← Back to Sessions

Leveraging the Iran Regional Moment Tensor database to estimate parameter uncertainties in 1D and 3D Earth models

Description: 

The Arabia-Eurasia plate convergence is accommodated in a wide deforming zone spanning most of Iran resulting in frequent moderate-to-large earthquakes. We used in-country data to build a regional moment tensor database that includes over 2,000 events from 2004 to 2022 ranging in size from Mw 3.4 to 7.8. The magnitude of completeness is near Mw 4.2 since 2013 when broadband data from the Iranian Seismological Center became available in addition to data from the Iranian National Seismic Network. We used a single 1-D velocity model for all events adjusting the modeling frequency band based on event size and event-station distances when inverting 3-component data for the deviatoric moment tensor. To better understand and constrain source parameter uncertainties and the effects of unmodeled 3D structural effects, we are now re-analyzing selected events using the MTUQ (moment tensor uncertainty quantification) software, which performs a grid search over the full parameter space - double couple, deviatoric, or full moment tensor - for 1D and 3D velocity models. The computationally expensive calculation of 3D synthetics and fine-scale grid searches are performed on an HPC facility at USF. Initial results indicate that double couple and deviatoric grid searches usually agree well with earlier inversion results, while full moment tensor searches using surface waves only may result in significant, spurious non-deviatoric contributions. The use of appropriate 3D models significantly improves waveform fits and seems to reduce the non-double couple source contributions. Grid searches also reveal that a wide range of source parameters often results in misfits relatively close to a best-fit solution. This is particularly true for long-period analyses or datasets with limited azimuthal coverage and/or data with a low signal-to-noise ratio.

Session: Earthquake Source Parameters: Theory, Observations and Interpretations

Type: Oral

Date: 4/18/2023

Presentation Time: 08:00 AM (local time)

Presenting Author: Jochen Braunmiller

Student Presenter: No

Invited Presentation: 

Authors