A glimpse into the Unobserved: Runoff simulation for ungauged catchments with LSTMs

NeurIPS 2018 workshop paper, showing first results on using LSTMs for prediction in ungauged basins.

Abstract

Runoff predictions of a river from meteorological inputs is a key task in the field of hydrology. However, current hydrological models require a substantial amount of parameter tuning on basis of historical records. If no historical runoff observation sare available it is very challenging to produce good predictions. In this study we explore the capability of LSTMs for simulating the runoff for these ungauged cases. A single LSTM is trained to learn a general hydrological model from hundreds of catchments throughout the contiguous United States of America and evaluated against catchments not used during training. Our results suggest that LSTMs a) are able to learn a general hydrological model and b) in the majority of catchments outperform an established hydrological model, which was especially trained for these catchments.

Paper

Kratzert, F., Klotz, D., Herrnegger, M. and Hochreiter, S.: A glimpse into the Unobserved: Runoff simulation for ungauged catchments with LSTMs, Workshop on Modeling and Decision-Making in the Spatiotemporal Domain, 32nd Conference on NeuralInformation Processing Systems (NeuRIPS 2018), Montréal, Canada, 2018

Citation

@inproceedings{kratzert2018glimpse,
  title={A glimpse into the Unobserved: Runoff simulation for ungauged catchments with LSTMs},
  author={Kratzert, Frederik and Klotz, Daniel and Herrnegger, Mathew and Hochreiter, Sepp},
  booktitle={Workshop on Modeling and Decision-Making in the Spatiotemporal Domain, 32nd Conference on Neural Information Processing Systems (NeuRIPS 2018)},
  venue={Montréal, Canada},
  date={2-8 Dec},
  year={2018}
}