LAUSR

Abstract Ground-level ozone ( O 3 ) is a major air pollutant which is strongly links to weather conditions. To our knowledge, there are few analysis of O 3 behavior in Caribbean region. The aim of this paper is to study the correlation between O 3 and meteorological parameters (i.e., solar radiation, temperature, relative humidity and wind speed) in a multi-scale way. Firstly, we estimate their Fourier power spectra and consider their scaling properties in the physical space. The power spectra computed follows a power law behavior for all datasets. Thereafter, the multiple scale dynamics and the correlations between simultaneous measurements of O 3 and meteorological parameters was analyzed using the Empirical Mode Decomposition (EMD) based methods. After EMD decomposition of each time series, we investigate their local correlations using a new cross-correlation technique termed Time Dependent Intrinsic Correlation method (TDIC). The time evolution and the scale dependence of their correlation are determined at different time scales and for different intrinsic modes functions. All measured TDICs have highlighted strong correlations between O 3 and meteorological parameters. The impact of weather conditions (e.g., drought, rainfall, windy period) have been locally observed on O 3 formation for all the time scales. This study shows that TDIC method is a powerful tool for perform multi-scale cross-correlation analysis in air pollution field.