Variation dynamics of the complex topology of a seismicity network

Reduction in the scaling exponent of the frequency–magnitude power law of regional seismic activity as a precursor to sizable earthquakes has been widely documented and debated. Recently, postulation based on a modified sand-pile model has been proposed as a plausible explanation. The model demonstrates systematic variations in the frequency–size scaling exponent of avalanches through the introduction of varying degrees of randomness to the conventional regular, nearest-neighbor network connection. In this study, we examined a connection network of successive events in the Taiwan seismicity, in an attempt to shed lights on the behavior of the actual earthquake network. The revealed nature of connection is indeed quite different from the nearest-neighbor interaction usually presumed in most conventional seismicity modeling. However, monthly variations in the statistics of the connection degree, the connection time and the connection distance that reflect important transition dynamics of the regional seismicity network, are inconsistent with the postulation based on the modified sand-pile model that attributes the scaling exponent variation to the varying degree of long range connections.