Glob Model The Glob Model study initial findings

Glob. Model The Glob. Model study, initial findings and objectives of the day Zofia Stott 13 September 2007 1

Objective of presentation/contents Background to the Glob. Model study Preliminary conclusions of the study Objectives of the day 2

Background to the Glob. Model study EO data-model fusion is a relatively new area for ESA DUE Glob-projects Summer schools Ad hoc collaborations, eg with ECMWF Fact finding Programmes, initiatives, organisations, people • European focus • Also international programmes, eg IGBP, WCRP • Analogies with US where appropriate Analysis Report Opinion seeking What are the issues for the European community? Strategy and implementation plan for ESA Where should ESA be involved? How should ESA be involved? 3 Workshop

Background to the Glob. Model study Scope Numerical Weather Prediction Re-analysis New (pre)-operational services, eg GMES Fast Track services • Ocean forecasting • Chemical weather forecasting Global change and Earth system science EO data-model fusion Data assimilation Ancillary surface data fields Model validation 4

Background to the Glob. Model study Glob. Model hypothesis Understanding, forecasting and predicting the behaviour of the Earth system depends on • Data and models working together • Satellite data are key Progress is accelerated by collaboration between the science base and operational services Objective is to create a “virtuous circle” • High scientific return • Linked to new operational services • Leading to investment in both new research and operational missions 5

Background to the Glob. Model study Specific requirements/issues • The role of OSSE and OSE in quantifying the impact of particular data streams • Concerns about data continuity over the next 10 years • Areas where new or improved instruments are required • Novel data products specifically tailored for model assimilation (eg radiances V retrievals V gridded fields) • Improved techniques for EO data-model fusion (eg development of new data assimilation techniques, observation operators) • Intercomparison and cross validation of different data sets • Improved model development environments which include consideration of EO data issues • Standardisation and harmonisation of EO data formats, data discovery and data access • Improved quality control • Software tools to support the use of EO data streams • Real time delivery and long term curation • Provision of high level products, eg model independent reanalyses • Shared high performance computing environment • Training. 6

Preliminary recommendations – OSE, OSSE “The Global Observing System”, Jean-Noël Thépaut, Data Assimilation Training Course, ECMWF Reading, 25 April- 4 May 2007 7

Preliminary recommendations – access to operational systems Make operational systems more readily available for research Mutual benefit Scientists work on topics of interest to operational agencies Benefit from operational facilities (models, computer resources, expert help) Operational agencies benefit from latest research results Increases chances to technology transfer from research base to operations 8

Preliminary recommendations – integrated data systems Increase emphasis on integrated data systems for new services Optimise in situ and satellite components • Eg What is the balance between Argo floats and altimeters? GODAE/GHRSST/Medspiration projects optimising sea surface temperature retrievals could be taken as an example of good practice 9

Preliminary recommendations Develop observation operators Fundamental link between data and models Essential to ensure early take up of data into operational systems Commit to long term continuity of re-analysis Develop the use of EO data in the land cryosphere components of the Earth system models Develop “climate” quality data sets 10

Preliminary recommendations - people Ensure that the right mix of people/institutions are brought together Experts on satellite data processing, retrievals Experts on operational data assimilation systems Experts on Earth system modelling in the research community Members of satellite instrument and/or science teams Participants in the cal-val effort Members of the satellite data management teams. 11

Preliminary conclusions – provide a science focus Address the big science issues Develop regional climate models able to identify “tipping points” in the climate system Understand link between physical and biological feedbacks in carbon cycle Understand links between climate change and atmospheric composition Develop coupled sea-ice and ocean circulation models Develop improved ability to model hydrological cycle and predict high impact weather Develop ecosystem and biodiversity models 12

Objectives of the day - Splinter sessions Where are we today? What are the key issues? What is your vision for Earth system modelling in 10 years time? What will we be able to do which we cannot do today? Eg • Forecast on an annual/decadal and regional basis? • Forecast high impact weather? • Identify and monitor all climate tipping points? What role should EO play in achieving our goals? What programmes and projects would you recommend to ESA to fulfil your objectives? 13

Backup slides 14

NWP I Developments driven by operational requirements of forecasting centres New services Seasonal and inter annual forecasts High impact weather New and improved services, based on Better models Better data Satellite data are key Innovation needs close links between R&D and operations 15

NWP II Pull through of satellite data for NWP, in Europe Strong for meteorological data sources Eg via EUMETSAT SAFs Weaker for non EUMETSAT data • Ad hoc • But good examples of transfer from research to operational status eg scatterometer, GOME, altimetry Key satellite requirements • • • Low level (1 B/C) radiances Some retrievals (eg Atmospheric Motion Vectors) Surface gridded fields Real time delivery (<1 hour) BUFR, GRIB High priority issues • • 16 Improved coupled models Use of satellite radiances over land, cloud Hydrological cycle Improved surface representation/assimilation

NWP III Increasing experience of OSE, OSSE Quantify impact of satellite data on NWP Comparison of Europe with USA JCSDA • NASA/NOAA initiative • To accelerate take up of new data sources 17

NWP IV “The Global Observing System”, Jean-Noël Thépaut, Data Assimilation Training Course, ECMWF Reading, 25 April- 4 May 2007 18

NWP V “The Global Observing System”, Jean-Noël Thépaut, Data Assimilation Training Course, ECMWF Reading, 25 April- 4 May 2007 19

NWP VI Messages from NWP key for operational data assimilation • 40 years of infrastructure and capability Need to work effectively with NWP centres • EUMETSAT, ECMWF, national met offices No equivalent of GMAO or JCSDA in Europe • No systematic mechanisms for accelerating transfer of research data sources to operations • ADM, SMOS already identified by ECMWF 20

Reanalysis I Long term (eg 40 years) global data sets of past climate using data assimilation Reliant of latest NWP model + historical data ECMWF leads in Europe Key for Understanding climate trends Improving both models and data (biases) Challenges Need for improved coupled models Inhomogeneities in data records 21

Reanalysis II Messages from reanalysis Long term missions needed • Repeats • Overlaps Long term curation of data – a major challenge European reanalysis projects are • “Add on” to existing activities, not core business • Funding ad hoc No sustained European effort in reanalysis 22

“New” (pre)-operational forecasting I Ocean forecasting Chemical weather forecasting Learning from current practice in NWP Reliant on NWP either through loosely or tightly coupled models GMES Core Services providing a European delivery structure Far less technically mature than NWP Requirements less precise Techniques more experimental 23

“New” (pre)-operational forecasting II Data types Ocean forecasts • Broad correspondence between GMES Sentinel 3 and ocean forecasts (altimetry, SST, ocean colour) • Also ocean salinity (SMOS), sea ice thickness (Cryosat), gravity/geoid (GRACE/GOCE), wind/waves (scatterometer) Chemical weather forecasting • Broad correspondence between GMES Sentinels 4/5 and chemical weather forecasting • Also METOP, MSG, ENVISAT, AURA instruments PLUS NWP outputs (forcing fields) 24

“New” (pre)-operational forecasting III Messages Continued development through close research/operational interactions Models immature in key areas of user interests, eg • boundary layer chemical forecasts • coupled physical-biogeochemical models and assimilation of ocean colour data Need for better comparison between data and models Standards, data formats are still evolving etc • GMES and INSPIRE are addressing this Tools, training, common research hub to exchange data and models Important to work with emerging structures Eg EUROGOOS for ocean forecasting 25

Earth system science I Developing GCMs What’s new • Shorter timescales (from centuries to decades), more local impacts (from global to regional) Representation of energy and hydrological cycle Ocean variability and climate change signals Developing land surface models in GCMs Developing models of coupled atmosphere/ ocean/cryosphere 26

Earth system science II Global carbon cycle Quantifying surface fluxes Quantifying role played by fire Identifying weights of key processes in tropics for post-Kyoto negotiations Atmospheric composition Understanding interactions between climate change and atmospheric composition Cryosphere Strongest signals of climate change, but key processes poorly represented in models Predictability of high impact weather Monitoring, understanding, predicting behaviour of ecosystems Impacts of natural resource depletion 27