PM10 fluctuation modeling in the Caribbean area using the universal multifractals framework

Every year, the Caribbean basin is strongly impacted by sand mists from African deserts. The health impact is significant for the population living there. This is the reason why a better understanding of the behavior of particulate matter that have an aerodynamic diameter less than 10 µm diameter (PM10) is crucial to predict their fluctuations. The aim of this study is to characterize the PM10 fluctuations in the fully developed turbulence framework. For that, this analysis is carried out using PM10 datasets sampled at daily basis during six years period for three Caribbean islands (Martinique–Guadeloupe-Puerto Rico). After a multifractal analysis, the results obtained show that the log-Lévy model is suitable, in comparison to the log normal model, to fit the scaling exponent function ζ(q) and the multifractal spectrum f(α). Under this basis, a PM10 fluctuations characterization for each island is proposed using the three universal multifractal parameters [1,2]. Hence, stochastic simulations can be envisaged to mimic the stochastic behavior of PM10 data.

References

[1] Schertzer, D., Lovejoy, S., 1987. Physical modeling and analysis of rain and clouds by anistropic multiplicative processes. J. Geophys. Res. 92 (D8), 9693-9714.

[2] Schertzer, D., Lovejoy, S., Schmitt, F., Chigirinskaya, Y., Marsan, D., 1997. Multifractal cascade dynamics and turbulent intermittency. Fractals 5 (3), 427-471.