Time series analysis of wildfire propagation in Chile: A complex networks approach

Abstract

Wildfires are a recurrent and devastating phenomenon that has expanded in Chile and various places worldwide, with complex propagation patterns influenced by environmental, climatic, and anthropogenic factors. In this study, we employ a complex network framework to analyze the temporal and spatial dynamics of fire propagation throughout Chile. Using publicly available data on wildfires, including their locations and dates of occurrence, we map the position and timing of events onto an undirected spatial complex network to characterize fire propagation in Chile’s central and southern zones from a multifractal point of view. We construct a time series associated with the connectivity of nodes, where each event is assigned to a node by position match. Traversing the network sequentially, based on the date and time of the events, generates the connectivity time series, which captures spatiotemporal information. Our analysis demonstrates that the graph exhibits hallmark complexity features. In addition, through an analysis of the connectivity time series, we uncover the multifractal nature of the system. In particular, we show that the central zone exhibits a more remarkable persistence of multifractal behavior than the southern zone, which could provide insights into the underlying dynamics of the system and be valuable for identifying regions with heightened structural or temporal complexity. We expect our findings will contribute to understanding the spatial heterogeneity of fire distribution in Chile and serve as a precedent for future research.