Title:Multifractal detrended fluctuation analysis of rainfall time series in the Guadeloupe archipelago

Abstract:Due to the vulnerability of the Caribbean islands to the climate change issue, it is important to investigate the behavior of rainfall. In addition, the soil of the French West Indies Islands has been contaminated by an insecticide (Chlordecone) whose decontamination is mainly done by drainage water. Thus, it is crucial to investigate the fluctuations of rainfall in these complex environments. In this study, 19 daily rainfall series recorded in different stations of Guadeloupe archipelago from 2005 to 2014 were analyzed with the multifractal detrended fluctuation analysis (MF-DFA) method. The aim of this work is to characterize the long-range correlations and multifractal properties of the time series and to find geographical patterns over the three most important islands. This is the first study that addresses the analysis of multifractal properties of rainfall series in the Caribbean islands. This region is typically characterized by the almost constant influence of the trade winds and a high exposure to changes in the general atmospheric circulation. 12 stations exhibit two different power-law scaling regions in rainfall series, with distinct long-range correlations and multifractal properties for large and small scales. On the contrary, the rest of stations only show a single region of scales for relatively small scales. Hurst exponents reveal persistent long-range correlations. In the most eastern analyzed areas, larger scales exhibit higher persistence than smaller scales, which suggests a relationship between persistence and the highest exposure to the trade winds. Stronger conclusions can be drawn from multifractal spectra, which indicate that most rainfall series have a multifractal nature with higher complexity and degree of multifractality at the smallest scales. Furthermore, a clear dependence of multifractal nature on the latitude is revealed.