Forecast skill of the INCOIS operational ocean general circulation models – case study for 2020

The Indian National Centre for Ocean Information Services (INCOIS) provides daily 7-day ocean forecasts for the Indian Ocean as a part of Regional Integrated Multi-Hazard Early Warning System (RIMES) initiative using ROMS and HYCOM models. This study validates first-day forecasts of sea surface temperature (SST) and currents for 2020 against moored buoy data, satellite observations, and analysis products, applying Allan-Murphy’s forecast verification framework. Subsurface currents were assessed using ADCP observations. Both models captured SST variability with high spatial correlation; ROMS showed slightly better SST skill. HYCOM performed better for basin-scale surface and subsurface current forecasts, realistically reproducing mesoscale eddies and the East India Coastal Current, while buoy comparisons indicated ROMS outperformed HYCOM for local surface currents. Despite configuration differences, both systems demonstrated statistically comparable performance on first day SST and current, with discrepancies linked to unresolved processes, model physics, atmospheric forcing, and observational limitations. Skill score analysis revealed substantial value addition of SST forecasts relative to climatology, with root mean square deviation (RMSD) lower than standard deviation (STD). Current forecasts showed lower skill, with RMSD exceeding STD. Overall, the study highlights the strengths and limitations of INCOIS operational ocean forecasting systems and emphasises the importance of continuous verification to enhance prediction services across the Indian Ocean.