Model inputs for machine learning models forecasting streamflow drought at gaged and ungaged locations across the conterminous United States

This dataset contains 4,952 feather files with time series of all model inputs for machine learning models predicting streamflow drought across the conterminous United States (CONUS). These data are designed to support both gaged model training and ungaged prediction through the inclusion of WaterWatch gridded runoff estimates. Files contain weekly time series of streamflow percentiles, meteorology, snow water equivalent, forecast meteorology, estimated water use, soil moisture, WaterWatch runoff estimates, as well as lagged versions of these datasets. Values in these files were assembled from existing published datasets as explained in the data quality and processing steps sections.