MODELLING AND FORECASTING FACILITY

Numerical models are essential tools both for research and for forecasting oceanic weather and climate, particularly for their capacity to describe the ocean’s evolution over time from the surface to the sea bed.The modelling uses the computers capacity of calculation to resolve the mathematic equations which describe the physical phenomena occurring in the ocean.

The models integrate the combined actions of gravity, the Earth's rotation, the atmospheric effects as well as the input from rivers to represent the state of marine currents, temperature, salinity and sea level.

WHY ARE NUMERICAL MODELS USED FOR FORECASTING?

Our seas and oceans cover over 70% of the planet’s surface but we still don´t understand many of the physical processes which occur in them.  Thanks to the different observation systems it is possible to know the ocean state and the variability.  Since these observations don´t have complete spatial and temporal coverage, it is necessary to formulate numerical models to understand and predict the ocean state. 

These numerical forecasting models combine observational data and the basic principles of physical oceanography. The oceanic models are used to simulate the ocean’s state and its evolution and to make predictions of temperature, salinity, sea level, currents and waves.

HOW OCEANIC MODELS ARE MADE: FROM OBSERVATION TO FORECAST

1. Observing and collecting data about the sea state.  The data (temperature, salinity, currents, wave height and direction, etc.) comes from the different observation systems such as satellites, coastal stations, buoys (fixed and drifting), ocean research vessels and underwater gliders.

2. Processing the data obtained.  The increase in computer power and capacity has enabled more realistic models to be built.  The process consists in combining different types of data to obtain the most precise image possible of the sea.

3. Designing the simulation.  Similar to meteorological forecasting models, oceanic models use specific mathematical formulae to resolve equations used to generate predictions.  To make a model, scientists use a three dimensional network where the equations are applied and the results evaluated.

4. Verifying the model.  The model is evaluated with the available observations to verify its reliability and to check the quality of the model. This process permits improved forecasting.

5. Forecasting.  The model obtained is combined with real-time data observations to generate a forecast.