Unlike European, Asian or American megacities, the high concentrations of fine particles with a diameter of less than 10 µm (PM10) in Guadeloupe are mainly due to sand mists. In partnership with the University of the West Indies, the KaruSphère laboratory studied the statistical behavior of PM10. For the first time in the Caribbean region, the empirical distribution of daily PM10 concentrations is modelled. In the literature, the statistical modeling of these concentrations is classically carried out with unimodal models by probability density functions of the log-normal or Weibull type. Our results showed that the Burr & Weibull is the best distribution with a first statistical behavior corresponding to the low dust season (October – April) and another to the high dust season (May – September). Analyzing the statistics of extreme events with the classical power law distribution, we observe that the Burr & Weibull could also improve the modeling of these events. In summary, the Burr & Weibull is suitable for modeling both classical and extreme events.