Comparison of analyzed and measured wind speeds in the perspective of oceanic simulations over the Mediterranean basin: Analyses, QuikSCAT and buoy data

Surface vector wind datasets from different assimilation systems and from scatterometers have been recently made available over the entire Mediterranean basin and for a large spectrum of spatial and temporal resolution. In this work, we compare wind vector analyses, derived from different routine assimilation systems and from blended products, to wind vectors obtained from QuikSCAT satellite sensor and to those directly measured by buoy-mounted anemometers. The analysis has been performed to verify the accuracy of the analyzed data, when the specific objective is the generation of surface winds field to force Mediterranean Sea simulations. The inter-comparison covers the period 2000–2005. Our analysis demonstrated that the spatial resolution of the data sets represents one of the main relevant sources of error in the analyzed wind fields, explaining the worst results of the reanalysis data and the relative accuracy of the ECMWF. This work also confirms the usefulness of blending QuikSCAT and reanalysis products, which could be used to force oceanic simulations. The blended data cover the period from July 1999 to present when QuikSCAT wind data are available. Before this period, blended products are not produced and different solutions to correct wind speed from routine assimilation systems have to be investigated. A simple empirical method to adjust the ERA40 wind speed product is then proposed. The analysis of the difference between the annual Mediterranean heat budget computed using the adjusted and the original ERA40 winds suggests that the impact of the correction is not negligible. Considering the year 2000, the annual average heat budget for the whole basin is modified from ∼ 34 W/m2 to ∼ − 6 W/m2.