations User support

Common software:  OASIS coupler  CDO data processing software  NEMO, Nucleus for European Modeling of the Ocean

Tutorial, FAQ, developments

IPSL CNRM EC-EARTH CMCC

http://is.enes.org

EU FP7 : IS-ENES & METAFOR

Requested : 2.3 Pb Earth System Grid Adapted from Taylor, WGCM, 2010

Access to distributed ESM data network

High-Performance Computing

HPC task force : JC André, R. Budich, MA Foujols, S. Joussaume, A. Navarra, G. Aloisio

• Strategy

Scientific case, requirements, benchmarks

• Improve model performance

Portability, parallel I/O,OASIS4

• Prepare for future architectures

IESP & European Exascale Software Initiative, answers to G8 call

• HPC ecosystem

Interface with PRACE and DEISA2

Eg « virtual community » by DEISA2 Collaboration with PRACE on EC-Earth

Unified HPC environment

ENES collaboration with PRACE IS-ENES / PRACE workshop

Paris 01/12/2010

3 coupled models run at high resolution on o(1000) processors Hadley, IPSL, EC-Earth 1 year project : based upon EC-Earth as a benchmark Share experience ; common issues : coupled model; OASIS 4 / NEMO ; I/O issues Data storage

Background

 unique foresight : EUROCLIVAR Concerted Action (1998)

http://www.knmi.nl/euroclivar/frsum.html

“a better integration of the European modelling effort with respect to human potential, hardware and software”

• PRISM project and ENES

 WCRP : COPES strategic framework 2005-2015 move towards “a common climate and Earth system modelling infrastructure”  World Climate Modelling Summit, Reading May 2008

Infrastructure strategy for ENES For next 10 years

Foresight exercice to prepare the strategy

Resp : John Mitchell & Jochem Marotzke with ENES SB

Foresight meetings Montvillargennes, March 2010 & Hamburg, Feb 2011

 Drivers : Science & Society From understanding to the development of “Climate Services”  Key science questions

What is needed to provide reliable predictions of regional changes in climate? How predictable is climate ? What is the sensitivity of climate (feedbacks, nonlinear behaviours) ? Can we model and understand glacial-interglacial cycles ? Can we attrivute observed signals and understand processes ?

Why ? Improve research efficiency / convince about our needs For who ? For us & for funding agencies (national, FP8, JPI Climate)

Infrastructure strategy for ENES for the next 10 years

ENES Towards a European Climate Infrastructure Initiative : a sustainable virtual laboratory Grand Challenge project: ~ km ESM convective cloud resolving Recommandations: 1) Access to world-class HPC for climate 2) Develop the next generation of climate models

To address: Exascale HPC & improved physics

3) Set up data infrastructure (global and regional models) for large range of users from impact community 4) Improve physical network (eg link national archives) 5) Strengthen european expertise and networking

IS-ENES, Circle2 Eranet and EEA Data needs for the impact community

Copenhaguen, 11-12 january 2011

• Access to both global and regional climate change simulations
• Need for processing tools and processed data
• Provide guidance on uncertainties and how to use climate models

results

• Eventually provide different sources of information in one linked

system (virtual portal)

• Improve access to data and training

From Rob Swart

IS-ENES, Circle2 Eranet and EEA Data needs for the impact community

Copenhaguen, 11-12 january 2011

• Access to both global and regional climate change simulations
• Need for processing tools and processed data
• Provide guidance on uncertainties and how to use climate models

results

• Eventually provide different sources of information in one linked

system (virtual portal)

• Improve access to data and training

A timely foresight !!

 preparation of FP8  possibility for a follow-up of IS-ENES - call 2012  Joint Programming on Climate (Clik’EU)  discussions on Climate Services: role of climate modelling  start of PRACE

Joint Programming JPI Climate (CliK’EU) Climate Knowledge for Europe

Germany, Austria, Finland, Netherlands, France and Italy Belgium, Denmark, Ireland, Norway, Sweden, and United Kingdom

• Joint Programming :
• Strategic and long lasting (> FP)
• Challenges which pan European research will be facing next

10-20 yrs

• Coordination of national research, development of common

strategy, joint programs and research infrastructures

• Linked with EU FP and other initiatives

Connecting climate knowledge and decision making

Moving towards decadal climate predictions Climate Services Research Sustainable Transformation

• f Society

Decision Making Tools

Climate system knowledge 1 2 3 4 Socio-ecological context knowledge

Innovation Multi- sectoral, multi-level Integration

Climate system variability Understanding key processes Decadal and regional prediction Science-Society interfaces Capacity of response Socio-ecological constraints Synergies and trade-offs Model comparison Strategic planning Knowledge integration User needs Communication tools

Stakeholder participation

Integrated Climate Knowledge and Decision Support Services for Societal Innovation

Action perspectives

3

Module 1 : Moving towards reliable decadal climate predictions

• Investigate climate predictability on seasonal to decadal timescales
• Provide reliable climate information for the next few decades and

up to the centennial scale

• Improve the observation, understanding and modelling of key

processes and mechanisms

• Promote and develop a european climate modelling

collaboration environment

• Promote and develop a european collaboration environment for

long-term monitoring and analysis of the earth system

CONCLUSIONS

• ENES : strengthen the European collaboration (eg COMBINE) and

infrastructure (IS-ENES)

• ENES : key role to play in JPI Climate

JPI Climate : key role to reinforce ENES (national & EC)

Next call on « seasonal to decadal prediction »

• ENES strategy : key for FP8, JPI Climate, PRACE

Future models: still to clarify how far we want to go on collaboration Strategy will need regular revisions

• IS-ENES : importance of e-infrastructure for climate modelling

At starting point, Further developments : include RCMs, model developments, service to impact community … FEEDBACKS !