WaterBench: A Large-scale Benchmark Dataset for Data-Driven Streamflow Forecasting
- 1Department of Civil and Environmental Engineering, University of Iowa, Iowa City, 52246 USA
- 2Department of Electrical and Computer Engineering, University of Iowa, Iowa City, 52246 USA
- 3Interdisciplinary Graduate Program in Informatics, University of Iowa, Iowa City, 52246 USA
- 1Department of Civil and Environmental Engineering, University of Iowa, Iowa City, 52246 USA
- 2Department of Electrical and Computer Engineering, University of Iowa, Iowa City, 52246 USA
- 3Interdisciplinary Graduate Program in Informatics, University of Iowa, Iowa City, 52246 USA
Abstract. This study proposes a comprehensive benchmark dataset for streamflow forecasting, WaterBench, that follows FAIR data principles that is prepared with a focus on convenience for utilizing in data-driven and machine learning studies, and provides benchmark performance for state-of-art deep learning architectures on the dataset for comparative analysis. By aggregating the datasets of streamflow, precipitation, watershed area, slope, soil types, and evapotranspiration from federal agencies and state organizations (i.e., NASA, NOAA, USGS, and Iowa Flood Center), we provided the WaterBench for hourly streamflow forecast studies. This dataset has a high temporal and spatial resolution with rich metadata and relational information, which can be used for varieties of deep learning and machine learning research. We defined a sample streamflow forecasting task for the next 120 hours and provided performance benchmarks on this task with sample linear regression and deep learning models, including Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and S2S (Sequence-to-sequence). To some extent, WaterBench makes up for the lack of unified benchmarks in earth science research. We highly encourage researchers to use the WaterBench for deep learning research in hydrology.
Ibrahim Demir et al.
Status: open (extended)
-
CC1: 'Include mention of Iowa in title and abstract', Mathis Messager, 03 Mar 2022
reply
Brief comment: I applaud the authors for putting together this useful dataset and manuscript.
Without going into depth with its content, I highly encourage the authors to mention in the title and abstract that this dataset is entirely within Iowa. This would help readers understand its content from the get-go. In addition, this would allow for future WaterBench versions in other regions (just like CAMEL; this could be renamed WaterBench-Iowa or WaterBench-IA).
Best wishes.
-
RC1: 'Comment on essd-2022-52', Anonymous Referee #1, 06 Apr 2022
reply
General comment
This paper describes a dataset with time series of 125 monitored catchments and some examples of how it can be used to evaluate forecasting models. Science is a cumulative effort and this dataset fills a gap in the publicly available to hydrology and AI community. The dataset itself is open and in a good condition. It would be nice to have the watershed boundaries and coordinates of each measuring point as shapefiles available together with the data. However, this is not a major concern since they are available from Iowa Flood Center.
I do not see any major flaws In this paper and recommend that it gets accepted.
Specific comments Individual scientific questions/issues
Open data is important to the scientific community and I welcome this contribution by the authors.
Technical corrections
I am a non-native English speaker and struggled with this sentence on row 25 and it could be rephrased for clarity. This is the sentence in question:
“The power of deep learning in problem-solving has opened ways to advancements in many fields that machine learning has been a go-to solution for predictive modeling, such as image recognition and synthesis (Demiray et al., 2021), speech recognition, language modeling and time-series prediction.”
Ibrahim Demir et al.
Data sets
WaterBench Ibrahim Demir, Zhongrun Xiang, Bekir Demiray, Muhammed Sit https://github.com/uihilab/WaterBench
Model code and software
WaterBench Ibrahim Demir, Zhongrun Xiang, Bekir Demiray, Muhammed Sit https://github.com/uihilab/WaterBench/tree/main/examples
Ibrahim Demir et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
339 | 99 | 12 | 450 | 5 | 7 |
- HTML: 339
- PDF: 99
- XML: 12
- Total: 450
- BibTeX: 5
- EndNote: 7
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1