Dataset: Rising global riverine deoxygenation rates and GHG emissions driven by the synergistic effects of warming and anthropogenic land use expansion

This dataset was generated using random forest models driven by remote sensing observations. It provides modeled annual means of riverine greenhouse gas (GHG) saturations and associated water quality parameters across 5,084 globally distributed catchments from 2002 to 2022. Catchment IDs are similar to those from level 12 of the hydrobasins database (HydroBASINS).

The dataset includes both the compiled training and validation field data used to develop the random forest models, as well as the modeled outputs (provided as Excel files). For the modeled outputs, three files are available: one containing only the modeled variables, and two additional files that incorporate ancillary data, including upstream land use (forest, cropland, and urban cover, %) and mean annual precipitation.

Further details on data generation, model assumptions, and limitations are provided in the associated publication. For additional information, please contact the corresponding author.