A 44-year (1958-2001) high-resolution atmospheric hindcast for the whole Mediterranean Basin was performed within the EU-funded Hindcast of Dynamic Processes of the Ocean and Coastal Areas of Europe (HIPOCAS) Project. The long-term hindcasted data set, which comprises several atmospheric parameters at different levels, was produced by means of dynamical downscaling from the NCEP/NCAR global reanalysis using the atmospheric limited area model REMO. The REMO hindcast has been exhaustively validated. On that score, various hindcasted surface parameters, such as 10-m wind field, 2-m temperature and mean sea level pressure, have been compared to satellite data (ERS-1/2 scatterometer) and in-situ measurements from offshore stations. In addition, two ocean models (waves and sea level) have been forced with REMO hindcasted fields (mean sea level pressure and 10-m wind field). The validation of these ocean runs, performed through comparisons of simulated waves and sea level with oceanographic measurements, allows to evaluate "indirectly" the quality of the REMO hindcasted data used as atmospheric forcing. Once the quality of the hindcasted data was verified, the efficiency of the regional enhancement performed through dynamical downscaling on the NCEP global reanalysis was assessed. The regional improvement was evaluated through comparisons of REMO and NCEP performance in reproducing observations. The important improvement obtained in the characterization of extreme wind events is particularly remarkable.