Скачать презентацию Transitioning space weather models to operations at the

3c22be6b103ce5de93de86122704f5b4.ppt

• Количество слайдов: 22

Transitioning space weather models to operations at the UK Met Office Suzy Bingham, David Jackson, Catherine Burnett and Mark Gibbs Met Office, UK ESWW, Thurs. 20 th Nov. 2014 © Crown copyright Met Office

Table of Contents • Numerical Weather Prediction (NWP) • Met Office Space Weather Operations Centre (MOSWOC) • Research models to operations: process & challenges • Using other NWP methods for space weather • Summary © Crown copyright Met Office The Met Office

Numerical weather prediction • NWP: when current weather observations are assimilated into numerical computer models & processed using mathematical equations to generate a forecast of the future state of the weather. • Met Office has great experience in NWP methods which can be used for space weather forecasting. Transfer NWP methods for terrestrial weather, to space weather © Crown copyright Met Office

NWP building blocks Building blocks of NWP & some experiences to build upon: • OBSERVATIONS: forecasts depend on knowledge of the current state of the environment. (1) Research in instruments & obs. , (2) real-time acquisition of obs. , (3) implementation of research instruments to operational use. • DATA ASSIMILATION: an accurate representation of the initial state of the environment is vital so assimilatest obs. (1) Fast comms required to gather obs. , (2) expense of computational process. • PREDICTION: use of a mathematical model to generate a forecast. (1) Use of approximations to evaluate equations, (2) model validation. © Crown copyright Met Office

NWP building blocks • UNCERTAINTY/ENSEMBLES: creation of multiple versions of a forecast from different initial states to give info on uncertainty of forecast. (1) Small errors can quickly grow to large, (2) resulting variation gives indication of confidence in forecast. • RISK ASSESSMENT & COMMUNICATION: understand risks of impacts of the forecast & communicate to users. (1) If a large impact is forecast, a group of expert forecasters/scientist/users can be involved in decision making, (2) use of different communication methods (emails, emergency response meetings). • VERIFICATION: to continually improve operational systems. (1) Use metrics to show well the model predicts/ value added by forecaster, (2) work within WMO guidelines. © Crown copyright Met Office

Met Office Space Weather Operations Centre (MOSWOC) • Delivers space weather alerts, warnings & forecasts 24/7. Public forecasts & sector specific forecasts are available: http: //www. metoffice. gov. uk/publicsector/emergencies/space-weather • Sits within terrestrial weather ops centre. • Provides a platform for running space weather models operationally. © Crown copyright Met Office MOSWOC

Operational system at MOSWOC • ‘Operational’ means an automated end-to-end system with: • a real-time observations stream, • model predictions, • product dissemination & • quality control monitoring. • Models in an operational system are: (1) rigorously tested, (2) imbedded into a robust infrastructure & (3) supported 24/7. • Reasons to run models operationally: • to provide unfailing & timely model output to forecasters, • to enable a research model to be run with 24/7 support (reliably), • to run model after a project has been completed, • resilience/backup to another centre, • one place of focus for information. © Crown copyright Met Office

Research models to operations: the models Operational: • WSA-Enlil, solar wind speed • REFM, Relativistic Electron Forecasting Model, daily electron fluence forecast at GEO orbit. • D-RAP, D Region Absorption Predictions Model, global map of prediction of radio propagation conditions. REFM (SWPC) © Crown copyright Met Office Enlil (D. Odstrcil) D-RAP (SWPC)

Research models to operations: the models Semi- operational: • Multi Instrument Data Analysis System, Bath University, Total Electron Content. • SPACECAST, British Antarctic Survey/EU project, 0 -3 hr forecast of >2 Me. V electron flux. MIDAS © Crown copyright Met Office SPACECAST

Research models to operations: the process 1. Run research model on desktop PC. 2. Run revised model on development environment. 3. Run model on operational suite and view output on forecaster webpages. © Crown copyright Met Office • Review licence and terms of use of model and input data. • Understand model processes, input data and output data. • Rewrite in Met Office standard language if necessary. • Adapt code to Linux platform if written in Windows. • Adapt code to run using Met Office database input if necessary. • Change model area covered e. g. to Europe, depending on customer requirements. • Change cadence to meet customer requirements, e. g. REFM changed to 3 hours when originally ran daily. • FTP input files through firewall. • Write ROSE script to auto-run model on operational suite. • Test model doesn’t affect other processes on IBM. • Model resolution can depend upon speed/expense on IBM. • Some commands may need to be changed to run on IBM. • Adapt visualisation code to Met Office style. • Verify output. • Complete Met Office documentation to allow model to run on operational suite.

Operational process Process to follow to run a model or service operationally in the Met Office. © Crown copyright Met Office

Research models to operations: output Once a model is running operationally, output is viewed on internal Met Office webpages forecaster use. Forecasts are then provided on the external webpages for public/sector specific use. MO internal webpages: WSA-Enlil & REFM © Crown copyright Met Office External webpage: http: //www. metoffice. gov. uk/publi csector/emergencies/spaceweather

Further work following NWP methods • Note: need to understand space weather needs & not just use off the shelf NWP solution. • Ongoing: implementing models operationally. • WSA-Enlil ensemble forecasting – research stage. • Verification – initial stages. Evaluation of models & forecasts to check they meet & continue to meet requirements. © Crown copyright Met Office

Proposal for “Carrington” – a UK Space Weather Mission • A Sun-Earth Sentinel at L 5. • First Operational Space Weather mission – addresses MOSWOC requirements. • High technology readiness, low risk, low cost. • High UK heritage: instruments developed by UK scientists. • Fast transfer to L 5 for a 10 -year mission. • 24/7 operations, 100% coverage, continuous data. • Excellent research output. • UKMO, UK Cabinet office, protects UK infrastructure hence growth. • Very large potential for UK industry and UK science to lead the field. • Excellent opportunity for UK/US bilateral. © Crown copyright Met Office

Summary • Met Office has built upon its experience using NWP methods, for transitioning space weather models to operations. • Challenges which arise when implementing models operationally include: rewriting code & ftping input data through firewalls. • Continue to transition research models to operations. • Continue to use NWP methods for space weather forecasting. © Crown copyright Met Office