Tutorial from TIDES advanced training school in Sesimbra, 2017
==============================================================
Both parts of the tutorial are contained in the Virtual Box. The Virtual Box creates a "guest operating system" on your computer, which allows you to use the packages and codes we used for the tutorial without further installation. If you intend to use the virtual box to go through the tutorial, you have to download and install a virtual box manager at https://www.virtualbox.org. See also practical setup session slides.
However, should you just want to explore the codes from the tutorial themselves without using Virtual Box, they can be obtained from github:
https://github.com/lermert/ants_2.git (clone the master branch)
https://github.com/sagerk/fd2d_noise.git (clone the Teaching branch with git clone --branch Teaching ) (or find the code in the zip file for fd2d_noise).
Part 1.
In part 1 of the tutorial, you will find a step-by-step tour from continuous noise data to a map of noise sources using python.
You can experiment with different preprocessing techniques (e.g. one-bit, spectral whitening) and observe how this changes the resulting correlations.
The next step illustrates how you project a measurement of cross-correlation asymmetry back towards the noise sources. This is done with simplified ray-theoretical kernels which you can compare to the corresponding finite frequency kernels. The finite frequency kernels will be discussed in more detail in part 2.
In the final step, you can apply the mapping introduced in the previous step to a large selection of global cross-correlations to reveal the seasonal sources of the Earth's hum.
Part 2.
Part 2 of the tutorial covers the modeling of synthetic correlation functions that accounts for the distribution of noise sources, heterogeneous Earth structure and seismic wave propagation physics. In addition, finite-frequency kernels for the source distribution and for Earth structure are computed.
The tutorial uses the software package fd2d_noise. It is based on adjoint techniques and on a 2D finite-difference discretization of the acoustic wave equation as an analogue for fundamental-mode surface wave propagation.
Requirements: MATLAB or Octave