Author(s): Antonietta Simone; Carlo Tucci; Gabriele Freni; Luigi Berardi; Daniele Biagio Laucelli
Linked Author(s): Gabriele Freni
Keywords: Climate change; Rainfall time series; Time series analysis; Time series transformation; Visibility domain
Abstract: The growing impacts of climate change are pushing the scientific community to develop more effective strategies for interpreting natural phenomena. This study introduces a novel strategy to gain deeper insights into the behaviour of rainfall time series and to enhance their classification. The method is based on geometric visibility relationships between data points within a time series. The idea is to represent the time series not in its original domain but in a visibility domain, obtained through a transformation of the series. This can be done with respect to either maxima or minima, effectively treating visibility as an extreme value transform that maps ordered data to highlight anomalies and extreme values, providing a meaningful classification of the phenomenon. The results have proven promising, showing that the mapping enhances the discriminating power of the analysis. Pearson correlation highlighted the potential of the approach for classifying rainfall time series and, more generally, complex systems/ phenomena. The analysis was carried out on rainfall time series from the island of Ischia (Italy), covering the past 16 years.
Year: 2026