Published February 6, 2026 | Version 0.1
Dataset Open

Ancillary dataset for Global River BankFull Discharge (GQBF)

  • 1. ROR icon University of Oxford
  • 2. ROR icon Loughborough University
  • 3. University of oxford
  • 4. ROR icon European Centre for Medium-Range Weather Forecasts

Contributors

Research group:

Researcher:

Description

This is a collection of the ancillary data of the Global River Bankfull Discharge (GQBF). It consists of 6 columns, [domain, reach_id, cv, darea_flag, grwl_width_flag, mapped].

This data should be matched with the GQBF dataset with [domain, reach_id]. 

The cv (coefficient of variation) indicates the relative uncertainty in the global estimation of bankfull river discharge (GQBF). The CV represents the variability among the tree-level estimations produced by the deployed random forest model and is calculated as the ratio of the standard deviation to the mean value across all tree estimations.

The darea_flag, grwl_width_flag indicate whether the drainage area, width of the river reach is out-of-distribution of the training samples of the bankfull discharge, respectively.

The mapped column flags river reaches as 0 when they either have width narrower than 30 m (or their overlap with the GRWL_WIDTH <0.5), or their estimated bankfull discharge is null, otherwise it is 1.  Note, the out-of-distribution river reaches are not considered as a rule to calculate this field. Those out-of-distribution river reaches should be identified with the darea_flag, grwl_width_flag.

Change log

  • v0.1 - 2024-09-29
    • First globally complete dataset published
  • v0.1 - 2024-11-19
    • Add vector files for regions SI, SP
  • v0.1 - 2025-10-27
    • Add the coefficient of variation of bankfull discharge estimations
  • v0.1 - 2026-02-06
    • Add compiled ancillary data for bankfull discharge estimations 

Files

GLOBAL_reach_QBFestuncertainty_FLAGout-of-distribution.csv

Files (528.4 MB)

Additional details

Related works

Is published in
Preprint: 10.21203/rs.3.rs-5185659/v1 (DOI)
References
Preprint: 10.1029/2024WR038308 (DOI)

Funding

UK Research and Innovation
THE EVOLUTION OF GLOBAL FLOOD HAZARD AND RISK [EVOFLOOD] NE/S015728/1

Dates

Available
2025-10-27

Software

Repository URL
https://github.com/YinxueLiu/G-QBF
Programming language
R , Python
Development Status
Active