ResidualPDF

Plot residuals of ground-motion models (GMMs) to visualize GMM performance with period

Setting up your conda environment:

conda env create -f env_respdf.yml

The name of your new environment is defined in the first line of the .yml file (name: ), so you can edit this BEFORE creating the environment, if you want to call it something different.

Activate the respdf environment with

conda activate respdf

Remember, you will need to run the conda activate respdf command every time you open a new terminal or tab!

NOTES:

• If you have an older version of conda you may need to use source activate respdf instead.
• If you get an error message about Python not being installed as a framework, try modifying your matplotlibrc file to use a different backend. If you don't have one, you can create a very simple one with

echo "backend: TkAgg" > matplotlibrc

Residual PDF plots for individual ground motion models

To make a residual PDF for the A10 ground motion model, with residuals stored in file A10_PDF.out and sigmas in A10_sigma.out in current working directory:

python respdf.py A10

this will create a file called A10pdf.png

You can quickly regenerate plots for all of the GMMs in your directory by running the script remakeplots.sh which will loop over all *_PDF.out files in the directory.

Summary plots showing means and sigmas for many models

If you have lots of *_PDF.out and *_sigma.out files in your current working directory, you can produce figures summarizing the mean residuals with:

python resmeans.py 

This will create a heat map of the GMM means (resmeans_heat.png) and a plot of residuals and sigmas vs. period (resmeans_lines.png)