Improving Storm surge modeling along the French (Atlantic and English Channel) coast

Research on storm surge modeling along the French coast was conducted in the framework of the PREVIMER project aiming at providing observations, modeling tools and real-time forecasts for coastal zone uses and management. One of PREVIMER project's deliverable is the development of a coastal operational forecasting system providing hydrodynamics and hydrological data in French maritime areas. The objective of this study is to better understand the storm surge dynamics along the French coasts, as well as to improve high frequency sea level variations modeling (and specifically storm surge) within the PREVIMER system, based on the regional ocean Model for Applications at Regional Scale (MARS 2D). First, storm surge generation and propagation processes (tide and surge interaction, contribution of wind and pressure to the storm surge) are studied. Comparisons between modeled storm surges issued from MARS with or without tides shows that interactions can reach about several tens of centimeters (up to 63 cm at Dunkerque). They can be positive - and amplify surges implied by wind and pressure - or negative. The analysis shows that most of the French coast is submitted to non-negligible interactions between storm surges and waves, at the exception of some islands, Brittany, and a part of the Aquitania coast. The areas where wind and pressure terms are predominant in storm surge signal are also identified. It appears that the areas where storm surge and tide interactions are negligible correspond to the areas where pressure-induced storm surge is dominant, compared to wind-induced storm surge. Such result allows designing a modeling strategy to forecast the water level (tide and storm surge), optimizing computational time and water level prediction using the better tidal data available. Besides, a sensitivity study on sea surface drag parameterization, based on comparisons with 4 tidal gauges data (Dunkerque, Saint-Malo, Le Conquet and La Rochelle), during 4 storm events and over about 7.5 years are performed to evaluate the quality of the modeled storm surge. Tidal gauges data are post processed using the same pattern as with MARS results (tidal component analysis), so as to obtain measured storm surge data. The tested surface drag coefficient parameterizations are mainly a constant one, Wu (1982) and Charnock (1955). Charnock's parameterization, relying on a full statistical description of the sea state (wavewatch III simulations) provided by the IOWAGA project, enables to reproduce satisfying storm surges with 10 cm differences between the storm surge peak and with the ones obtained with the other drag coefficient formulations. For the events where the meteorological inputs are the closest to spatial observations, the model predicts storm surges quite close (differences about a few centimeters) to the observations. The results of this study allow not only to improve the existing storm surge model used in the PREVIMER system, but also to provide a modeling strategy to optimize computational cost and water level forecast along the French Atlantic and English Channel coast. References Charnock, H. (1955) Wind stress on a water surface. Quart. J. Roy. Meteor. Soc., 81, 639-640. Wu, J. (1982) Wind-stress coefficients over sea surface from breeze to hurricane. J.Geophys. Res. 87, 9704-9706.