Due to the fragility of the Caribbean islands in the face of climate change, it is crucial to understand precipitation patterns. Additionally, the soils of the French Antilles islands have been contaminated by an organochlorine insecticide (Chlordane), which undergoes decontamination primarily through natural soil leaching. Therefore, it is essential to study the spatiotemporal variations of precipitation in these complex environments. This study focused on analyzing data from 19 meteorological stations in the Guadeloupe archipelago over the period 2005-2014, using the multifractal detrended fluctuation analysis (MF-DFA) method. The results indicate that 12 stations exhibit two distinct regions in terms of the scale of power law in their precipitation series, with persistent correlations over long periods and multifractal properties on both large and small scales. In contrast, the other stations show only a single scaling region for relatively restricted scales.
Interestingly, in the easternmost areas, larger scales exhibit more pronounced persistence compared to smaller scales, suggesting a link between this persistence and higher exposure to trade winds. Multifractal spectrum analyses reveal that most rainfall datasets have a multifractal nature, with higher complexity and degree of multifractality at smaller scales. Furthermore, a clear dependence of the multifractal nature on latitude is demonstrated. These findings highlight the significance of microclimates in island environments, providing a clear perspective to better understand these specific contexts in the comprehension and management of climate change.
