AbstractWe present ice-free and ice-included statistics for the Baltic Sea using a wave hindcast validated against data from 13 wave measurement sites. In the hindcast 84% of wave events with a significant wave height over 7 m occurred between November and January. The effect of the ice cover is largest in the Bay of Bothnia, where the mean significant wave height is reduced by 30% when the ice time is included in the statistics. The difference between these two statistics are less than 0.05 m below a latitude of 59.5°. The seasonal ice cover also causes measurement gaps by forcing an early recovery of the instruments. Including the time not captured by the wave buoy can affect the estimates for the significant wave height by roughly 20%. The impact below the 99th percentiles are still under 5%. The significant wave height is modelled accurately even close to the shore, but the highest peak periods are underestimated in a narrow bay. Sensitivity test show that this underestimation is most likely caused by an excessive refraction towards the shore. Reconsidering the role of the spatial resolution and the physical processes affecting the low-frequency waves is suggested as a possible solution.