Climatologies and climatological anomalies have been calculated at CEDA from CMIP5 model output.

Time periods:
- periods are defined as being:
- Relative to 2000 for future scenarios (RCPy.x)
- Relative to 1900 for historical run
- Relative to the beginning of the experiment for the piControl and 1pctCO2 runs

- Twenty year averaging periods: +20 to +39, +46 to +65, +80 to +99, +180 to +199

Anomaly specification:
All values are taken as anomalies from the 1986->2005 mean (base line values).
For the Historical and RCP simulations these base line values are taken from the Historical experiment.
For the 1pctCO2 and piControl simulations the base line values are taken from the last 20 years of the piControl experiment.

Output:
- Period mean of the monthly mean of the variables
- 12 monthly values per climatology
- Climatology is a 2D field
- One climatology per model per scenario per variable per time period
- CF compliant netCDF files

Directory Structure:
The output directory structure follows the DRS specification with the root directory being /derived.
Each file has the extended attribute: clim-anom-ref5. This indicates that it is a climatological mean anomaly (clim-anom) and that the reference period matches the AR5 reference period (ref5).

Method:
The algorithm follows the process below:

- For every model:
- For every variable (if available):
calculate the 1986->2005 climatological (period average) monthly means (baseline) for the Historical experiment

- For every model:
- For every variable (if available):
calculate the period average monthly means (baseline) of the last 20 years of the piControl run

- For every model:
- For every experiment:
- For every time period:
- For every variable (if available):
calculate the climatological monthly mean for this (model, experiment, time period, variable) and subtract the relevant baseline

Code:
https://github.com/cedadev/ipcc_ddc_cmip5_clim_anoms/releases/tag/1.0