## Summary

California is experiencing an “earthquake drought”: no ground-rupturing earthquake has occurred during the last century on the principal faults of the San Andreas system. The 2019 Grand Challenge will focus on two main research questions: (1) Is a hiatus of 100 years or longer consistent with earthquake forecasting models? (2) What are the implications of the hiatus for future earthquake activity in California? Your investigations will focus on the following tasks: Estimate the probability of such a hiatus from two earthquake forecasting models, (a) the Uniform California Earthquake Rupture Forecast, Version 3 (UCERF3) and (b) earthquake catalogs generated by running the RSQSim rupture simulator.
Calculate both state-independent and state-dependent probabilities. Condition the state-dependent probabilities on the occurrence of large earthquakes on the northern (1906-type) and southern (1857-type) San Andreas Fault during the 50-100 years before the hiatus.
Assess the chances of large California earthquakes that might occur during the next 30 years using (i) frequencies computed directly from RSQSim catalogs and (ii) probabilities computed by applying machine-learning techniques to RSQSim catalogs.
They will also illustrate these 30-year probabilities with representative scenarios that include estimates of expected ground motions, economic losses, and human casualties.

## Challenge Statement

Estimate the probability of such a hiatus from two earthquake forecasting models, (a) the Uniform California Earthquake Rupture Forecast, Version 3 (UCERF3) and (b) earthquake catalogs generated by running the RSQSim rupture simulator. Calculate both state-independent and state-dependent probabilities. Condition the state-dependent probabilities on the occurrence of large earthquakes on the northern (1906-type) and southern (1857-type) San Andreas Fault during the 50-100 years before the hiatus. Assess the chances of large California earthquakes that might occur during the next 30 years using (i) frequencies computed directly from RSQSim catalogs and (ii) probabilities computed by applying machine-learning techniques to RSQSim catalogs. They will also illustrate these 30-year probabilities with representative scenarios that include estimates of expected ground motions, economic losses, and human casualties.

## Project Teams

 Forecasting and Simulation Team Mentors: Scott Callaghan, Kevin Milner SCEC-VDO Team Mentors: Kevin Milner, John Yu, Harsh Waghela Machine Learning Team Mentors: William Savran Hazard and Risk Visualization Team Mentors: Gabriela Noriega

Back