Data Download & Management

Section under Construction The WAS_Download module provides a unified interface to retrieve climate data required for seasonal forecasting. It handles authentication (CDS API), protocol management (HTTP/FTP), file format conversion (TIFF/GRIB to NetCDF), and spatiotemporal standardization.

Three types of data can be downloaded: 1. GCM Data: Seasonal forecasts (Hindcasts and Real-time). 2. Reanalysis: Historical baselines (ERA5, ERSST). 3. Observations: Satellite-based products (CHIRPS, TAMSAT, AgERA5).

Prerequisites: For C3S data (ECMWF, UKMO, Météo-France, etc.), you must have a .cdsapirc file configured in your home directory. See CDS API How-to.

1. Seasonal GCM Forecasts

This section handles data from the Copernicus Climate Change Service (C3S) and the North American Multi-Model Ensemble (NMME).

Monthly GCM Data

Method: WAS_Download_Models

Downloads monthly mean hindcasts or forecasts. It automatically handles the differences between C3S (NetCDF via API) and NMME (NetCDF/CPT via FTP).

Parameters:

dir_to_save (str): Target directory.
center_variable (list): Format "CENTER_SYSTEM.VAR".
- Centers: ECMWF_51, UKMO_604, METEOFRANCE_8, NCEP_2, CMCC_35, DWD_21, JMA_3, ECCC_4.
- NMME: CFSV2_1, CMC1_1, CMC2_1, GFDL_1, NASA_1, NCAR_CCSM4_1.
- Variables: PRCP, TEMP, TMAX, TMIN, SST, SLP, UGRD10, VGRD10.
month_of_initialization (int): 1-12.
lead_time (list): List of lead months (e.g., ['01', '02', '03']).
year_forecast (int, optional): If provided, downloads real-time forecast. If None, downloads hindcasts.

from wass2s import WAS_Download

downloader = WAS_Download()

# Download Hindcasts (1993-2016) for ECMWF and NCEP Precipitation
downloader.WAS_Download_Models(
    dir_to_save="./data/GCM",
    center_variable=["ECMWF_51.PRCP", "NCEP_2.PRCP"],
    month_of_initialization=5,  # May starts
    lead_time=["01", "02", "03"], # Jun, Jul, Aug
    year_start_hindcast=1993,
    year_end_hindcast=2016,
    area=[20, -20, 0, 10]  # [N, W, S, E]
)

Daily GCM Data

Method: WAS_Download_Models_Daily

Downloads daily or sub-daily data (e.g., for heatwave or dry spell analysis). Note that daily data is voluminous.

Parameters: Adds day_of_initialization and uses leadtime_hour (e.g., “24”, “48”…) instead of months.

# Download Daily Forecast for specific initialization
downloader.WAS_Download_Models_Daily(
    dir_to_save="./data/GCM_Daily",
    center_variable=["ECMWF_51.TMAX"],
    month_of_initialization=5,
    day_of_initialization=1,
    leadtime_hour=["24", "48", "72", "96"], # First 4 days
    year_start_hindcast=2000,
    year_end_hindcast=2020,
    area=[20, -20, 0, 10]
)

2. Reanalysis Data

Used for model calibration and verification. Supports ERA5 (Atmosphere), ERA5-Land (Surface), and NOAA ERSST (Ocean).

ERA5 & ERSST

Method: WAS_Download_Reanalysis

Downloads monthly means. It handles cross-year seasons (e.g., DJF) by downloading appropriate months from adjacent years and aggregating them.

ERA5: Downloads from CDS.
NOAA ERSST: Downloads V5/V6 from NCEI/IRIDL.

Usage Example:

# Download SST for Nino 3.4 calculation
downloader.WAS_Download_Reanalysis(
    dir_to_save="./data/Reanalysis",
    center_variable=["NOAA.SST"],
    year_start=1981,
    year_end=2020,
    area=[5, -170, -5, -120], # Pacific box
    seas=["11", "12", "01"],  # NDJ Season
    force_download=False
)

ERA5-Land

Method: WAS_Download_ERA5Land & WAS_Download_ERA5Land_daily

Higher resolution (9km) land data, ideal for hydrological applications (Runoff, Soil Moisture).

# Download monthly ERA5-Land Soil Moisture
downloader.WAS_Download_ERA5Land(
    dir_to_save="./data/Reanalysis",
    center_variable=["ERA5Land.SOILWATER1"],
    year_start=1981,
    year_end=2020,
    area=[15, -18, 4, 10], # West Africa
    seas=["06", "07", "08"] # JJA
)

3. Observational Data (Satellite)

High-resolution gridded observations for calibration and verification.

Agro-Climatic Indicators (AgERA5)

Method: WAS_Download_AgroIndicators

Derived from ERA5, corrected against stations. Good for temperature and general indices.

Variables: AGRO.PRCP, AGRO.TMAX, AGRO.TMIN, AGRO.DSWR (Solar Radiation), AGRO.ETP (Evapotranspiration).

CHIRPS Precipitation

Method: WAS_Download_CHIRPSv3_Seasonal & _Daily

Downloads high-resolution (0.05°) precipitation data from the Climate Hazards Group. * Fetches TIFF files from UCSB servers. * Merges, reprojects, and saves as NetCDF.

# Download Seasonal (aggregated) CHIRPS
downloader.WAS_Download_CHIRPSv3_Seasonal(
    dir_to_save="./data/Obs",
    variables=["PRCP"],
    year_start=1981,
    year_end=2020,
    region="africa",
    season_months=["06", "07", "08", "09"], # JJAS
    area=[20, -20, 0, 15]
)

TAMSAT Precipitation

Method: WAS_Download_TAMSAT_Seasonal & _Daily

Downloads rainfall estimates (RFE) or Soil Moisture from the University of Reading (TAMSAT). * Product: rfe (Rainfall) or soil_moisture.

# Download Daily TAMSAT Rainfall
downloader.WAS_Download_TAMSAT_Daily(
    dir_to_save="./data/Obs",
    product="rfe",
    year_start=2023,
    year_end=2023,
    area=[20, -20, 0, 15]
)