===== inv-conj-geophy: Joint Inversions in Geophysics =====

Website: http://inv-conj-geophy.sciencesconf.org


==== Lecturers ====

  * Anthony Sladen, CNRS and University of Nice Sophia-Antipolis
  * Henriette Sudhaus, University of Potsdam
  * Zacharie Duputel, CNRS and University of Strasbourg


==== Schedule ====

  * Tuesday 9:00-10:30
  * Tuesday 11:00-12:30
  * Tuesday 16:00-17:30
  * Wednesday 14:00-15:30


==== Program ====

Introduction (Zacharie)

S1: Geodesy data - Tuesday 9:00-10:30 (Henriette)
      * What are GPS, InSAR, tsunami data
      * Where to get them
      * How to process them (providing a list of tools, websites...)
      * How to predict them (providing a list of tools, websites...)

S2: Implement/test different optimization techniques - Tuesday 11:00-12:30 (Henriette and Anthony)
      * linearized
      * stochastic
      * regularization
      * resolution tests
      * strategies to combine different data sets
      * Hands-on: Kleifarvatn, 2000 Icelandic earthquake 
          * **Part of S2, or just example on data pred./proc.?** (Anthony) -> **I moved it to S2** (Zacharie) 

S3: Seismology(/tsunami?) - Tuesday 16:00-17:30 (Anthony)
      * Where to get data (providing a list of tools, websites...)
      * How to process (providing a list of tools, websites...)
      * Theory behind source inversion 
      * Brief extension to slow-slip events?

S4: Bayesian source inversion - Wednesday 14:00 - 15:30 (Zacharie)
      * choices of prior distribution (for static and kinematic parameters), 
      * stochastic forward model (geodetic and seismic data), 
      * solving the source inversion problem (MCMC, transitional MCMC, hierarchical modelling, model class selection, BIC/ABIC, etc).
      * Simple toy problem to illustrate MCMC
==== Questions/Ideas ====

   * ** Virtual machines? ** 
     * Alessia says it should be possible, files can be uploaded on the website
     * We should discuss with Max Moorkamp
   * ** Content of theoretical lectures? **
     * Tarantola formalism 
     * deterministic inversion 
     * adjoint-based inversion