GODAE Final Symposium, 12 – 15 November 2008, Nice, France Marine Pollution Monitoring and Prediction Bruce Hackett, met.no Eric Comerma, Applied Science Assoc. Pierre Daniel, Météo-France Hitoshi Ichikawa, Japan Met. Agency
Jan 15, 2016
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
Marine Pollution Monitoring and Prediction
Bruce Hackett, met.no
Eric Comerma, Applied Science Assoc.
Pierre Daniel, Météo-France
Hitoshi Ichikawa, Japan Met. Agency
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
Outline
• What is the marine pollution problem?– Focus → marine oil pollution
• What use are GODAE products?
• Some examples from oil spill forecasting
• Future perspectives
Aim: Demonstrate the use and benefits of using GODAE ocean data
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
What is the marine pollution problem?
• A wide range of pollutants in the marine environment
• Main impacts are: - harm to life, property and commerce
- environmental degradation
Chocolate mousse à la Prestige – an emulsion
• Marine oil pollution is a good paradigm for pollution monitoring and prediction:– Indisputably human-caused– Catastrophic events (spectacular)– … and long-term dosages (insidious)– Can be ugly, deadly, illegal and costly– Often transnational– Well-established preparedness; national
and international services
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
What’s the role of GODAE? (1)
• We need monitoring and prediction for: 1. emergency response to spill
events2. impact assessment, both for
specific events and for scenarios
• Both need a good description of the metocean conditions– Weather, sea-state, ocean
circulation– Observations, hindcast, nowcast
and forecastHeavy seas during Statfjord A spill, Dec
2007
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
What’s the role of GODAE? (2)
GODAE is important since:
• GODAE systems are providers of observations and ocean circulation forecasts.
• GODAE can also provide archival data for past events and statistical studies.
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
What’s the role of GODAE? (3)The simple view of emergency forecast systems
Drift model
Initial and forcing data
Output data
Userinterface
GODAE!
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
Emergency response servicesCritical constraints
• Emergency oil spill response services depend on quick and reliable access to drift prognoses – response time <30 min– 24/7/365 availability
• Critical component for drift forecasting is real-time access to accurate prognostic forcing data:– NWP models: wind, air temp – Wave models: Hs, Stokes drift – Ocean models: currents, temp and salinity
• Access & Accuracy: main challenges to operational ocean forecasting!
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE dataWhat use are they to oil spill forecasting?
• “Problem”– GODAE provides data at global to basin scales– Oil spill events are typically near-coastal– Oil spill services are national, with local data sources
• Answer: Use GODAE data to help improve the prediction service by – providing boundary conditions for nesting local models– providing for direct forcing (alternatives, backup, ensembles)– extending the spatial and temporal reach of services
Examples follow
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
• ASA-SA as a commercial middle-user: – providing MetOcean data downstream services– facilitator, aggregating added value into the data
stream, providing dedicated forecasting services– to understand end-client needs, to deal with
constraints and data availability limitations– bridging the gap between global data providers
(GODAE) and local, very specific user needs (harbour, offshore platforms, etc.)
GODAE & oil spill forecasting – South AmericaASA-SA
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – South America ASA-SA
ASA-SA ocean and oil spill forecasting system for
Petrobras, based on nesting in global ocean data products
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – South America ASA-SA Aggregation to cover data requirements
Global ProductGlobal Product Coastal Coastal ProductProduct
Aggregated Aggregated ProductProduct
Regional Regional ProductProduct
Surface RadarSurface RadarSLDMB’sSLDMB’s
Global NCOM and ADCIRC via EDS Aggregation Service
Aggregation of short and long range CODAR
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – South America ASA-SA buoy validation
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – South America ASA-SA buoy validation
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA forecasting service
Information
for Japan Coast Guard - distribution of spilled oil
Information
for Japan Coast Guard - distribution of spilled oil
Surface winddirection / speed
Wave conditionsignificant wave heightpredominant wave period predominant wave direction
Land / Sea distribution
Current & SST
Ocean data assimilation model
(COMPASS-K MOVE)
Transport by wind Stokes driftAdvection by current Diffusion EvaporationEmulsification
Initial step
Results
Data input
Time integral
Numerical weathermodel
Global Spectral Model(GSM)
Numerical wavemodel
Global Wave Model (GWM)Coastal Wave Model (CWM)
Oil Spill Prediction Model
Initial conditiontime / location of a incidentSpilled oil amount / type
Initial conditiontime / location of a incidentSpilled oil amount / type
Calculation of positions of the pollution source
Since Sep. 2008
Grid resolution : 2 - 30 km Time step : 1 minute Max. calculation time : 192 hour
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA forecasting service
Information
for Japan Coast Guard - distribution of spilled oil
Information
for Japan Coast Guard - distribution of spilled oil
Surface winddirection / speed
Wave conditionsignificant wave heightpredominant wave period predominant wave direction
Land / Sea distribution
Current & SST
Ocean data assimilation model
(COMPASS-K MOVE)
Transport by wind Stokes driftAdvection by current Diffusion EvaporationEmulsification
Initial step
Results
Data input
Time integral
Numerical weathermodel
Global Spectral Model(GSM)
Numerical wavemodel
Global Wave Model (GWM)Coastal Wave Model (CWM)
Oil Spill Prediction Model
Initial conditiontime / location of a incidentSpilled oil amount / type
Initial conditiontime / location of a incidentSpilled oil amount / type
Calculation of positions of the pollution source
Since Sep. 2008
Grid resolution : 2 - 30 km Time step : 1 minute Max. calculation time : 192 hour
MOVE ocean model:– A GODAE
contribution– Development aided
by GODAE collaboration
– Direct forcing and nesting
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA ocean forecast model (GODAE)
COMPASS-K MOVE
Kuroshio route
Tokai regionTocharian strait
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA oil drift validation
Standard : drifting buoy for oil spill tracking radius : 20 cm thickness : 5 cm weight : 4 kg material : expanded polyester
1999 Jan.
1999 Nov.
2001 Feb.
2001 Nov.
2002 Feb.
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA oil drift validation
Case of East China Sea in November, 2001
:
: simulation with COMPASS-K
: simulation with MOVE-WNP
: oil pursuit buoy
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
MOTHY
Observations
Cedre, MRCC,
PREMAR
Atmospheric models(ALADIN, ARPEGE, IFS)
Tide
Ocean models(MERCATOR, MFS)
ForecastsImages, GIS, Google Earth
Internet, email, fax
Bathymetry
GODAE and oil spill forecasting – EuropeMétéo-France forecasting service
Schematic of Météo-France general framework with access to local and GODAE prognostic ocean data sets.
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – EuropeMeteo-France – Prestige simulations
Prestige case snapshot 2002-12-13 12 UTC. Prestige trajectory = black line MOTHY (wind only) = blue MOTHY+Mercator = greenMOTHY+FOAM = red.
Observed slicks = black triangles Positive impact of adding GODAE current data (Mercator and FOAM) in waters where the large-scale circulation has a significant impact.
However, there are large differences between the current data sets (Mercator and FOAM).
First use of GODAE / Mersea data was most encouraging
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – EuropeMeteo-France, met.no, UCY – Mersea simulations
6 PTR buoys Western Mediterranean sea
3 Argosphere buoys Eastern Mediterranean sea
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – EuropeMeteo-France, met.no – Mersea simulations
Findings:- Often large differences in the predicted current fields- Drifter trajectories are better reproduced in stable and well marked circulation (left)- Large discrepancies between the predicted currents in the open ocean reflects the fact
that the current field is dominated by unstable mesoscale dynamics (right)- The most accurate results are obtained when applying currents from a local, fine-scale
ocean model nested in basin-scale data.
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – EuropeMeteo-France, met.no – Mersea simulations
MFS 1/16th Mercator MED 1/15th
Mercator Global 1/4th All fields are daily means
for 2007-09-24 00 UTC
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – Future perspectivesData access and accuracy
• Access to ocean data– Immensely improved during GODAE lifetime!– Global GODAE models → global coverage for any
oil spill model.– Demonstrated in projects (e.g. Mersea) and by
individual middle users (e.g. ASA) …– … still need to improve standards for access and
products– Any data policy hindrances remaining?
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
• Accuracy of ocean data– Still the main source of oil spill forecast error.– GODAE assimilation systems continue to improve forecast
accuracy in the blue ocean (e.g., Kuroshio)… – … but still gross errors in some eddy-rich areas.– Need more detailed currents in coastal and shelf seas →
• Higher resolution models
• Nesting is preferable to direct use
• Tides must be included at some stage
• Improved forcing from atmosphere (high-resolution, high freq data)
• Assimilation of current data (HF radar, drifters)
– Need measures of accuracy to propagate into oil spill forecast results!
GODAE & oil spill forecasting – Future perspectivesData access and accuracy
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – Future perspectivesAddressing accuracy
MF: Merc15th
MF: WindOnly
met.no: Nordic4
met.no: Bio4
met.no: Merc15th
met.no: Merc4th
met.no: TOPAZ
MF: Merc15th
MF: WindOnly
met.no: Nordic4
met.no: Bio4
met.no: Merc15th
met.no: Merc4th
met.no: TOPAZ
Using ensembles to address accuracy / confidence issues:• Single-model – perturbations on one of the models (oil drift and/or
forcing) costly!• Multi-model – aggregating several independent oil spill forecasts• Multi-forcing – same oil spill model, different forcing data (enabled by
GODAE!)• A combination of the above• Requires good analysis and/or interpretation methods
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
End
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
Comparison between MOVE-WNP and COMPASS-K
IAU (Incremental Analysis Update; Bloom et al. 1996)
Nudging methodscheme of data assimilation
altimeter data of satellitesgrid point data of sea altitudetarget datafor assimilation
3-dimentional variation method(in which the EOF analyses of the vertical profiles of temperature and salinity are employed ; Fujii and Kamachi 2003)
4-dimentional opiticalinterporlation
scheme ofdata analysis
Assimi-lation
system
etc.
introducednot usedmixture layer process
introducednot usedsea ice model
free surfacerigid ridsurface of sea
54 layers (0-5500m)21 layers (0-4500m)vertical resolution
0.1 x 0.1 degree0.25 x 0.25 degreehorizontal resolution
Oceanmodel
MOVE-WNPCOMPASS-Kelementmodel/system
IAU (Incremental Analysis Update; Bloom et al. 1996)
Nudging methodscheme of data assimilation
altimeter data of satellitesgrid point data of sea altitudetarget datafor assimilation
3-dimentional variation method(in which the EOF analyses of the vertical profiles of temperature and salinity are employed ; Fujii and Kamachi 2003)
4-dimentional opiticalinterporlation
scheme ofdata analysis
Assimi-lation
system
etc.
introducednot usedmixture layer process
introducednot usedsea ice model
free surfacerigid ridsurface of sea
54 layers (0-5500m)21 layers (0-4500m)vertical resolution
0.1 x 0.1 degree0.25 x 0.25 degreehorizontal resolution
Oceanmodel
MOVE-WNPCOMPASS-Kelementmodel/system
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE & oil spill forecasting – South America ASA Environmental Data System
Environmental
Data Store
Professional Client Thin/Web Client
Internet
Web Services
Web Service XML Request
NetCDF Data
Catalog ServerModelling Engine
• Spatial / Temporal Aggregation
• Modelling Services (Oil, SAR)
• Resources Monitoring
OGC WMS/WxS
EDS
Winds
Currents
Drifters
Weather
Remote Sources
(public/private, e.g. NOAA, Global/Regional OOS)
GODAE Final Symposium, 12 – 15 November 2008, Nice, France
GODAE and oil spill forecasting – West PacificJMA model drift validation
February 18, 2002February 19, 2002February 20, 2002February 21, 2002February 22, 2002February 23, 2002February 24, 2002February 25, 2002
:
: track of oil pursuit buoy
: simulation of MOVE-WNP
Case of southeast sea of Hokkaido in February 2002