pySeasonal

A Python package to verify seasonal prediction systems against reanalysis data or observations. It also provides real-time seasonal forecasts, expressed as probabilities for a given variable to occur in the lower, centre or upper tercile of the hindcast period. For skill and tercile calculations, the hindcasts can be optionally combined with the past forecasts. The package currently works with Copernicus Data Store products. All input parameters are set at the start of the scripts and their use is commented therein.

Author: Swen Brands, [email protected] or [email protected]

The scripts have to be run sequentially in the following order:

1. regrid_obs.py #brings monthly observations / reanalysis data to the model grid and cuts out the target region (Medcof)

2. aggregate_hindcast.py #aggregates daily model data to monthly-mean data and generates a single netCDF file per variable with dimensions time x lead x member x lat x lon from the netCDF files provided by Predictia. This script also brings the netCDF variable names in the model files to ERA5 standard (e.g. tas -> t2m or pr -> tp) and performs the associated unit transformations.

3. get_skill_season.py #calculates hindcast skill measures for the selected spatial domain, period, season and lead time. Also calculates the climatological tercile thresholds needed in pred2tercile.py

4. plot_seasonal_validation_results.py #plots the results obtained with get_skill_season.py and generates a netCDF file containing binary skill masks

5. pred2tercile.py #aggregates the daily GCM data of a given foreccast to seasonal average predictions and transforms them into forecast probabilities per tercile. These are then put into netCDF format

6. functions_seasonal.py #contains the functions used by the afore- mentioned scripts

Secondary scripts:

• downloadme_era5_monthly_single_level.py #script to download monthly ERA5 data form CDS with the CDS API; for variables on single levels

• downloadme_era5_monthly_pressure_level.py #script to download monthly ERA5 data form CDS with the CDS API; for variables on pressure levels

Deprecated scripts:

• get_skill.py
• map_results.py

Credits

This ongoing research work is being funded by the Ministry for Ecological Transition and Demographic Challenge (MITECO) and the European Commission NextGenerationEU (Regulation EU 2020/2094), through CSIC's Interdisciplinary Thematic Platform Clima (PTI-Clima).