Expanding Accessibility and Scalability of Ambient Noise Seismic Data Processing Tools Through an Open-Source Cloud-Based Software Application

Session: Applications and Technologies in Large-Scale Seismic Analysis
Type: Oral
Date: 4/28/2020
Time: 09:00 AM
Room: 120 + 130
Description: 

Seismology has historically spearheaded the use of high-performance computing (HPC) resources for numerical modeling. However, most currently available seismic noise processing software packages are not yet adequately equipped to fully exploit such resources nor address the challenges in processing continuously-sampled seismic data on a petabyte scale. Those that are equipped are often not transparent to the scientific community. Furthermore, few tools are accessible to researchers who desire to take advantage of HPC, but do not have access to a dedicated supercomputing facility. In a new project supported by Earth Science Information Partners (ESIP), we are developing an open-source, cloud-based application in order to democratize HPC-enabled processing tools for high-volume seismic data. We host the application on Amazon Web Services (AWS) S3 and interface with the Incorporated Research Institutions for Seismology (IRIS) data services to query station data and metadata for seismic noise cross-correlation. To maximize efficiency on large datasets, the correlation processing is parallelized in Julia, a modern computing language developed to make full use of HPC resources from the cloud. In this way, the cross-correlations are able to be conducted from a laptop or workstation. This project is a gateway to further development beyond noise cross-correlation so that ultimately, anyone can process and interpret high-volume seismic data. In this presentation we report on the current state of this project and discuss future directions.

Presenting Author: Theophilia Sukianto

Authors