GDNat: a global, daily, high-resolution, natural-forcing only temperature data set for attribution research

The Quantile Delta Mapping (QDM) bias adjustment method is used to compute quantile-based adjustment factors from GCM historical and historical-nat simulations of surface temperature variables from CMIP6. These adjustment factors are applied to ERA5 surface temperature variables to produce an ensemble of ERA5 simulations without anthropogenic forcing.

Currently daily average surface temperature (tas) is completed. tasmin and tasmax are planned.

Code

• notebooks
  • Data_retrieval: notebooks to retrieve raw GCM data from ESGF and ERA5 data from Google public datasets
  • Dataset_generation: notebooks to produce the GDNat dataset. Steps include 1. cleaning, 2. training, 3. regridding, 4. adjustment
  • Figures_and_diagnostics: notebooks that produce diagnostic figures and figures for the upcoming manuscript
  • Prototyping: method prototyping and proof-of-concept
• resources: includes spreadsheet of available GCM simulations