32nd Conference on Broadcast Meteorology/31st Conference on Radar Meteorology/Fifth Conference on Coastal Atmospheric and Oceanic Prediction and Processes

Thursday, 7 August 2003: 2:10 PM
Verification of Surface Current Predictions from the BIO Ocean Forecasting System
C. L. Tang, Bedford Institute of Oceanography, Dartmouth, NS, Canada; and E. Dunlap
Poster PDF (199.9 kB)
The BIO ocean forecasting system - Ice-Ocean Forecasts for the East Coast of Canada (http://www.mar.dfo-mpo.gc.ca/science/ocean/icemodel/ice_ocean_forecast.html) produces 2-day forecasts of surface currents, waves, water level and ice cover for eastern Canadian seaboard. The model used for ice-ocean forecasts is a coupled sea-ice model and Princeton Ocean Model. The model domain includes the Grand Banks, N.E. Newfoundland Shelf, Labrador Shelf and Labrador Sea. The forcings are six hourly surface winds, air temperature, dew point temperature, and cloud cover. On the open boundaries, temperature, salinity, sea surface elevation, volume transport are fixed for each season. Forecast currents are generated on a 20 km by 20 km grid at 16 sigma levels in the vertical.

To verify the surface current predictions, trajectories calculated from the model current fields are compared with trajectories of four surface drifters deployed over the Grand Banks by Canadian Coast Guard in October 2002. Model trajectories are computed from the model at the first sigma level, and 10 m and 30 m averages. The differences among them are not large. The agreement between the model and the drifter trajectories depends on wind conditions and location. The largest discrepancies occur during high winds where predicted directions deviate from drifter directions by up to 80 deg. In the ocean model, wave effects are not considered. The most important contribution of waves to surface current is the Stokes drift. A simple parametrization for wave induced currents is used and the resultant currents are superimposed on the model currents. A significant improvement is achieved. Average separation of the model and drifter trajectories after two days is reduced from 26 km without waves to 20 km with waves. The correlation coefficient between the model distance and drifter distance is increased form 0.58 to 0.79. A general formulation to incorporate the wave effects into the ocean model is proposed.

Supplementary URL: