Improved boundary layer height measurement using a fuzzy logic method: Diurnal and seasonal variabilities of the convective boundary layer over a tropical station

Abstract This paper presents the efficacy of a “tuned” fuzzy logic method at determining the height of the boundary layer using the measurements from a 1280 MHz lower atmospheric radar wind profiler located in Gadanki (13.5°N, 79°E, 375 mean sea level), India, and discusses the diurnal and seasonal variations of the measured convective boundary layer over this tropical station. The original fuzzy logic (FL) method estimates the height of the atmospheric boundary layer combining the information from the range‐corrected signal‐to‐noise ratio, the Doppler spectral width of the vertical velocity, and the vertical velocity itself, measured by the radar, through a series of thresholds and rules, which did not prove to be optimal for our radar system and geographical location. For this reason the algorithm was tuned to perform better on our data set. Atmospheric boundary layer heights obtained by this tuned FL method, the original FL method, and by a “standard method” (that only uses the information from the range‐corrected signal‐to‐noise ratio) are compared with those obtained from potential temperature profiles measured by collocated Global Positioning System Radio Sonde during years 2011 and 2013. The comparison shows that the tuned FL method is more accurate than the other methods. Maximum convective boundary layer heights are observed between ~14:00 and ~15:00 local time (LT = UTC + 5:30) for clear‐sky days. These daily maxima are found to be lower during winter and postmonsoon seasons and higher during premonsoon and monsoon seasons, due to net surface radiation and convective processes over this region being more intense during premonsoon and monsoon seasons and less intense in winter and postmonsoon seasons.