The training set consists of 2,154,021 rows and the test set of 280,961 rows. The data is provided in two formats — a NetCDF file for features and a CSV file for labels. Each record is uniquely identified by its coordinates (latitude, longitude, and date) and contains the following features:

• TWS_t — Total Water Storage for month t
• SPEI_01 to SPEI_12 — Standardised Precipitation Indices at 1, 3, 6, and 12 month timescales timescale
• SOIL_MOISTURE_t and SOIL_MOISTURE_tp1 — Near-surface soil moisture for month t and month t+1

The corresponding target values (TWS for t+1) for the training set are provided in a separate CSV file.