Agent-based models of human response to natural hazards: systematic review of tsunami evacuation.

This systematic review provides a comprehensive overview of tsunami evacuation models. The review covers scientific studies from the last decade (2012-2021) and is explicitly focused on models using an agent-based approach. The PRISMA methodology was used to analyze 171 selected papers, resulting in over 53 studies included in the detailed full-text analysis. This review is divided into two main parts: (1) a descriptive analysis of the presented models (focused on the modeling tools, validation, and software platform used, etc.), and (2) model analysis (e.g., model purpose, types of agents, input and output data, and modeled area). Special attention was given to the features of these models specifically associated with an agent-based approach. The results lead to the conclusion that the research domain of agent-based tsunami evacuation models is quite narrow and specialized, with a high degree of variability in the model attributes and properties. At the same time, the application of agent-specific methodologies, protocols, organizational paradigms, or standards is sparse. Supplementary Information: The online version contains supplementary material available at 10.1007/s11069-022-05643-x.

Short-Term Foreshocks as Key Information for Mainshock Timing and Rupture: The Mw6.8 25 October 2018 Zakynthos Earthquake, Hellenic Subduction Zone.

Significant seismicity anomalies preceded the 25 October 2018 mainshock (Mw = 6.8), NW Hellenic Arc: a transient intermediate-term (~2 yrs) swarm and a short-term (last 6 months) cluster with typical time-size-space foreshock patterns: activity increase, b-value drop, foreshocks move towards mainshock epicenter. The anomalies were identified with both a standard earthquake catalogue and a catalogue relocated with the Non-Linear Location (NLLoc) algorithm. Teleseismic P-waveforms inversion showed oblique-slip rupture with strike 10°, dip 24°, length ~70 km, faulting depth ~24 km, velocity 3.2 km/s, duration 18 s, slip 1.8 m within the asperity, seismic moment 2.0 × 1026 dyne*cm. The two largest imminent foreshocks (Mw = 4.1, Mw = 4.8) occurred very close to the mainshock hypocenter. The asperity bounded up-dip by the foreshocks area and at the north by the foreshocks/swarm area. The accelerated foreshocks very likely promoted slip accumulation contributing to unlocking the asperity and breaking with the mainshock. The rupture initially propagated northwards, but after 6 s stopped at the north bound and turned southwards. Most early aftershocks concentrated in the foreshocks/swarm area. This distribution was controlled not only by stress transfer from the mainshock but also by pre-existing stress. In the frame of a program for regular monitoring and near real-time identification of seismicity anomalies, foreshock patterns would be detectable at least three months prior the mainshock, thus demonstrating the significant predictive value of foreshocks.