Exploring the efficiency of bias corrections of regional climate model output for the assessment of future crop yields in Europe

A.M.R. Bakker, J.J.E. Bessembinder, A.J.W. de Wit, B.J.J.M. van den Hurk, S.B. Hoek

Research output: Contribution to journalArticleAcademicpeer-review

13 Citations (Scopus)

Abstract

Excessive summer drying and reduced growing season length are expected to reduce European crop yields in future. This may be partly compensated by adapted crop management, increased CO2 concentration and technological development. For food security, changes in regional to continental crop yield variability may be more important than changes in mean yields. The assessment of changes in regional and larger scale crop variability requires high resolution and spatially consistent future weather, matching a specific climate scenario. Such data could be derived from regional climate models (RCMs), which provide changes in weather patterns. In general, RCM output is heavily biased with respect to observations. Due to the strong nonlinear relation between meteorological input and crop yields, the application of this biased output may result in large biases in the simulated crop yield changes. The use of RCM output only makes sense after sufficient bias correction. This study explores how RCM output can be bias corrected for the assessment of changes in European and subregional scale crop yield variability due to climate change. For this, output of the RCM RACMO of the Royal Netherlands Meteorological Institute was bias corrected and applied within the crop simulation model WOrld FOod STudies to simulate potential and water limited yields of three divergent crops: winter wheat, maize and sugar beets. The bias correction appeared necessary to successfully reproduce the mean yields as simulated with observational data. It also substantially improved the year-to-year variability of seasonal precipitation and radiation within RACMO, but some bias in the interannual variability remained. This is caused by the fact that the applied correction focuses on mean and daily variability. The interannual variability of growing season length, and as a consequence the potential yields too, appeared even deteriorated. Projected decrease in mean crop yields is well in line with earlier studies. No significant change in crop yield variability was found. Yet, only one RCM is analysed in this study, and it is recommended to extend this study with more climate models and a slightly adjusted bias correction taking into account the variability of larger time scales as well
Original languageEnglish
Pages (from-to)865-877
JournalRegional Environmental Change
Volume14
Issue number3
DOIs
Publication statusPublished - 2014

Fingerprint

crop yield
regional climate
climate modeling
crop
growing season
weather
water yield
sugar beet
technological development
Europe
food security
wheat
maize
timescale
climate change
food
winter
climate
summer
simulation

Keywords

  • weather generator
  • global radiation
  • change scenarios
  • era-interim
  • variability
  • precipitation
  • projections
  • simulation
  • uncertainty
  • circulation

Cite this

@article{cedfda0f4bb24128baf10e38ac573d39,
title = "Exploring the efficiency of bias corrections of regional climate model output for the assessment of future crop yields in Europe",
abstract = "Excessive summer drying and reduced growing season length are expected to reduce European crop yields in future. This may be partly compensated by adapted crop management, increased CO2 concentration and technological development. For food security, changes in regional to continental crop yield variability may be more important than changes in mean yields. The assessment of changes in regional and larger scale crop variability requires high resolution and spatially consistent future weather, matching a specific climate scenario. Such data could be derived from regional climate models (RCMs), which provide changes in weather patterns. In general, RCM output is heavily biased with respect to observations. Due to the strong nonlinear relation between meteorological input and crop yields, the application of this biased output may result in large biases in the simulated crop yield changes. The use of RCM output only makes sense after sufficient bias correction. This study explores how RCM output can be bias corrected for the assessment of changes in European and subregional scale crop yield variability due to climate change. For this, output of the RCM RACMO of the Royal Netherlands Meteorological Institute was bias corrected and applied within the crop simulation model WOrld FOod STudies to simulate potential and water limited yields of three divergent crops: winter wheat, maize and sugar beets. The bias correction appeared necessary to successfully reproduce the mean yields as simulated with observational data. It also substantially improved the year-to-year variability of seasonal precipitation and radiation within RACMO, but some bias in the interannual variability remained. This is caused by the fact that the applied correction focuses on mean and daily variability. The interannual variability of growing season length, and as a consequence the potential yields too, appeared even deteriorated. Projected decrease in mean crop yields is well in line with earlier studies. No significant change in crop yield variability was found. Yet, only one RCM is analysed in this study, and it is recommended to extend this study with more climate models and a slightly adjusted bias correction taking into account the variability of larger time scales as well",
keywords = "weather generator, global radiation, change scenarios, era-interim, variability, precipitation, projections, simulation, uncertainty, circulation",
author = "A.M.R. Bakker and J.J.E. Bessembinder and {de Wit}, A.J.W. and {van den Hurk}, B.J.J.M. and S.B. Hoek",
year = "2014",
doi = "10.1007/s10113-013-0557-9",
language = "English",
volume = "14",
pages = "865--877",
journal = "Regional Environmental Change",
issn = "1436-3798",
publisher = "Springer Verlag",
number = "3",

}

Exploring the efficiency of bias corrections of regional climate model output for the assessment of future crop yields in Europe. / Bakker, A.M.R.; Bessembinder, J.J.E.; de Wit, A.J.W.; van den Hurk, B.J.J.M.; Hoek, S.B.

In: Regional Environmental Change, Vol. 14, No. 3, 2014, p. 865-877.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Exploring the efficiency of bias corrections of regional climate model output for the assessment of future crop yields in Europe

AU - Bakker, A.M.R.

AU - Bessembinder, J.J.E.

AU - de Wit, A.J.W.

AU - van den Hurk, B.J.J.M.

AU - Hoek, S.B.

PY - 2014

Y1 - 2014

N2 - Excessive summer drying and reduced growing season length are expected to reduce European crop yields in future. This may be partly compensated by adapted crop management, increased CO2 concentration and technological development. For food security, changes in regional to continental crop yield variability may be more important than changes in mean yields. The assessment of changes in regional and larger scale crop variability requires high resolution and spatially consistent future weather, matching a specific climate scenario. Such data could be derived from regional climate models (RCMs), which provide changes in weather patterns. In general, RCM output is heavily biased with respect to observations. Due to the strong nonlinear relation between meteorological input and crop yields, the application of this biased output may result in large biases in the simulated crop yield changes. The use of RCM output only makes sense after sufficient bias correction. This study explores how RCM output can be bias corrected for the assessment of changes in European and subregional scale crop yield variability due to climate change. For this, output of the RCM RACMO of the Royal Netherlands Meteorological Institute was bias corrected and applied within the crop simulation model WOrld FOod STudies to simulate potential and water limited yields of three divergent crops: winter wheat, maize and sugar beets. The bias correction appeared necessary to successfully reproduce the mean yields as simulated with observational data. It also substantially improved the year-to-year variability of seasonal precipitation and radiation within RACMO, but some bias in the interannual variability remained. This is caused by the fact that the applied correction focuses on mean and daily variability. The interannual variability of growing season length, and as a consequence the potential yields too, appeared even deteriorated. Projected decrease in mean crop yields is well in line with earlier studies. No significant change in crop yield variability was found. Yet, only one RCM is analysed in this study, and it is recommended to extend this study with more climate models and a slightly adjusted bias correction taking into account the variability of larger time scales as well

AB - Excessive summer drying and reduced growing season length are expected to reduce European crop yields in future. This may be partly compensated by adapted crop management, increased CO2 concentration and technological development. For food security, changes in regional to continental crop yield variability may be more important than changes in mean yields. The assessment of changes in regional and larger scale crop variability requires high resolution and spatially consistent future weather, matching a specific climate scenario. Such data could be derived from regional climate models (RCMs), which provide changes in weather patterns. In general, RCM output is heavily biased with respect to observations. Due to the strong nonlinear relation between meteorological input and crop yields, the application of this biased output may result in large biases in the simulated crop yield changes. The use of RCM output only makes sense after sufficient bias correction. This study explores how RCM output can be bias corrected for the assessment of changes in European and subregional scale crop yield variability due to climate change. For this, output of the RCM RACMO of the Royal Netherlands Meteorological Institute was bias corrected and applied within the crop simulation model WOrld FOod STudies to simulate potential and water limited yields of three divergent crops: winter wheat, maize and sugar beets. The bias correction appeared necessary to successfully reproduce the mean yields as simulated with observational data. It also substantially improved the year-to-year variability of seasonal precipitation and radiation within RACMO, but some bias in the interannual variability remained. This is caused by the fact that the applied correction focuses on mean and daily variability. The interannual variability of growing season length, and as a consequence the potential yields too, appeared even deteriorated. Projected decrease in mean crop yields is well in line with earlier studies. No significant change in crop yield variability was found. Yet, only one RCM is analysed in this study, and it is recommended to extend this study with more climate models and a slightly adjusted bias correction taking into account the variability of larger time scales as well

KW - weather generator

KW - global radiation

KW - change scenarios

KW - era-interim

KW - variability

KW - precipitation

KW - projections

KW - simulation

KW - uncertainty

KW - circulation

U2 - 10.1007/s10113-013-0557-9

DO - 10.1007/s10113-013-0557-9

M3 - Article

VL - 14

SP - 865

EP - 877

JO - Regional Environmental Change

JF - Regional Environmental Change

SN - 1436-3798

IS - 3

ER -