CSEP2

Roadmap for CSEP2 and understood by me. This document contains my notes and progress on a collection of CSEP related things. This represents my unofficial sounding board.

Legend

Bold text: High Priority
Italic text: Medium Priority
Normal text: Low Priority
Strikethrough: Complete

CSEP2 Roadmap

Milestones

1. paper on testing stochastic event sets
   • ucerf3 simulations in Landers sequence
   • ucerf3 simulations for other large eqs
   • implement n-test, m-test, s-test for stochastic event sets
   • determine power of statistical tests using simulation based approach
   • run csep1 ETAS simulations in csep2 experiment
2. paper on testability of earthquake forecasts that involve faults
   • needs thought!
3. release csep2 testing center distribution
   • write docs for csep2 github
   • implement methods to run experiments (db, jobs to hpc, post-processing)

Upcoming Conferences

1. EGU 2019 (Abstracts due: 01/11/19)
2. SSA 2019 (Abstracts due: 01/11/19)
3. Statistical Seismology 2019 (Abstracts due: Unknown)

Project Overview

Testing Stochastic Event Sets

• Read Gneiting work on proper scoring.
• Cramer-von Mises test
• Kolmogorov Smirnov test

Testability of Earthquake Forecasts on Faults

• power of evaluations
• how long will we have to perform an experiement so that we

CSEP2 Testing Center Release

• backend logic to store model data/parameters
• store catalogs
• store results
• store generated forecasts

General To-do

• maximum likelihood estimation for b-values
• m-test and s-test
• documentation on csep (start with tools to run ucerf3-etas)
• build data model for csep2 runs
• revisit prototypes of end-to-end system; need more runs done before april.