A scanning lidar measures a projection of the actual wind speed to the line-of-sight. Reconstruction of the original field is possible via various retrieval algorithms, especially if several lidars are deployed in the area. However, this may not be the case due to financial constraints. Additionally, conventional algorithms like Volume Velocity Processing (VVP) do not perform well for non-homogeneous flow fields, e.g., smooth wind turbine wakes in a retrieved field. A 2D-VAR algorithm using VVP as an intermediate step was suggested allowing the retrieval for consecutive scans at a low elevation angle. We validate this algorithm for scans containing prominent wakes under elevation angles within 5° using FINO1 mast and supervisory control and data acquisition (SCADA) data. We show that prominent wakes scanned at the hub height introduce substantial disturbances in the VVP retrieval output and affect the retrieved field. The unwanted disturbance may be overcome by masking the wakes. In point measurements, the wind speed magnitude is retrieved with an acceptable accuracy, but the wind direction appears to be sensitive to the weights and initial guess chosen. According to our sensitivity analysis, the weight assigned to the radial velocity residuals affects the outcome most, while other weights act as corrections.