Wind lidars present advantages over meteorological masts, including simultaneous multipoint observations, flexibility in measuring geometry, andreduced installation cost. But wind lidars come with the “`cost” of increased complexity in terms of data quality and analysis. Carrier-to-noiseratio (CNR) has been the metric most commonly used to recover reliable observations from lidar measurements but with severely reduced datarecovery. In this work we apply a clustering technique to identify unreliable measurements from pulsed lidars scanning a horizontal plane, takingadvantage of all data available from the lidars – not only CNR but also line-of-sight wind speed (VLOS), spatial position, andVLOS smoothness. The performance of this data filtering technique is evaluated in terms of data recovery and data quality against botha median-like filter and a pure CNR-threshold filter. The results show that the clustering filter is capable of recovering more reliable data innoisy regions of the scans, increasing the data recovery up to 38 % and reducing by at least two-thirds the acceptance of unreliablemeasurements relative to the commonly used CNR threshold. Along with this, the need for user intervention in the setup of data filtering is reducedconsiderably, which is a step towards a more automated and robust filter.