Wave modeling performance in the Gulf of Mexico and Western Caribbean: Wind reanalyses assessment

This paper evaluates the wave modeling performance in the Gulf of Mexico and Western Caribbean Sea employing three different wind reanalysis data. Wind reanalysis is employed as the main forcing in wave generation/propagation numerical models. While the National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) and the European Centre for Medium-Range Weather Forecasts (ECMWFs) ERA-40 reanalyses have been previously assessed in the performance for wave modeling, ECMWF's ERA-interim and in particular NCEP's North American Regional Reanalysis (NARR) are more recent reanalyses. They both provide better resolution and description of the wind fields and have not been evaluated for long-term wave modeling. Therefore, the aim of this work is the assessment of the three different wind reanalyses on the wave hindcast performance. Attention is drawn on the wind reanalysis capability for predicting both mean and extreme wave conditions during two different periods: (i) an anomalous year where cyclonic events dominate the extreme wave climate in the region (2005); and (ii) a year with the wave climate dominated by synoptic events (2006). A third generation wave model, forced by the different wind reanalysis data, is calibrated with National Data Buoy Center (NDBC) buoys observations. Wind reanalysis data allow a consistent implementation of third generation wave models in order to predict the mean wave climate (correlation coefficient similar to 0.84 for NCEP/NCAR, 0.94 for ERA-interim, 0.92 for NARR) for applied ocean studies. Numerical results revealed that both ERA-interim and NARR improve the wave modeling performance with respect to NCEP/NCAR (for extreme and non-extreme conditions), whereas the high- (spatial and temporal) resolution NARR data are more suitable for modeling extreme cyclonic events (i.e., hurricanes) in this region. (C) 2012 Elsevier Ltd. All rights reserved.