Validation of Seismic Crustal and Mantle Models of the Contiguous U.S.

Date: 4/26/2019

Time: 09:00 AM

Room: Vashon

Since the completion of the Transportable Array deployment across the contiguous US, there have been various seismic models with improved image resolution inferred for the crust and (mostly) mantle beneath US (e.g., DNA13, US-SL-2014, and US.2016). However, discrepancies exist between these models due to the differences in both data sets and tomographic methods. In order to build a unified seismic model for the contiguous US, it is an essential (often missing) step to validate these models and to find out which model achieves the optimal predictability, i.e., best predicting complex wave propagation phenomena at higher frequencies. In this study we systematically compared a total of eight (regional and global) models publically available through the IRIS EMC-EarthModels, in terms of the waveform fitting between data and synthetics. The criteria for model predictive capability include waveform similarities and travel time misfits for different phases (P, S, and surface waves) for different epicentral distances using USArray stations. The synthetics are calculated based on the spectral-element method with grid spacing of about 5 km in the crust and 10 km in the mantle within a model region of 36 by 48 degrees, which includes the western, central, and part of eastern US. Both synthetics and data are filtered between 10 and 40 seconds and low signal-to-noise-ratio (SNR < 10) data traces are excluded from the comparison. All the 3-D models without good crustal imaging resolution are implemented in the simulation region with CRUST1.0 embedded except US.2016, which instead is combined with model AK135 for the mantle below the depth of 150 km. The results of the model predictive capabilities will be presented along with the proposed solution toward a unified model of the contiguous US.

Presenting Author: Min Chen

Authors