RIASSUNTO
An ice-ocean forecasting system for eastern Canadian waters has been developed to provide short-term forecasts of ice concentration, ice thickness, sea surface temperature and other oceanographic variables. In this paper, we describe the forecast model, forcing data, methods of data assimilation and operational procedure. The method by which to compute the forecast skills is outlined. As an example, April 2008 ice forecasts are evaluated by comparing the mean square errors of the forecast and the persistence.
INTRODUCTION
Knowledge of the present and future sea-ice conditions off Canada's east coast is important for the operations of the offshore oil industry, marine shipping, fishing and weather forecasting. Timely and reliable ice forecasting requires accurate ice models. Ice models have been developed since the mid-1970s (Semtner, 1976). For the east coast of Canada, coupled ice-ocean models (Tang and Gui, 1996; Yao et al., 2000) were developed and used to study the seasonal variation of ice cover and ice motion. These models were later implemented in an ice-ocean forecasting system at the Bedford Institute of Oceanography (BIO). Improvements of the models and the forecasting system led to the development of Canadian East Coast Ocean Model (CECOM) and a new forecasting system at BIO. The new system has been run operationally since 2007. Daily forecasts are provided to users in the offshore oil industry, sea-ice research institutes, and government agencies responsible for search-and-rescue operation and marine shipping. Forecast results in graphic forms are posted on the Internet accessible by the general public at the following: http://www.mar.dfompo.gc.ca/science/ocean/icemodel/ice_ocean_forecast.html. In this paper, we describe the forecast model and the BIO forecasting system, the methods of data assimilation for sea surface temperature and ice concentration, and the forecast schedule. The forecast skills are evaluated using ice data provided by the Canadian Ice Service (CIS).