UNDERGRADUATE STUDIES IN EARTHQUAKE INFORMATION TECHNOLOGY

 

ABOUT GRAND CHALLENGES INTERNS PROJECTS

Summary
The 2018 intern cohort developed a computational system for evaluating how well the Uniform California Earthquake Rupture Forecast, version 3 (UCERF3), can predict long-term rates of M ≥ 7 ruptures on the southern San Andreas Fault. In lieu of real data, they tested the UCERF3 model against long synthetic seismicity catalogs for Southern California generated by running the RSQSim rupture simulator on the Blue Waters supercomputer.
Using the SCEC Virtual Display of Objects (SCEC-VDO), they visualized the full RSQSim catalog, as well as the RSQSim catalog reduced to M ≥ 7 ruptures on the southern San Andreas fault system. They compared the skill of UCERF3 relative to the ideal RSQSim forecast and a spatially independent reference forecast. A group of interns investigated how machine learning might be used to derive a statistical earthquake forecast from deterministic RSQSim simulations. From the RSQSim rupture set, they selected three earthquake scenarios on the southern San Andreas Fault that are among the most threatening in terms of annualized expected loss, and they illustrated their hazard and risk with maps of expected ground motions, economic losses, and human casualties.

Challenge Statement
Develop a computational system for evaluating how effective the Uniform California Earthquake Rupture Forecast, version 3 (UCERF3), can predict long-term rates of M ≥ 7 ruptures on the southern San Andreas Fault. In lieu of real data, test the UCERF3 model against long synthetic seismicity catalogs for Southern California generated by running the RSQSim rupture simulator on the Blue Waters supercomputer. Use SCEC-VDO to visualize the full RSQSim catalog as well as RSQSim catalog reduced to M ≥ 7 ruptures on the southern San Andreas fault system. Compare the skill of UCERF3 relative to ideal RSQSim forecast and a spatially independent reference forecast. Explore applications of machine learning to derive a statistical earthquake forecast. Illustrate hazards and risks of ruptures on the San Andreas Fault.
 

Intern Class of 2018

 

 

 

Project Teams


 		   High Perfomance Computing Team

Task: Run RSQSim on Blue Waters to generate long seismicity catalogs for analysis

Team: Varduhi Kababjyan, Shril Panchigar, Anthony Lopez, Anthony Guerra, Tomoe Mizutani

Mentors: Scott Callaghan, Jacqui Gilchrist

   
         

 		   SCEC-VDO Development Team

Task: Refine and update plugins, create visualizations, and formally release SCEC-VDO

Team: Brandon Ho, Shalani Weerasooriya, Trent Jones, Bill Addo, Alejandro Narvaez, Tiffany Streitenberger, Dian Zhu

Mentors: Kevin Milner, John Yu

   
         

 		   Probabilistic Forecasting Team

Task: From the synthetic catalog statistics, determine probabilities for M≥7 aftershocks following three scenarios (M6 Parkfield, M7 Mojave, and M6 Bombay beach) and compare these probabilities to UCERF3 probabilities

Team: Lewis Wang, Jordan Cortez, Guillermo Beas, Cynthia Tong, Katia Ascencio, Sebastien Roussouw

Mentors: Kevin Milner, Jacqui Gilchrist

    
         

 		   Machine Learning Team

Task: Investigate how machine learning can be applied to current software to forecast earthquakes

Team: Anthony Guerra, Lewis Wang, Geuillermo Beas, Jordan Wolz, Shalani Weerasooriya, Tomoe Mizutani, Ramon Mei, Varduhi Kababjyan, Brandon Ho, Tiffany Streitenberger

Mentors: Abhijit Kashyap, Jacqui Gilchrist

    
         

 		   Hazard and Risk Visualization Team

Task: Determine which multi-event scenarios are a threat to Los Angeles and illustrate their hazard and risk with visualizations, and summarize the impact of these scenarios

Team: Ashlee Trotter, Ramon Mei, Paige Given, Chrissy Polcino, Elvis Carrillo, Jordan Wolz

Mentors: Resherle Verna

    
         

Back