Simulation models, generating meteo-oceanic conditions and submerged areas scenarios, are a major asset for forecasting coastal floods and prepare at risk populations. Forecasting is still performed by human operators, using visualization interfaces. In order to take model uncertainty into account in the determination of alert levels, advances in computing allow to perform ensemble simulation, by generating a set a possible scenarios. The visual interpretation of ensemble simulation outputs however requires specific approaches of visualization, allowing human operators to determine trends, identify outliers, and assess uncertainty, from many simulated scenarios at different spatial scales. While geospatial ensemble visualization approaches mainly use 2D representation of aggregated indicators, 3D interfaces could be a way to depict more precisely the distribution of ensemble at different spatial locations. We present 2D and 2D+1 visualization approaches aiming at helping forecasters to visually analyze ensemble simulation of submerged coastal areas, at different spatial scales and for different levels of complexity.
