Statistical analysis of model wind data for the Mediterranean Sea

The focus of this study is to analyze the probability distribution functions of model wind data over the Mediterranean Sea. The atmospheric wind data set is composed by ECWMF analyses for the period 2010-2019. A single grid point statistical method is applied to the Mediterranean Sea for both wind components and amplitude. The pdf (probability distribution function) of the wind components is Gaussian while the amplitude is Weibull. In addition, sensitivity experiments are done to compare the Weibull with the Exponential Weibull pdfs, showing almost identical patterns for both distributions. The use of two parameters Weibull distribution is widely accepted to represent the statistical structure of surface wind, while three parameters Exponential Weibull distribution mostly refers to extreme events. The pdf parameter distribution in the Mediterranean Sea is shown for the first time to be associated with specific wind structures such as Mistral and Etesian winds. This study confirms the previous results from Chu (2009) for oceanic currents and by Drobinski (2015) for wind station data, both cases showing the two parameter Gaussian pdf for wind components and Weibull pdf for wind amplitude. The knowledge of these distributions will help to improve the ensemble ocean forecast as for the setting of initial conditions of ocean forecasts where atmospheric forcing is crucial to quantify the forecast errors.