Abstract. The development of large wind turbines and airborne wind energy (AWE) systems requires reliable wind speed datasets at heights above the atmospheric surface layer. Traditional measurement approaches relying on met masts (meteorological masts) fall short of addressing these needs. In this study, we validate three different model-based datasets, namely the 3 km Norwegian Hindcast archive (NORA3), the New European Wind Atlas (NEWA), and ECMWF Reanalysis v5 (ERA5), using Doppler wind lidar data from several locations in Norway and the North Sea. The validation focuses on altitudes from 100 to 500 m above ground, covering the operational range of large wind turbines and AWE systems. Our findings indicate that ERA5 and NORA3 perform similarly well in offshore locations in terms of bias, correlation coefficient, root-mean-square error, and Earth mover's distance. The choice of an appropriate wind speed database depends on the topography, altitude and error metrics of interest. However, NORA3 outperforms the other two models in two coastal sites and one complex-terrain site. In most cases, the agreement between the models and lidar measurements increases with height.