On the use of the NCEP/NCAR surface winds for modeling barotropic circulation in the Río de la Plata Estuary

Abstract

Atmospheric products such as the National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis seem to constitute a solution for the lack of direct observations in the Río de la Plata estuary. Forcing numerical ocean models with those data allows the simulation of the estuarine system for long periods including the effect of high frequency wind variability, which in turn constitutes a powerful tool for understanding and evaluating wind forced climate variability in the region. Nevertheless, given the serious limitations of those data sets in the Southern Hemisphere, a question that must be answered before this kind of simulation explains to what extent ocean models forced by a reanalysis represent the real system. In this paper a model validation for the Rio de la Plata inner and outer estuary including adjacent shelf regions is presented. The HAMburg Shelf Ocean Model (HamSOM) code was forced with NCEP/NCAR reanalysis data to perform four hindcasts covering about 1 month in different seasons and five extreme events of storm surges. Three nested unilaterally coupled model domains with increasingly finer grids were used for the simulations. The role of wind forcing over the outer, large-scale shelf region was studied by both including and excluding the forcing for the outer domain. It was found that weak winds in the NCEP/NCAR reanalysis are under-estimated, and a correction factor is needed to adjust winds for different magnitudes. The comparison of simulated surface elevations and currents with observed ones yields high correlations, which improve even further when a meteorological forcing is included in the outer domain. The validation comprises state of the art accuracy with regards to problems where stratification plays a minor role. Numerical solutions satisfactorily reproduce both the timing and amplitude of the observed estuarine waters’ response to wind, implying that the quality of the forcing is in turn good. Consequently, our simulations constitute an indirect validation of the NCEP/NCAR reanalysis 10 m winds. Even though they tend to underestimate wind speed, direction and variability seem to be properly represented, at least in the atmospheric synoptic to intra-seasonal time scales.