Change of scaling of background activity events in complex systems before extreme events

  1. International Institute of Earthquake Prediction Theory and Mathematical Geophysics, Russian Academy of Sciences

Socio-economic and natural complex systems persistently generate extreme events also known as disasters,
crises, or critical transitions. Here we analyze patterns of background activity preceding extreme events in four
complex systems: the U.S. economy, a megacity (Los Angeles), the Earth magnetosphere, and the Earth litho-
sphere. Economic recessions, surges in homicides in a megacity, magnetic storms, and strong earthquakes are
considered as extreme events in these systems. We use as a starting point the indicators describing the system’s
behavior. Rather than analyze the whole time series of an indicator, we identify sufficiently strong changes in
an indicator’s trend. Those changes constitute our background events (BE). We demonstrate a premonitory
pattern common to all four systems considered: relatively large magnitude BEs become more frequent before
an extreme event. A premonitory change of scaling has been found in various models and observations. Here
we demonstrate this change in scaling of uniformly defined BEs in four real complex systems, their enormous
differences notwithstanding.