Using an atomic/molecular vapor as an aerosol blocking filter for atmospheric temperature measurements with a Cabannes lidar is revisited. Different problems in previously used barium and iodine filters prevented them from delivering the 78 times signal advantage (8.8 times less uncertainty) over rotational Raman lidar. We conclude that, despite the sensitivity optimization in rotational Raman lidar, a proposed Cabannes lidar utilizing potassium vapor filters can have 6.1 times less temperature uncertainty. By tuning the laser frequency cyclically to above and below the potassium D1 transition, the lidar system can measure temperature and wind simultaneously.