@article{Clark-Vogel-2021-The-Abuse,
title = "The Abuse of Popular Performance Metrics in Hydrologic Modeling",
author = "Clark, Martyn and
Vogel, Richard M. and
Lamontagne, Jonathan and
Mizukami, Naoki and
Knoben, Wouter and
Tang, Guoqiang and
Gharari, Shervan and
Freer, Jim and
Whitfield, Paul H. and
Shook, Kevin and
Papalexiou, Simon Michael and
Clark, Martyn and
Vogel, Richard M. and
Lamontagne, Jonathan and
Mizukami, Naoki and
Knoben, Wouter and
Tang, Guoqiang and
Gharari, Shervan and
Freer, Jim and
Whitfield, Paul H. and
Shook, Kevin and
Papalexiou, Simon Michael",
journal = "Water Resources Research, Volume 57, Issue 9",
volume = "57",
number = "9",
year = "2021",
publisher = "American Geophysical Union (AGU)",
url = "https://gwf-uwaterloo.github.io/gwf-publications/G21-57001",
doi = "10.1029/2020wr029001",
abstract = "The goal of this commentary is to critically evaluate the use of popular performance metrics in hydrologic modeling. We focus on the Nash-Sutcliffe Efficiency (NSE) and the Kling-Gupta Efficiency (KGE) metrics, which are both widely used in hydrologic research and practice around the world. Our specific objectives are: (a) to provide tools that quantify the sampling uncertainty in popular performance metrics; (b) to quantify sampling uncertainty in popular performance metrics across a large sample of catchments; and (c) to prescribe the further research that is, needed to improve the estimation, interpretation, and use of popular performance metrics in hydrologic modeling. Our large-sample analysis demonstrates that there is substantial sampling uncertainty in the NSE and KGE estimators. This occurs because the probability distribution of squared errors between model simulations and observations has heavy tails, meaning that performance metrics can be heavily influenced by just a few data points. Our results highlight obvious (yet ignored) abuses of performance metrics that contaminate the conclusions of many hydrologic modeling studies: It is essential to quantify the sampling uncertainty in performance metrics when justifying the use of a model for a specific purpose and when comparing the performance of competing models.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="Clark-Vogel-2021-The-Abuse">
<titleInfo>
<title>The Abuse of Popular Performance Metrics in Hydrologic Modeling</title>
</titleInfo>
<name type="personal">
<namePart type="given">Martyn</namePart>
<namePart type="family">Clark</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Richard</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Vogel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jonathan</namePart>
<namePart type="family">Lamontagne</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoki</namePart>
<namePart type="family">Mizukami</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wouter</namePart>
<namePart type="family">Knoben</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guoqiang</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shervan</namePart>
<namePart type="family">Gharari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jim</namePart>
<namePart type="family">Freer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="given">H</namePart>
<namePart type="family">Whitfield</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Shook</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="given">Michael</namePart>
<namePart type="family">Papalexiou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<genre authority="bibutilsgt">journal article</genre>
<relatedItem type="host">
<titleInfo>
<title>Water Resources Research, Volume 57, Issue 9</title>
</titleInfo>
<originInfo>
<issuance>continuing</issuance>
<publisher>American Geophysical Union (AGU)</publisher>
</originInfo>
<genre authority="marcgt">periodical</genre>
<genre authority="bibutilsgt">academic journal</genre>
</relatedItem>
<abstract>The goal of this commentary is to critically evaluate the use of popular performance metrics in hydrologic modeling. We focus on the Nash-Sutcliffe Efficiency (NSE) and the Kling-Gupta Efficiency (KGE) metrics, which are both widely used in hydrologic research and practice around the world. Our specific objectives are: (a) to provide tools that quantify the sampling uncertainty in popular performance metrics; (b) to quantify sampling uncertainty in popular performance metrics across a large sample of catchments; and (c) to prescribe the further research that is, needed to improve the estimation, interpretation, and use of popular performance metrics in hydrologic modeling. Our large-sample analysis demonstrates that there is substantial sampling uncertainty in the NSE and KGE estimators. This occurs because the probability distribution of squared errors between model simulations and observations has heavy tails, meaning that performance metrics can be heavily influenced by just a few data points. Our results highlight obvious (yet ignored) abuses of performance metrics that contaminate the conclusions of many hydrologic modeling studies: It is essential to quantify the sampling uncertainty in performance metrics when justifying the use of a model for a specific purpose and when comparing the performance of competing models.</abstract>
<identifier type="citekey">Clark-Vogel-2021-The-Abuse</identifier>
<identifier type="doi">10.1029/2020wr029001</identifier>
<location>
<url>https://gwf-uwaterloo.github.io/gwf-publications/G21-57001</url>
</location>
<part>
<date>2021</date>
<detail type="volume"><number>57</number></detail>
<detail type="issue"><number>9</number></detail>
</part>
</mods>
</modsCollection>
%0 Journal Article
%T The Abuse of Popular Performance Metrics in Hydrologic Modeling
%A Clark, Martyn
%A Vogel, Richard M.
%A Lamontagne, Jonathan
%A Mizukami, Naoki
%A Knoben, Wouter
%A Tang, Guoqiang
%A Gharari, Shervan
%A Freer, Jim
%A Whitfield, Paul H.
%A Shook, Kevin
%A Papalexiou, Simon Michael
%J Water Resources Research, Volume 57, Issue 9
%D 2021
%V 57
%N 9
%I American Geophysical Union (AGU)
%F Clark-Vogel-2021-The-Abuse
%X The goal of this commentary is to critically evaluate the use of popular performance metrics in hydrologic modeling. We focus on the Nash-Sutcliffe Efficiency (NSE) and the Kling-Gupta Efficiency (KGE) metrics, which are both widely used in hydrologic research and practice around the world. Our specific objectives are: (a) to provide tools that quantify the sampling uncertainty in popular performance metrics; (b) to quantify sampling uncertainty in popular performance metrics across a large sample of catchments; and (c) to prescribe the further research that is, needed to improve the estimation, interpretation, and use of popular performance metrics in hydrologic modeling. Our large-sample analysis demonstrates that there is substantial sampling uncertainty in the NSE and KGE estimators. This occurs because the probability distribution of squared errors between model simulations and observations has heavy tails, meaning that performance metrics can be heavily influenced by just a few data points. Our results highlight obvious (yet ignored) abuses of performance metrics that contaminate the conclusions of many hydrologic modeling studies: It is essential to quantify the sampling uncertainty in performance metrics when justifying the use of a model for a specific purpose and when comparing the performance of competing models.
%R 10.1029/2020wr029001
%U https://gwf-uwaterloo.github.io/gwf-publications/G21-57001
%U https://doi.org/10.1029/2020wr029001
Markdown (Informal)
[The Abuse of Popular Performance Metrics in Hydrologic Modeling](https://gwf-uwaterloo.github.io/gwf-publications/G21-57001) (Clark et al., GWF 2021)
ACL
- Martyn Clark, Richard M. Vogel, Jonathan Lamontagne, Naoki Mizukami, Wouter Knoben, Guoqiang Tang, Shervan Gharari, Jim Freer, Paul H. Whitfield, Kevin Shook, Simon Michael Papalexiou, Martyn Clark, Richard M. Vogel, Jonathan Lamontagne, Naoki Mizukami, Wouter Knoben, Guoqiang Tang, Shervan Gharari, Jim Freer, et al.. 2021. The Abuse of Popular Performance Metrics in Hydrologic Modeling. Water Resources Research, Volume 57, Issue 9, 57(9).
