TY - JOUR
T1 - Forecasting italian seismicity through a spatio-temporal physical model
T2 - Importance of considering time-dependency and reliability of the forecast
AU - Chan, Chung Han
AU - Sørensen, Mathilde B.
AU - Stromeyer, Dietrich
AU - Grünthal, Gottfried
AU - Heidbach, Oliver
AU - Hakimhashemi, Amir
AU - Catalli, Flaminia
PY - 2010
Y1 - 2010
N2 - We apply here a forecasting model to the Italian region for the spatiotemporal distribution of seismicity based on a smoothing Kernel function, Coulomb stress variations, and a rate-and-state friction law. We tested the feasibility of this approach, and analyzed the importance of introducing time-dependency in forecasting future events. The change in seismicity rate as a function of time was estimated by calculating the Coulomb stress change imparted by large earthquakes. We applied our approach to the region of Italy, and used all of the cataloged earthquakes that occurred up to 2006 to generate the reference seismicity rate. For calculation of the time-dependent seismicity rate changes, we estimated the rate-and-state stress transfer imparted by all of the ML ≥ 4.0 earthquakes that occurred during 2007 and 2008. To validate the results, we first compared the reference seismicity rate with the distribution of ML ≥ 1.8 earthquakes since 2007, using both a non-declustered and a declustered catalog. A positive correlation was found, and all of the forecast earthquakes had locations within 82% and 87% of the study area with the highest seismicity rate, respectively. Furthermore, 95% of the forecast earthquakes had locations within 27% and 47% of the study area with the highest seismicity rate, respectively. For the time-dependent seismicity rate changes, the number of events with locations in the regions with a seismicity rate increase was 11% more than in the regions with a seismicity rate decrease.
AB - We apply here a forecasting model to the Italian region for the spatiotemporal distribution of seismicity based on a smoothing Kernel function, Coulomb stress variations, and a rate-and-state friction law. We tested the feasibility of this approach, and analyzed the importance of introducing time-dependency in forecasting future events. The change in seismicity rate as a function of time was estimated by calculating the Coulomb stress change imparted by large earthquakes. We applied our approach to the region of Italy, and used all of the cataloged earthquakes that occurred up to 2006 to generate the reference seismicity rate. For calculation of the time-dependent seismicity rate changes, we estimated the rate-and-state stress transfer imparted by all of the ML ≥ 4.0 earthquakes that occurred during 2007 and 2008. To validate the results, we first compared the reference seismicity rate with the distribution of ML ≥ 1.8 earthquakes since 2007, using both a non-declustered and a declustered catalog. A positive correlation was found, and all of the forecast earthquakes had locations within 82% and 87% of the study area with the highest seismicity rate, respectively. Furthermore, 95% of the forecast earthquakes had locations within 27% and 47% of the study area with the highest seismicity rate, respectively. For the time-dependent seismicity rate changes, the number of events with locations in the regions with a seismicity rate increase was 11% more than in the regions with a seismicity rate decrease.
UR - http://www.scopus.com/inward/record.url?scp=78549251405&partnerID=8YFLogxK
U2 - 10.4401/ag-4761
DO - 10.4401/ag-4761
M3 - 期刊論文
AN - SCOPUS:78549251405
SN - 1593-5213
VL - 53
SP - 129
EP - 140
JO - Annals of Geophysics
JF - Annals of Geophysics
IS - 3
ER -