Impact of different initialisation methods on the quality of sea forecasts for the Sicily Channel

An operational ocean forecasting system provided numerical predictions on the main
hydrodynamics in the seas around Sicily. The system is evaluated using an independent dataset
of sea surface temperature (SST) satellite observations. A set of specific numerical experiments
are carried and the simulated SST data was compared with the satellite data observed at
selected dates. Two kinds of initialisation techniques are adopted to generate different initial
conditions: 'slave mode' and 'partially active mode'. The accuracy of both approaches is analysed
to individuate the optimal nesting strategy. The results show that the partially active mode
improves the forecasts, resulting in a model accuracy increase of up to about 30%.