[Paper Review] Towards Evidences of Long-Range Correlations in Seismic Activity

Douglas Ferreira, Jennifer Ribeiro+2 more|arXiv (Cornell University)|May 1, 2014

Complex Systems and Time Series Analysis3 citations

TL;DR

This paper proposes a novel time-window-based network model to analyze seismic activity, avoiding limitations of traditional methods in global earthquake catalogs. The model reveals small-world topology and q-Gaussian degree distribution, indicating long-range spatial and temporal correlations in seismic events, particularly in highly active regions.

ABSTRACT

In this work, we introduce a new methodology to construct a network of epicenters that avoids problems found in well-established methodologies when they are applied to global catalogs of seisms. The new methodology involves essentially the introduction of a time window which works as a temporal filter. Our approach is more generic and for small regions the results coincide with previous findings. The network constructed with that model has small-world properties and the distribution of node connectivity follows a non-traditional $q$-Gaussian function, where scale-free properties are present. The vertices with larger connectivity in the network correspond to the areas with very intense seismic activities in the period considered. These new results strengthen the hypothesis of long spatial and temporal correlations between earthquakes.

Motivation & Objective

• To address methodological limitations in existing global seismic network models.
• To develop a more generic and robust approach for constructing seismic epicenter networks.
• To investigate the presence of long-range correlations in seismic activity across space and time.
• To validate the model on small regions where prior results are established.

Proposed method

• A time window is introduced as a temporal filter to construct a network of earthquake epicenters.
• The network is built by connecting epicenters within a defined temporal window, reducing noise from distant or unrelated events.
• The model preserves small-world network properties, enabling analysis of connectivity patterns.
• Node connectivity distribution is analyzed using a q-Gaussian function to detect scale-free characteristics.
• The method is validated on regional data, showing consistency with prior findings.
• The approach allows identification of highly active seismic zones through high-degree nodes.

Experimental results

Research questions

• RQ1 Can a time-window-based network model improve the detection of long-range correlations in global seismic catalogs?
• RQ2 Do the connectivity patterns in the seismic network exhibit scale-free or non-traditional statistical distributions?
• RQ3 Are the most active seismic regions consistently identified as high-degree nodes in the network?
• RQ4 How does the model's performance compare to established methods in small, well-studied regions?
• RQ5 What evidence does the network structure provide for spatial and temporal correlations in earthquake occurrences?

Key findings

• The constructed seismic network exhibits small-world properties, indicating efficient information propagation across the system.
• The distribution of node connectivity follows a non-traditional q-Gaussian function, suggesting complex, non-Poissonian dynamics.
• Scale-free characteristics are observed in the network, implying the presence of highly influential seismic nodes.
• Vertices with the highest connectivity correspond to regions of intense seismic activity during the study period.
• The results support the hypothesis of long-range spatial and temporal correlations between earthquakes.
• The model's consistency with prior findings in small regions validates its reliability and generality.