@article{Gauch-2020-Rainfall–Runoff,
title = "Rainfall{--}Runoff Prediction at Multiple Timescales with a SingleLong Short-Term Memory Network",
author = "Gauch, Martin and
Kratzert, Frederik and
Klotz, Daniel and
Nearing, Grey and
Lin, Jimmy and
Hochreiter, Sepp",
journal = "",
year = "2020",
publisher = "Copernicus GmbH",
url = "https://gwf-uwaterloo.github.io/gwf-publications/G20-3001",
doi = "10.5194/hess-2020-540",
abstract = "Abstract. Long Short-Term Memory Networks (LSTMs) have been applied to daily discharge prediction with remarkable success. Many practical scenarios, however, require predictions at more granular timescales. For instance, accurate prediction of short but extreme flood peaks can make a life-saving difference, yet such peaks may escape the coarse temporal resolution of daily predictions. Naively training an LSTM on hourly data, however, entails very long input sequences that make learning hard and computationally expensive. In this study, we propose two Multi-Timescale LSTM (MTS-LSTM) architectures that jointly predict multiple timescales within one model, as they process long-past inputs at a single temporal resolution and branch out into each individual timescale for more recent input steps. We test these models on 516 basins across the continental United States and benchmark against the US National Water Model. Compared to naive prediction with a distinct LSTM per timescale, the multi-timescale architectures are computationally more efficient with no loss in accuracy. Beyond prediction quality, the multi-timescale LSTM can process different input variables at different timescales, which is especially relevant to operational applications where the lead time of meteorological forcings depends on their temporal resolution.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="Gauch-2020-Rainfall–Runoff">
<titleInfo>
<title>Rainfall–Runoff Prediction at Multiple Timescales with a SingleLong Short-Term Memory Network</title>
</titleInfo>
<name type="personal">
<namePart type="given">Martin</namePart>
<namePart type="family">Gauch</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Frederik</namePart>
<namePart type="family">Kratzert</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daniel</namePart>
<namePart type="family">Klotz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Grey</namePart>
<namePart type="family">Nearing</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jimmy</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sepp</namePart>
<namePart type="family">Hochreiter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<genre authority="bibutilsgt">journal article</genre>
<relatedItem type="host">
<titleInfo>
<title/>
</titleInfo>
<originInfo>
<issuance>continuing</issuance>
<publisher>Copernicus GmbH</publisher>
</originInfo>
<genre authority="marcgt">periodical</genre>
<genre authority="bibutilsgt">academic journal</genre>
</relatedItem>
<abstract>Abstract. Long Short-Term Memory Networks (LSTMs) have been applied to daily discharge prediction with remarkable success. Many practical scenarios, however, require predictions at more granular timescales. For instance, accurate prediction of short but extreme flood peaks can make a life-saving difference, yet such peaks may escape the coarse temporal resolution of daily predictions. Naively training an LSTM on hourly data, however, entails very long input sequences that make learning hard and computationally expensive. In this study, we propose two Multi-Timescale LSTM (MTS-LSTM) architectures that jointly predict multiple timescales within one model, as they process long-past inputs at a single temporal resolution and branch out into each individual timescale for more recent input steps. We test these models on 516 basins across the continental United States and benchmark against the US National Water Model. Compared to naive prediction with a distinct LSTM per timescale, the multi-timescale architectures are computationally more efficient with no loss in accuracy. Beyond prediction quality, the multi-timescale LSTM can process different input variables at different timescales, which is especially relevant to operational applications where the lead time of meteorological forcings depends on their temporal resolution.</abstract>
<identifier type="citekey">Gauch-2020-Rainfall–Runoff</identifier>
<identifier type="doi">10.5194/hess-2020-540</identifier>
<location>
<url>https://gwf-uwaterloo.github.io/gwf-publications/G20-3001</url>
</location>
<part>
<date>2020</date>
</part>
</mods>
</modsCollection>
%0 Journal Article
%T Rainfall–Runoff Prediction at Multiple Timescales with a SingleLong Short-Term Memory Network
%A Gauch, Martin
%A Kratzert, Frederik
%A Klotz, Daniel
%A Nearing, Grey
%A Lin, Jimmy
%A Hochreiter, Sepp
%D 2020
%I Copernicus GmbH
%F Gauch-2020-Rainfall–Runoff
%X Abstract. Long Short-Term Memory Networks (LSTMs) have been applied to daily discharge prediction with remarkable success. Many practical scenarios, however, require predictions at more granular timescales. For instance, accurate prediction of short but extreme flood peaks can make a life-saving difference, yet such peaks may escape the coarse temporal resolution of daily predictions. Naively training an LSTM on hourly data, however, entails very long input sequences that make learning hard and computationally expensive. In this study, we propose two Multi-Timescale LSTM (MTS-LSTM) architectures that jointly predict multiple timescales within one model, as they process long-past inputs at a single temporal resolution and branch out into each individual timescale for more recent input steps. We test these models on 516 basins across the continental United States and benchmark against the US National Water Model. Compared to naive prediction with a distinct LSTM per timescale, the multi-timescale architectures are computationally more efficient with no loss in accuracy. Beyond prediction quality, the multi-timescale LSTM can process different input variables at different timescales, which is especially relevant to operational applications where the lead time of meteorological forcings depends on their temporal resolution.
%R 10.5194/hess-2020-540
%U https://gwf-uwaterloo.github.io/gwf-publications/G20-3001
%U https://doi.org/10.5194/hess-2020-540
Markdown (Informal)
[Rainfall–Runoff Prediction at Multiple Timescales with a SingleLong Short-Term Memory Network](https://gwf-uwaterloo.github.io/gwf-publications/G20-3001) (Gauch et al., GWF 2020)
ACL
- Martin Gauch, Frederik Kratzert, Daniel Klotz, Grey Nearing, Jimmy Lin, and Sepp Hochreiter. 2020. Rainfall–Runoff Prediction at Multiple Timescales with a SingleLong Short-Term Memory Network.