DEVELOPING AN OCEAN PREDICTING SYSTEM WITH HYCOM FOR THE SOUTH ATLANTIC REGION PROJETO: Rede Temática de Modelagem e Observação Oceanográfica

In this study, the Hybrid Coordinate Ocean Model (HYCOM) has been tested for the Atlantic Basin in order to produce forecasts of ocean state variables for the South Atlantic and the South America coastal region. Initially, the model was run for a 30-year period forced with COADS climatological atmospheric fields in order to produce a climatological ocean state. After that, the 6-hour 2007 NCEP reanalysis atmospheric forcing fields were used to produce more realistic oceanic conditions for actual time. The fields used to force HYCOM were the surface air temperature and specific humidity, precipitation rate, net shortwave radiation, net radiation and wind stress. Monthly means were calculated from the 6-hour NCEP reanalysis atmospheric data and the model was forced with this data for one year during 2007 initialized with the last output of the climatological run. After that, the model was reinitialized with the last output from the 2007 run and re-integrated for two more years forced with the 6-hour NCEP reanalysis for 2007 to impose a better adjustment of the model state for this period. Then, the model was run for the period between 1 January and 31 July 2008, using the NCEP 6-hour reanalysis atmospheric data to produce a near real-time ocean state. Finally, the output on 31 July 2008 was used as initial condition to produce an experimental 7-day forecast from 00 UTC 1 August 2008 to 00 UTC 8 August, forced by the NCEP Global Forecasting System (GFS) atmospheric prediction fields. The model results were compared with observed data obtained from the satellite retrievals and from in situ measurements by the Pilot Research Moored Array in the Tropical Atlantic (PIRATA) moorings, and with the Levitus oceanic climatological data. The preliminary results show that the model is able to simulate the mean oceanic features. Also, when the model was forced with the NCEP reanalysis, it was able to produce features associated with the interannual variability, i.e., the model was sensitive to the NCEP dated fields. A comparison of the two 2007 runs - one forced with the atmospheric monthly mean fields and the other forced with the 6-hour fields - indicated that the sea surface temperature (SST) produced by the run with the high-frequency forcing was more realistic than the simulation with the monthly mean forcing.