Template for internal Assessment of bias correction methods

[dingaaling]

We agreed with our Met Office colleagues the metrics to be used to assess bias correction methods (there's a summary of that here).

• MVP: 1 method
• Stretch Goal: 3 methods

I will put together an example notebook of what to be applied. Once that's done the tasks would be:

• Rewrite in python (be interesting to compare assessment metrics across R and python)
• Incorporate python data into existing R markdown by Ruth (here: https://github.com/alan-turing-institute/clim-recal/blob/Update.Rpipeline/notebooks/Assessing_bc_data/Assessing_BC_Data.RMD). Previously in issue Incorporate Python pipeline output debiased data to assessment notebook #60
• Run for debiased data (separately for each method or package) for the 3 cities
• Check R dependencies management with @RuthBowyer
• Set up configurations for running the R notebookes
• Apply the specification in Decide the cases we want to showcase/make available for playing around with at the workshop #90
• Provide users options to specify what they want to look at

Note: Don't worry about caching results

Tess urgent for December workshop:

• There are further considerations about the variable specific metrics to use (eg how to calc drought, and precip)

Following workshop (note before any of this can be done, need to re=run the

• Add trend figure
• Metrics should be calculated by rankings the variables rather than day-to-day matching

[dingaaling]

Next steps:

• Upload HTML to Github to share first notebook draft

[griff-rees]

Would it make sense to tie this to #97 ?