Marine wind is a major input for storm surge and waves. As the construction of an offshore wind observation facility is quite expensive, its number in bays is still small. By this reason, the marine surface wind field as an input of storm surge and wave simulation should be estimated by a certain wind model.Empirical typhoon models and the MASCON model could not estimate precisely the wind field in bays.On the other hand, meteorological models have been developed for the description of meteorological processes in various scales. Among the meteorological models the Mesoscale Model MM5 with typhoon bogus is one of the models for the estimation of marine surface wind in bays, yet its applicability should be examined in detail.
　In this paper, therefore, recent six major typhoons were selected among those having significant storm surge in Seto Inland Sea. The marine surface wind field in each typhoon was estimated with 2D empirical typhoon models, the MASCON model, and the Mesoscale Model MM5 with JMA’s RANAL and typhoon bogus. Then the storm surge in the sea was simulated by using each wind field. The following results were drawn.
(1) The MM5 model with typhoon bogus estimates a marine surface wind field more accurately than 2D empirical typhoon models and also improves the accuracy of storm surge simulation.
(2) Among 2D empirical typhoon models, the model taking account typhoon movement into the equation of pressure gradient wind and introducing the wind speed coefficient for super gradient wind estimates the wind field near a typhoon center more accurately than the others.
(3) The MASCON model overestimates the effect of land topography on the horizontal component of surface wind and consequently the model tends to underestimate the marine surface wind speed in bays surrounded with mountains.
　These results show that the MM5 model with typhoon bogus is useful for the wind estimation for storm surge simulation.