Nowadays, wind energy is becoming a major contribution to balance the load with power production. However, wind production is highly fluctuating and its output can not be guaranteed at any particular time. This weather dependence of wind power makes difficult the planning between production and load. Moreover, it can also involve, in case of high wind penetration, problematic situations such as unpredicted stopping/start of high powered classical thermal or nuclear units. In this article, based on a self developed algorithm, the combined impact of wind geographical dispersion and prediction errors on the operation of classical production is studied. By the way, some interesting observations are drawn and show that the wind geographical dispersion does improve and facilitate the planning of the classical production system but that it has, however, to be nuanced in terms of produced wind energy. Finally, in order to define a new way to estimate wind production for the dispatch of classical units, a global mean wind distribution is also established. This last one is obtained via the calculation of hourly mean wind speeds based on the predictions made for each considered park. This article proposes such a study based on several important patents made in this wind area.