Complex network based techniques to identify extreme events and (sudden) transitions in spatio-temporal systems

Abstract

We present here two promising techniques for the application of the complex network approach to continuous spatio-temporal systems that have been developed in the last decade and show large potential for future application and development of complex systems analysis. First, we discuss the transforming of a time series from such systems to a complex network. The natural approach is to calculate the recurrence matrix and interpret such as the adjacency matrix of an associated complex network, called recurrence network. Using complex network measures, such as transitivity coefficient, we demonstrate that this approach is very efficient for identifying qualitative transitions in observational data, e.g., when analyzing paleoclimate regime transitions. Second, we demonstrate the use of directed spatial networks constructed from spatio-temporal measurements of such systems that can be derived from the synchronized-in-time occurrence of extreme events in different spatial regions. Although there are many possibilities to investigate such spatial networks, we present here the new measure of network divergence and how it can be used to develop a prediction scheme of extreme rainfall events.

An event synchronization method to link heavy rainfall events and large‐scale atmospheric circulation features

“…() and () identified spatial patterns of extreme rainfall events of the South American monsoon system and the impact of ENSO on moisture divergence in South America. Marwan and Kurths () demonstrated the potential of the ES approach to predict hydrological extremes.…”

Can recurrence networks show small-world property?

“…In this work, we consider the second method for the construction of the network where a reference node ı is connected to another node  if the Euclidean distance d ij between the corresponding points on the attractor in the reconstructed space is less than or equal to the recurrence threshold ǫ, that is, if d ij ≤ ǫ. The resulting complex network, called the ǫ -recurrence network or simply recurrence network (RN) [26,27], has been shown to have great potential for a wide range of practical applications, from identifying critical transitions in dynamical systems [28] to the classification of cardio-vascular time series [29]. Note that, by construction, the RN is an undirected and unweighted graph with a symmetric and binary adjacency matrix A, with elements A ij = 1 or 0, depending on whether the two nodes ı and  are connected or not.…”

Alignment of Lyapunov Vectors: A Quantitative Criterion to Predict Catastrophes?

“…Considerable work was done in the last years to understand the mechanisms behind the emergence of large-amplitude pulses3. An idea of the intense activity concerning outliers in several distinct lasers, in nanophotonic devices and media, in excitable systems, and in other key applications may be obtained by perusing a small selection of representative papers published in the last two years456789101112131415.…”