The Coordinated Regional Climate Downscaling Experiment (CORDEX) has helped paved the way for studying climate change at regional and local scales. Systematic model evaluation studies across the 14 domains (each with multiple simulations and many variables) have been limited, with differences in data publication requirements (e.g. ESGF vs local storage/ftp) and archiving standards (e.g. use of CF-compliant netCDF)

Project Goals

  1. Make it easy to perform systematic evaluations with CORDEX domains, models and variables
  2. Leverage NASA satellite observational datasets in evaluations
  3. Increase communication and collaboration between CORDEX domains



Why we need “Systematic Evaluation”

This Config file (namelist file) is necessary to run each evaluation combination (CORDEX Domain, Season and Variable), forming a large “evaluation matrix”.

14 variables x 13 domains x 3 seasons x ~10 models > 5000 evaluations. Writing that many Config files manually would be cumbersome/prohibitive.




Solution: Extract metadata from input filenames

We can group all models and obs datasets together by common attributes (domain and variable) to form a unique evaluation, and therefore automatically generate Config Files using only the dataset locations as user input.
















Analysis Plan / Future Work

  1. Initial Evaluations will focus on domains with the most simulations on the ESGF (present)
    • Publish these results on RCMES website (early 2018)
  2. Inform CORDEX POI’s of result publication
    • Include analysis and configuration files for 3D Temperature and Humidity fields from AIRS.
  3. Analyze and post remaining evaluations with simulations (not on ESGF but) made available to us (mid 2018)