Hyper-local Seasonal Forecasts

This repository contains the workflow to produce seasonal weather forecasts in collaboration with Sprout Open Content . The tool uses several data sources listed below to create historical baselines (1983 - 2016) as a reference to compare with the seasonal weather forecasts from ECMWF-S51 and calculate quantiles. The tool needs to be updated on a monthly basis (6th day of each month).

Requirements:

• Docker installed
• CDS API

Deployment:

1. Pull existing image:

docker pull egbendito/eia-transform:hlsf-last

2. Set-up:

docker run -it --rm --name hlsf -v /path/to/your/volume:/media egbendito/eia-transform:hlsf-last setup-hlsf $ISO3

For the first time (or any subsequent re-deployments) you will need to set-up the tool. This downloads historical data, and creates the baseline for the country you select ($ISO3). Please, substitute $ISO3 for the 3-letter code of your country of interest (see here). Make sure you mount a persistent volume on the container to be able to persist the data. It will be mapped to the /media folder of the container.

3. Update forecasts:

docker run -it --rm --name hlsf -v /path/to/your/volume:/media/hlsf egbendito/eia-transform:hlsf-last update-hlsf

This uses the CDS API, which should be placed in the directory you are mounting to the container (/path/to/your/volume) in a file called .cdsapirc with the following structure:

url: https://cds-beta.climate.copernicus.eu/api
key: your_api_key_here

4. Results:

You will be able to retrieve all the data, including the outputs from your /path/to/your/volume/data/. There you'll find:

• inputs:

├── chirps
│   ├── 1983.nc
│   ├── ...
│   └── 2016.nc
├── chirts
│   ├── 1983.nc
│   ├── ...
│   └── 1984.nc
├── gadm
│   ├── gadm41_$ISO3.gpkg
│   └── roi.gpkg
└── s5
    ├── ecmwf_s5_rain_$YEAR_$MONTH.nc
    ├── ...
    └── ecmwf_s5_tmax_$YEAR_$MONTH.nc

• intermediate:

├── climatic
│   ├── climatic_monthly_prec.nc
│   ├── climatic_monthly_tmax.nc
│   ├── climatic_monthly_ts_prec.nc
│   └── climatic_monthly_ts_tmax.nc
└── forecast
    ├── ecmwf_s5_rain_$YEAR_$MONTH.nc
    ├── ...
    └── ecmwf_s5_tmax_$YEAR_$MONTH.nc

• outputs:

├── sprout_forecast_s5_$YEAR_$MONTH-( $MONTH + 9 ).csv
└── ...

Data Sources:

Source	URL
C3S	https://cds-beta.climate.copernicus.eu/datasets/seasonal-original-single-levels
GADM	https://gadm.org/index.html
CHIRPS	https://chc.ucsb.edu/data/chirps
CHIRTS	https://chc.ucsb.edu/data/chirtsdaily

Acknowledgements: