It is common practice to evaluate mesoscale meteorological models against in-situ mast measurements of one or several met masts. With the NEWA Ferry Lidar Experiment and Benchmark conducted as part of the New European Wind Atlas (NEWA) project we further develop the concept of spatially distributed wind profile measurements. A vertically scanning Doppler lidar was placed on a ferryboat travelling on a regular route through the Southern Baltic Sea between Kiel (Germany) and Klaipeda (Lithuania). During four months between February and June 2017, the lidar continuously measured wind profiles between 65 m and 275 m above sea level along the ship track. The obtained wind profile data were corrected for ship motions (i.e. ship velocity and heading). To our knowledge this is the first dataset of wind profile measurements with lidar along such a long track. As the lidar moves with the ship and covers a distance that is comparable to a typical mesoscale model domain, this dataset is very suitable to be compared against meteorological mesoscale model data. In the associated modelling benchmark, participants from the international mesoscale modelling community were invited to simulate the period of the experiment with different mesoscale models and model setups. The participants were encouraged to either use their best practice setup or to test several setups. The objectives of the benchmark are: To assess how well today’s mesoscale models can reproduce the wind conditions offshore and, in particular, in the Southern Baltic Sea where quite often coastal effects are present. To compare different models as well as different model setups of the same model against the measurements. To gain experience with this unique kind of data (moving wind profiles) and explore its strengths and weaknesses. Several participants used the WRF model with different setups and forcing data but also ICON-EU operational forecasts and MM5 are participating in the benchmark. The simulated time series are compared with the measured time series in different heights, and vertical profiles are evaluated at specific locations. Furthermore, error statistics are presented.