Presentation on theme: "The Role of Computing in Climate Science Dr. Robert Bishop WMO Information System (WIS) Workshop on Information Access Enablers Geneva, Switzerland, 17-18."— Presentation transcript:
The Role of Computing in Climate Science Dr. Robert Bishop WMO Information System (WIS) Workshop on Information Access Enablers Geneva, Switzerland, 17-18 May 20010
Yokohama Earth Simulator Opened March 2002, NEC SX-6
Best Dedicated Climate Machines (TAKEN FROM THE NOVEMBER 2009 TOP500 SUPERCOMPUTER SITES) Worldwide Ranking Organisation Country Peak Teraflops Sustained Teraflops Supplier # 31 JAMSTEC JAPAN 131.07 122.40 NEC SX-9 # 33 ECMWF UK 156.42 115.90 IBM Power 575 # 34 ECMWF UK 156.42 115.90 IBM Power 575 # 35 DKRZ GY 151.60 115.90 IBM Power 575 # 54 NAVO USA 117.13 90.84 CRAY XT5
ICES: a peak performance facility Hardware Software Data Science People
ICES Top Priorities Maintain dedicated HPC in the top 5 of machines worldwide Supply HPC cycles and software engineering support to national and regional Earth & Climate centres worldwide Enable climate science to reach comparable levels of spatial and temporal resolution as NWP Evolve open source Earth system models by integrating elements from climate, bio, geo, space & social sciences Drive breakout hierarchy of nextgen Earth-Climate models Support training of next generation holistic thinkers Provide info-briefings on Earth & Climate to International Organisations & NGOs
ICES and Disaster Risk Management Community Resilience Adaptation & Mitigation Planning & Relief Strategies Precursor Signals ENVIRONMENT BIOSPHERE EARTH SYSTEM CLIMATE & WEATHER SOCIAL SYSTEMS SOLAR SYSTEM
ICES and Geoengineering ENVIRONMENT BIOSPHERE EARTH SYSTEM CLIMATE & WEATHER SOCIAL SYSTEMS SOLAR SYSTEM Climate Remediation CO 2 Removal Solar Radiation Management Unexpected Consequences
Proposed ICES Computing Resources Dedicated High Performance Computing - 20 year transition: petaflop(10 15 )-exaflop(10 18 )-zettaflop(10 21 flops) High-resolution 3D interactive immersion & image analysis - auditorium level viewing with remote viewing & remote steering Low cost power availability (nuclear, hydro, solar) Ultra-high-speed networking from ITU Green Computing, Cloud Computing Citizen Science Computing Google Earth, WolframAlpha
HPC Computing Architectures (We need to compute ~ 1000 x real-time) Homogeneous vs heterogeneous Multi-core, CPU-GPU, FPGA, ASICs or full custom Programming languages, software tools & middleware Cluster vs SMP, distributed vs shared memory Power management, flops/watt Silicon-Photonics. Quantum?
Earth Modelling Software Grid Size Parameterisation Algorithm development Coupling, linkages & feedbacks Representation of physical processes Integration of the socio-economic processes Initial & boundary condition determination Uncertainty estimates & management Statistical & ensemble methods Hierarchy of models Multi-models Stochastics Nextgen
Earth Data Challenges Data assimilation Historical data re-analysis Data access, archiving & meta-data Data quality control & harmonisation Data availability (in situ, remote sensing) Sparse data (Oceans, Africa, Antarctica) Model output-data validation & verification Model output-data storage (or re-compute?)
Observation & Data Sources Airborne & Satellite Remote Sensing: Envisat, MeteoSat, SMOS, GOCE, GOES-R, LandSat, SBIRS In Situ : AWS, Radar, Lidar, Broadband Seismic Mobile: Aircraft, Ships, Argo Buoys, Autos? Cell Phones? Socio-economic: GDP, Land Use, Food, Water Resources, Energy, etc.
Core Actors Network World Meteorological Organisation (WMO) - World Climate Research Programme (WCRP) - World Weather Research Programme (WWRP) Group on Earth Observations (GEO) European Centre Medium-Range Weather Forecasts (ECMWF) National Meteorology Bureaus National Geological Surveys National Climate Centres Research Universities
Extended Actors Network ESA, NASA, JAXA, EUMETSAT, CGMS IOC, 21 st C-OI, SIO, WHC, JAMSTEC BGS, BRGM, USGS, ERI GEOSS, GCOS, GMES ENES, ACRE, ESFRI CSIRO, CALIT2, NCAR, COLA CNRS (CC-IN2P3), PRACE UK Agencies: DEFRA, DECC US Agencies: NOAA, NSF, DOE, DOD
Why Public-Private Partnership? Fast Agile Simple Flexible Responsive Non-political Independent New sources of funding
Why Geneva? International city, neutral country, trusted Science literate, educational infrastructure Proximity to global policy bodies: UNEP, WBCSD, IUCN, WWF WHO, UNHCR, ICRC WMO (WCRP, WWRP), GEO WTO, WEF, UNCTAD, ILO, ITU, EBU Partnerships: CERN, ETH, Canton Universities
ICES Funding PHASE 1 (2010~2015) $350M from sources: 1/2 public, 1/2 private - overflow capacity for national and regional centres - development of nextgen integrated climate/Earth models PHASE 2 (2016~2020) $450M from sources: 1/3 public, 1/3 private, and 1/3 products and services, such as: - test bed for large scale construction projects - disaster risk management - industry specific services - policy-making support - decision support - what if scenarios - geoengineering
ICES Foundation Members Board members: Bob Bishop President, André Kaplun Secretary, Julien Pitton Treasurer Bankers: UBS Auditors: PricewaterhouseCoopers Expert Committee: Dr. Ghassem Asrar Director, World Climate Research Programme, WMO Prof. Martin Beniston Chair for Climate Research, University of Geneva Director, Institute for Environmental Sciences Prof. Marc Parlange Dean of the School of Architecture, Civil & Environmental Eng. Ecole Polytechnique Federal Lausanne Dr. Michael Rast Head of Programme Planning Office Directorate of Earth Observation Programmes European Space Agency Ethics Committee: tba
Helping guide the successful transformation of human society in an era of rapid climate change and frequent natural disasters.
Recent Major Natural Disasters Date Country Lives Lost Cause April-May 2010 Iceland - volcanic ash cloud April 2010 China 2000+ 6.9 quake April 2010 Brazil 200+ rain, mudslides March 2010 Uganda 350+ rain, mudslides Feb 2010 France 50+ tempest, sea walls Feb 2010 Chile 700+ 8.8 quake, tsunami Jan 2010 Haiti 250,000+ 7.0 quake April 2009 Italy 300+ 6.3 quake Feb 2009 Australia 250+ bushfires May 2008 China 70,000+ 8.0 quake Aug 2005 USA 1,800+ hurricane, levees Dec 2004 Indonesia 225,000+ 9.3 quake, tsunami Aug 2003 Europe 30,000+ heat wave
Before the Age of Computing In 1922, Lewis Fry Richardson, a British mathematician and meteorologist, proposed an immersive giant globe to numerically forecast weather. This factory would employ 64,000 human computers to sit in tiers around the interior circumference of a giant globe.
ICES and Society Research Integration Development Discovery & Innovation Communications Teaching Training ENVIRONMENT BIOSPHERE EARTH SYSTEM CLIMATE & WEATHER SOCIAL SYSTEMS SOLAR SYSTEM
ICES in a Nutshell Development of a transformative meta-science that integrates climate, weather, environmental, geo, bio, & socio-economic sciences Next-generation modelling and simulation techniques Support for national & regional climate centres Teaching, training, capacity building Decision support, communications Dedicated supercomputing Global networking Visual intensity ~200 professionals including seconded experts
Weather & Climate Communities (A Convergence of Methodologies) Numerical Weather Prediction (NWP): National Bureaus of Meteorology today: 3~5 days ECMWF today: 5~10 days Future Goal: increased accuracy, and on to monthly & seasonal level Climate modelling: WCRP today: 100~1000 years Future Goal : increased accuracy, and on to decadal & annual level The 10-year Challenge: Seasonal to inter-annual predictions Global to regional to local forecasting Coarse-grain to fine-grain spatial resolution Extreme weather early warning