/WaterBench

A Large-scale Benchmark Dataset for Data-Driven Streamflow Forecasting

MIT LicenseMIT

WaterBench

A Large-scale Benchmark Dataset for Data-Driven Streamflow Forecasting

Welcome to the main repository of the WaterBench, a large-scale benchmark dataset for streamflow forecasting. The dataset covers the 125 catchments in Iowa, US with seven different features, including precipitation, streamflow rate and ET with available data from October 1st, 2011 (the water year 2012) to September 30th, 2018 (the water year 2018).

Sample Codes

In the examples, four different models used in our paper is provided with ready to run notebooks.

Download the Dataset

The original files of the dataset can be downloaded from here as an archive file. The metadata is provided in the archive file.

The files that we used in sample codes can be downloaded from here as an archive file. The metadata is provided in the archive file.

Citation

If you use the specification or flood event inventory please cite the following publication:
Zhongrun Xiang, Bekir Z Demiray, Muhammet Sit and Ibrahim Demir. (2021). WaterBench: A Large-scale Benchmark Dataset for Data-Driven Streamflow Forecasting. EarthArXiv. DOI: https://doi.org/10.31223/X57W6T

Feedback

Feel free to send us feedback by filing an issue.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

This project is developed by the University of Iowa Hydroinformatics Lab (UIHI Lab):
https://hydroinformatics.uiowa.edu/.