Spectral characteristics of surface atmosphere in range of macroscale to microscale at Hong Kong

Abstract Many atmospheric processes are stochastic rather than completely deterministic, which limits the applicability of some traditional analysis techniques. Spectral analysis is a well-established and useful tool to analyze different time series from a statistical standpoint, which is particularly suited for the characterization of meteorological variables. Spectral analysis is attractive to meteorologist as it enables the identification of preferred periods/frequencies at which oscillation of meteorological variables are more likely to occur. Such information can practically aid the meteorological forecast and provide insights in various sectors of engineering practices. This study presents a comprehensive investigation of various meteorological variables (i.e., wind speed and direction, air temperature, pressure and relative humidity) using a 30-yr field measurement database available at 8 weather stations across Hong Kong, with particular emphasis on the characterization of the hidden periodic oscillation by means of spectral analysis. More importantly, the data involved in this study consist of various sampling frequencies and duration, which have been well used to describe the spectral behavior of meteorological variables over an extended range of frequencies. Typical oscillation periods of each meteorological variables have been identified individually, which are generally centered at 1-yr and 1-day. Some secondary oscillation periods have also been found in the spectra curves of different variables, which, however, dependent upon the specific meteorological component of concern.