Abstract. Wake measurements of a scanning Doppler lidar mounted on the nacelle of a yawed full-scale wind turbine are used for the characterization of the wake flow and the validation of analytical wake models. Inflow scanning Doppler lidars, a meteorological mast and the data of the wind turbine control system complemented the set-up. Results showed an increase of the wake deflection with the yaw angle that agreed with two of the three compared models. For yawed cases, the predicted power of a waked downwind turbine estimated by the two previously successful models had an error of 17 % and 24 % compared to the SCADA data and 12 % and 13 % compared to the power estimated from the Doppler lidar measurements. Shortcomings of the method to compute the power coefficient in an inhomogeneous wind field are likely the reason for disagreement between estimates using the Doppler lidar data versus SCADA data. Further, it was found that some wake steering cases were detrimental to the power output due to errors of the inflow wind direction perceived by the yawed wind turbine and the wake steering design implemented. Lastly, it was observed that the spanwise cross-section of the wake is strongly affected by wind veer, masking the kidney-shaped wake cross-sections observed from wind-tunnel experiments and numerical simulations.