Evaluation of methods for selecting climate models to simulate future hydrological change

Andrew C. Ross, Raymond G. Najjar

Research output: Contribution to journalArticle

Abstract

A challenge for climate impact studies is the selection of a limited number of climate model projections among the dozens that are typically available. Here, we examine the impacts of methods for climate model selection on projections of runoff change for five different watersheds across the conterminous USA. The results from an ensemble of 29 global climate models and 29 corresponding hydrological model simulations are compared with the results that would have been obtained by applying six different selection methods to the climate model data and using only the selected models to drive the hydrological model. We evaluate each selection method based on whether the runoff projections produced by the method meet the method’s objective and on whether the results are sensitive to the number of models chosen. The Katsavounidis–Kuo–Zhang (KKZ) method, which is intended to maximize the spread in the projected climate change, was the only method that met both criteria for suitability. Although the KKZ method generally performed well, the results from both it and the other methods varied somewhat unpredictably based on region and number of models chosen. This study shows that the methods and models used in similar top–down studies should be carefully chosen and that the results obtained should be interpreted with caution.

Original languageEnglish (US)
Pages (from-to)407-428
Number of pages22
JournalClimatic Change
Volume157
Issue number3-4
DOIs
StatePublished - Dec 1 2019

Fingerprint

hydrological change
climate modeling
evaluation
method
runoff
climate effect
global climate
watershed

All Science Journal Classification (ASJC) codes

  • Global and Planetary Change
  • Atmospheric Science

Cite this

@article{e8741c8d37e944a6b0e2557e73513fb0,
title = "Evaluation of methods for selecting climate models to simulate future hydrological change",
abstract = "A challenge for climate impact studies is the selection of a limited number of climate model projections among the dozens that are typically available. Here, we examine the impacts of methods for climate model selection on projections of runoff change for five different watersheds across the conterminous USA. The results from an ensemble of 29 global climate models and 29 corresponding hydrological model simulations are compared with the results that would have been obtained by applying six different selection methods to the climate model data and using only the selected models to drive the hydrological model. We evaluate each selection method based on whether the runoff projections produced by the method meet the method’s objective and on whether the results are sensitive to the number of models chosen. The Katsavounidis–Kuo–Zhang (KKZ) method, which is intended to maximize the spread in the projected climate change, was the only method that met both criteria for suitability. Although the KKZ method generally performed well, the results from both it and the other methods varied somewhat unpredictably based on region and number of models chosen. This study shows that the methods and models used in similar top–down studies should be carefully chosen and that the results obtained should be interpreted with caution.",
author = "Ross, {Andrew C.} and Najjar, {Raymond G.}",
year = "2019",
month = "12",
day = "1",
doi = "10.1007/s10584-019-02512-8",
language = "English (US)",
volume = "157",
pages = "407--428",
journal = "Climatic Change",
issn = "0165-0009",
publisher = "Springer Netherlands",
number = "3-4",

}

Evaluation of methods for selecting climate models to simulate future hydrological change. / Ross, Andrew C.; Najjar, Raymond G.

In: Climatic Change, Vol. 157, No. 3-4, 01.12.2019, p. 407-428.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Evaluation of methods for selecting climate models to simulate future hydrological change

AU - Ross, Andrew C.

AU - Najjar, Raymond G.

PY - 2019/12/1

Y1 - 2019/12/1

N2 - A challenge for climate impact studies is the selection of a limited number of climate model projections among the dozens that are typically available. Here, we examine the impacts of methods for climate model selection on projections of runoff change for five different watersheds across the conterminous USA. The results from an ensemble of 29 global climate models and 29 corresponding hydrological model simulations are compared with the results that would have been obtained by applying six different selection methods to the climate model data and using only the selected models to drive the hydrological model. We evaluate each selection method based on whether the runoff projections produced by the method meet the method’s objective and on whether the results are sensitive to the number of models chosen. The Katsavounidis–Kuo–Zhang (KKZ) method, which is intended to maximize the spread in the projected climate change, was the only method that met both criteria for suitability. Although the KKZ method generally performed well, the results from both it and the other methods varied somewhat unpredictably based on region and number of models chosen. This study shows that the methods and models used in similar top–down studies should be carefully chosen and that the results obtained should be interpreted with caution.

AB - A challenge for climate impact studies is the selection of a limited number of climate model projections among the dozens that are typically available. Here, we examine the impacts of methods for climate model selection on projections of runoff change for five different watersheds across the conterminous USA. The results from an ensemble of 29 global climate models and 29 corresponding hydrological model simulations are compared with the results that would have been obtained by applying six different selection methods to the climate model data and using only the selected models to drive the hydrological model. We evaluate each selection method based on whether the runoff projections produced by the method meet the method’s objective and on whether the results are sensitive to the number of models chosen. The Katsavounidis–Kuo–Zhang (KKZ) method, which is intended to maximize the spread in the projected climate change, was the only method that met both criteria for suitability. Although the KKZ method generally performed well, the results from both it and the other methods varied somewhat unpredictably based on region and number of models chosen. This study shows that the methods and models used in similar top–down studies should be carefully chosen and that the results obtained should be interpreted with caution.

UR - http://www.scopus.com/inward/record.url?scp=85074827424&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074827424&partnerID=8YFLogxK

U2 - 10.1007/s10584-019-02512-8

DO - 10.1007/s10584-019-02512-8

M3 - Article

AN - SCOPUS:85074827424

VL - 157

SP - 407

EP - 428

JO - Climatic Change

JF - Climatic Change

SN - 0165-0009

IS - 3-4

ER -