Variability of the sea surface salinity (SSS) in the western tropical Pacific: On the ability of an OGCM to simulate the SSS, and on the sampling of an operating merchant ship SSS network

The sea surface salinity (SSS) derived from a network developed at Institut de Recherche pour le Développement (IRD/Nouméa has been analyzed during the period 1995-1998 in the tropical western Pacific. The measurements were made with thermosalinographs installed on merchant ships selected for their regularity and routes. The western tropical Pacific was sampled mainly along three regular routes across the equator leading to an average of a one month frequency. We analyze here how such a network can be efficient in monitoring the SSS at time scales longer than one month. For this purpose we have used results derived from the Princeton Ocean Model (POM) which is forced by the surface flux of National Centers for Environmental Prediction (NCEP) reanalysis data. The interannual variability of the simulated SSS exhibits very similar features to (sub-sampled) observations despite its being weakly damped to a climatology in order to avoid biases. Even smaller time scale phenomena can be simulated, like the erosion/reconstruction of the region composed of low density waters lying within the Pacific warm pool. The agreement between the observational data and the simulations indicate that the network sampling is sufficient to monitor the SSS variability of the western tropical Pacific from three-month to interannual time scales.