05. Important notes
Some important note:
-
MCMC sampling is strongly influenced by the data weighting. If
you want to change the weighting for a single station, simply open
{Project}/data/{sta}_data/in.data_{sta} and edit it as a text
file. If you want to apply the change to all stations, update the
parameters in
SetupData/parameter.py, then rerun the workflow from01_prepare_data.shto02_do_MCMC.shto fully reprocess the inversion.
- When processing the MCMC workflow, it mainly runs in two steps. For each step, checkpoint messages are printed for tracking. The workflow proceeds from checkpoint 1 (MCMC sampling) to checkpoint 2 (post-process). Sometimes checkpoint 1 passes but checkpoint 2 fails. In that case, check the printed output to identify which part caused the error.
-
MCMC sampling is time comsuming. Incase you only need to change
the posterior distribution to see difference in final model. Then
change only parameter in {Project}/data/{sta}_data/in.connector
and call script to call the script
04_re_do_MCMC_post_process.shto do post process only.
-
If you want to change the plot style only, change the parameter
in {Project}/data/{sta}_data/in.connector and call scrip
05_do_replot_MCMC.sh.