Code and data to "Statistical learning and topkriging improve spatio-temporal low-flow estimation"

This data and software supports the manuscript "Statistical learning and topkriging improve spatio-temporal low-flow estimation" (https:://doi.org/10.1029/2024WR038329).

The dataset consists of:

• all produced predictions of the models (data/predictions.RDS and data/predictions_csv/*)
• observational data (data/observations.csv)
• additional catchment data (data/catchment_data.csv) used for presenting the figures
• state boundaries of Austria as a shape file (data/boundaries.*)
• partial predictions of a model-based boosting approach (data/partial_predictions.csv)
• Example output of number of EOF, due to long computational time (data/number_eofs.RDS)
• IDs of near natural catchments (data/ids_low_flow.csv)

Additionally, the code is provided to:

• Compute the number of EOFs (functions/number_eofs.R)
• Produce all the figures and tables in the paper (scripts/plotting_results.R)