Item

Abstract

A powerful seismic sequence struck the Turkey-Syria border on Feb. 6 with three earthquakes having M>6.5. At 01:17 UTC, a M7.8 earthquake occurred that was followed 11 minutes later by a M6.7, and around nine hours later by a M7.5. In the first 24 hours, 20 aftershocks with M>5 and 68 events with M>4 have been registered. Recently, we have improved and tested a Bayesian simulation-based workflow for spatio-temporal early seismicity forecasting based on ETAS model. It exploited the versatility of the Bayesian inference to adaptively update the forecasts based on the incoming information from the ongoing sequence. This workflow is demonstrated and verified through retrospective early seismicity forecasting of Central Italy 2016 and the 2017-2019 western Iran seismic sequences. We test this workflow to predict the spatial distribution of events and their uncertainties for various forecasting intervals within this seismic sequence. Bayesian updating is first employed to learn the ETAS model parameters conditioned on the registered events (that already took place). Then, plausible sequences of events during the forecasting interval are adaptively generated. To this end, we strive to simulate those plausible sequences by embedding a branching process formulation inside the proposed workflow as an alternative to the piece-wise stationary integration of the conditional rate. The latter could be a new feature to the forecasting workflow, while its efficiency in providing early forecasts is explored during this study. This work has been supported by PRIN-2017 MATISSE project No 20177EPPN2, funded by Italian Ministry of Education and Research.