Meteorologisk Institutt met.no Operational ocean forecasting in the Arctic (met.no) Øyvind Saetra Norwegian Meteorological Institute Presented at the ArcticGOOS.

Meteorologisk Institutt met.no

Operational ocean forecasting in the Arctic (met.no)

Øyvind SaetraNorwegian Meteorological Institute

Presented at the ArcticGOOS meeting, 12-13 September 2006

Bergen, Norway

Meteorologisk Institutt met.no

Ocean and sea-ice forecasts

• Large scale coupled ice-ocean model with 20 km horizontal resolution

• Nudging of OSISAF satellite observations of SST and ice concentration

• 168 hour forecasts with atmospheric forcing from ECMWF

Meteorologisk Institutt met.no

OSISAF 10km SST

Meteorologisk Institutt met.no

OSISAF 10m ice concentration

Meteorologisk Institutt met.no

Analysed SST and ice-concentration field

Meteorologisk Institutt met.no

High resolution ocean forecast for the Barents Sea

• Nudging of OSISAF SST data• Inflow form 20 km model for the Arctic• Atmospheric forcing from the 10 km HIRLAM model• Not coupled to ice model• 60 hour forecasts produced daily

Meteorologisk Institutt met.no

Meteorologisk Institutt met.no

Meteorologisk Institutt met.no

High resolution coupled ice-ocean model for the waters surrounding Svalbard

• Analysis based on OSISAF 10-kilometre product• In addition, the analysis has been improved by

subjective use of available satellite data: AVHRR, Modis and SAR

• This analysis is then nudged into the coupled ice-ocean model on 4 km resolution

• Boundary values provided by the 20km resolution moel

• A 48 hour forecast is performed daily

Meteorologisk Institutt met.no

Meteorologisk Institutt met.no

Maritime safety: emergency support forecasting

subsurface oil

stranded oil×

surface oil

“Prestige” hindcast simulationNow developing global capability (in Mersea) and moving source facility (coupling to ship drift model). Lessons learned from «Prestige» simulation under EU R&D project (Mersea Strand 1).

met.no supplies the Coastal Authority with 24/7 standby forecasts for oil spill fate anywhere in Norwegian waters, using best available atmosphere, ocean and wave prognoses.Similar service for Joint Rescue Centers for search-and-rescue and ship drift forecasts.

Draugen

oil slick

Froan

Realtime Draugen drift forecast

Example shows drifting object forecast results in JRC’s search and rescue tool.

Example shows real time forecast for wellhead oil leak at Draugen field May 2003.

Meteorologisk Institutt met.no

Environmental monitoring

met.no has developed a web-tool for monitoring the Norwegian Coastal Current environment. http://moncoze.met.no/ collects and presents all relevant information (observations, forecasts, background).Coop NERSC and IMR.

Daily current forecasts supply information on transport of contaminants along the the Norw. coast, e.g., from tanker traffic in the Barents Sea.

Trajectory forecast for coast Finnmark

Meteorologisk Institutt met.no

Wave forecasting: WAM50 & WAM8

WAM50 is run 4 times daily on a 0.45° grid giving forecasts to +60 hrs.WAM50 is run daily on a 0.45° grid giving forecasts to +240 hrs.WAM8 is run twice daily on an 8 km grid giving forecasts to +60 hrs.

Meteorologisk Institutt met.no

Storm surge forecasting:20km domain (topography)

The storm surge model is run daily on a 20 km grid giving forecasts to +60 hrs.

Meteorologisk Institutt met.no

Ensemble storm surge forecasting

Fields from ensemble runs are used to calculate exceedance probability maps.

’Southend-on-sea

Ensemble of 20 water level prognoses forced by down-scaled ECMWF targeted ensemble.

As a test, daily ensemble forecasts to +10 days are generated on a 30 km grid of the Arctic and Atlantic using ECMWF ensemble forcing directly.

thick black = control memberthick red = operational prognosisbox and whisker plots = ensemble members

prob(eta>0.5m) for +60h

Meteorologisk Institutt met.no

Future plans - challenges

• Despite the analysis, the skill of the ice forecasts is still relatively poor

• Model dynamics needs to be improved• Inconsistencies between the SST analysis and the ice

analysis• Ice forecasts on very high resolution ( down to 500 m)

for area like Hinlopenstredet• Present formulations for the dynamics are not valid at

this scale (e.g. rheology)

Meteorologisk Institutt met.no

Meteorologisk Institutt met.no

End