Testing Magnitude Dependence in Earthquake Occurrence: Correlation or Observation Bias?
Giuseppe Petrillo (CNRS, Chaire Prof. Junior ; LP ENSL)
| Quand ? |
Le 07/07/2026, de 13:00 à 14:00 |
|---|---|
| Où ? | M7 101 |
| Participants |
Giuseppe Petrillo |
| Ajouter un événement au calendrier |
vCal
iCal
|
Giuseppe Petrillo
Title: Testing Magnitude Dependence in Earthquake Occurrence: Correlation or Observation Bias?
Abstract: Earthquake magnitudes are commonly assumed to be independent marks in self-exciting point-process models such as ETAS (epidemic-type aftershock sequence model). However, several recent studies have reported apparent correlations between the magnitudes of triggering and triggered events. In this seminar, I present a statistical framework to test this hypothesis while accounting for observational biases arising from short-term catalog incompleteness. Combining stochastic declustering, rank-based transformations, and synthetic catalogs generated from a physics-based model, we show that the observed magnitude dependence can largely be explained by missing observations rather than genuine correlations. The results illustrate how incomplete data may generate spurious dependence structures in marked point processes and highlight the importance of carefully separating physical effects from sampling artifacts.
Website: https://www.linkedin.com/in/giuseppe-petrillo-42496563/ ; https://scholar.google.com/citations?user=PMH0LM4AAAAJ&hl=it
In Room M7 101, 1st floor, Monod campus, ENSL