Mixing height derivation from aerosol lidar using machine learning: KABL and ADABL algorithms

Overview & Synopsis

Abstract. Atmospheric boundary layer height (BLH) is a key parameter for several meteorological applications, for example air quality forecast. To measure it, a common practice is to use aerosol lidars: a strong decrease in the backscatter signal indicates the top of the boundary layer. This paper describes and compares two methods of machine learning to derive the BLH from backscatter profiles: the K-means algorithm and the AdaBoost algorithm. Their codes are available under a fully open access, with the name KABL (K-means for Atmospheric Boundary Layer) and ADABL (AdaBoost for Atmospheric Boundary Layer). Both methods are compared to the lidar manufacturer's software and to reference BLH derived from collocated radiosondes. The radiosondes are taken as the reference for all other methods. The comparison is carried out on a two-year long period (2017–2018) on 2 Meteo-France's operational network sites (Trappes and Brest). Results show that, although its training is limited, ADABL is performing better than KABL and can easily be improved by enhancing its training set. However, KABL can be easily adapted for other instrumental device and used to make instrument synergy, while ADABL must be fully re-trained at each change in the instrument settings.