Evaluation of the reanalysis products from GSFC, NCEP, and ECMWF using flux tower observations

Mark Decker, Michael A. Brunke, Zhuo Wang, Koichi Sakaguchi, Xubin Zeng, Michael G. Bosilovich

Research output: Contribution to journalArticle

176 Citations (Scopus)

Abstract

Reanalysis products produced at the various centers around the globe are utilized formany different scientific endeavors, including forcing land surface models and creating surface flux estimates. Here, flux tower observations of temperature, wind speed, precipitation, downward shortwave radiation, net surface radiation, and latent and sensible heat fluxes are used to evaluate the performance of various reanalysis products [NCEP-NCAR reanalysis and Climate Forecast System Reanalysis (CFSR) from NCEP; 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) and ECMWF Interim Re-Analysis (ERA-Interim) from ECMWF; and Modern-Era Retrospective Analysis for Research and Applications (MERRA) and Global Land Data Assimilation System (GLDAS) from the Goddard Space Flight Center (GSFC)]. To combine the biases and standard deviation of errors from the separate stations, a ranking systemis utilized. It is found thatERA-Interimhas the lowest overall bias in 6-hourly air temperature, followed closely by MERRA and GLDAS. The variability in 6-hourly air temperature is again most accurate in ERA-Interim. ERA-40 is found to have the lowest overall bias in latent heat flux, followed closely by CFSR, while ERA-40 also has the lowest 6-hourly sensible heat bias. MERRA has the second lowest and is close to ERA-40. The variability in 6-hourly precipitation is best captured byGLDAS and ERA-Interim, and ERA-40 has the lowest precipitation bias. It is also found that at monthly time scales, the bias term in the reanalysis products are the dominant cause of the mean square errors, while at 6-hourly and daily time scales the dominant contributor to the mean square errors is the correlation term. Also, it is found that the hourly CFSR data have discontinuities present due to the assimilation cycle, while the hourly MERRA data do not contain these jumps.

Original languageEnglish (US)
Pages (from-to)1916-1944
Number of pages29
JournalJournal of Climate
Volume25
Issue number6
DOIs
StatePublished - Mar 2012

Fingerprint

flight
weather
latent heat flux
data assimilation
climate
air temperature
timescale
shortwave radiation
analysis
evaluation
forecast
product
surface flux
sensible heat flux
ranking
land surface
discontinuity
wind velocity
temperature

Keywords

  • Atmosphere-land interaction
  • Land surface model
  • Model evaluation/performance

ASJC Scopus subject areas

  • Atmospheric Science

Cite this

Evaluation of the reanalysis products from GSFC, NCEP, and ECMWF using flux tower observations. / Decker, Mark; Brunke, Michael A.; Wang, Zhuo; Sakaguchi, Koichi; Zeng, Xubin; Bosilovich, Michael G.

In: Journal of Climate, Vol. 25, No. 6, 03.2012, p. 1916-1944.

Research output: Contribution to journalArticle

Decker, Mark ; Brunke, Michael A. ; Wang, Zhuo ; Sakaguchi, Koichi ; Zeng, Xubin ; Bosilovich, Michael G. / Evaluation of the reanalysis products from GSFC, NCEP, and ECMWF using flux tower observations. In: Journal of Climate. 2012 ; Vol. 25, No. 6. pp. 1916-1944.
@article{57a761350d9b4999ace486055f06d037,
title = "Evaluation of the reanalysis products from GSFC, NCEP, and ECMWF using flux tower observations",
abstract = "Reanalysis products produced at the various centers around the globe are utilized formany different scientific endeavors, including forcing land surface models and creating surface flux estimates. Here, flux tower observations of temperature, wind speed, precipitation, downward shortwave radiation, net surface radiation, and latent and sensible heat fluxes are used to evaluate the performance of various reanalysis products [NCEP-NCAR reanalysis and Climate Forecast System Reanalysis (CFSR) from NCEP; 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) and ECMWF Interim Re-Analysis (ERA-Interim) from ECMWF; and Modern-Era Retrospective Analysis for Research and Applications (MERRA) and Global Land Data Assimilation System (GLDAS) from the Goddard Space Flight Center (GSFC)]. To combine the biases and standard deviation of errors from the separate stations, a ranking systemis utilized. It is found thatERA-Interimhas the lowest overall bias in 6-hourly air temperature, followed closely by MERRA and GLDAS. The variability in 6-hourly air temperature is again most accurate in ERA-Interim. ERA-40 is found to have the lowest overall bias in latent heat flux, followed closely by CFSR, while ERA-40 also has the lowest 6-hourly sensible heat bias. MERRA has the second lowest and is close to ERA-40. The variability in 6-hourly precipitation is best captured byGLDAS and ERA-Interim, and ERA-40 has the lowest precipitation bias. It is also found that at monthly time scales, the bias term in the reanalysis products are the dominant cause of the mean square errors, while at 6-hourly and daily time scales the dominant contributor to the mean square errors is the correlation term. Also, it is found that the hourly CFSR data have discontinuities present due to the assimilation cycle, while the hourly MERRA data do not contain these jumps.",
keywords = "Atmosphere-land interaction, Land surface model, Model evaluation/performance",
author = "Mark Decker and Brunke, {Michael A.} and Zhuo Wang and Koichi Sakaguchi and Xubin Zeng and Bosilovich, {Michael G.}",
year = "2012",
month = "3",
doi = "10.1175/JCLI-D-11-00004.1",
language = "English (US)",
volume = "25",
pages = "1916--1944",
journal = "Journal of Climate",
issn = "0894-8755",
publisher = "American Meteorological Society",
number = "6",

}

TY - JOUR

T1 - Evaluation of the reanalysis products from GSFC, NCEP, and ECMWF using flux tower observations

AU - Decker, Mark

AU - Brunke, Michael A.

AU - Wang, Zhuo

AU - Sakaguchi, Koichi

AU - Zeng, Xubin

AU - Bosilovich, Michael G.

PY - 2012/3

Y1 - 2012/3

N2 - Reanalysis products produced at the various centers around the globe are utilized formany different scientific endeavors, including forcing land surface models and creating surface flux estimates. Here, flux tower observations of temperature, wind speed, precipitation, downward shortwave radiation, net surface radiation, and latent and sensible heat fluxes are used to evaluate the performance of various reanalysis products [NCEP-NCAR reanalysis and Climate Forecast System Reanalysis (CFSR) from NCEP; 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) and ECMWF Interim Re-Analysis (ERA-Interim) from ECMWF; and Modern-Era Retrospective Analysis for Research and Applications (MERRA) and Global Land Data Assimilation System (GLDAS) from the Goddard Space Flight Center (GSFC)]. To combine the biases and standard deviation of errors from the separate stations, a ranking systemis utilized. It is found thatERA-Interimhas the lowest overall bias in 6-hourly air temperature, followed closely by MERRA and GLDAS. The variability in 6-hourly air temperature is again most accurate in ERA-Interim. ERA-40 is found to have the lowest overall bias in latent heat flux, followed closely by CFSR, while ERA-40 also has the lowest 6-hourly sensible heat bias. MERRA has the second lowest and is close to ERA-40. The variability in 6-hourly precipitation is best captured byGLDAS and ERA-Interim, and ERA-40 has the lowest precipitation bias. It is also found that at monthly time scales, the bias term in the reanalysis products are the dominant cause of the mean square errors, while at 6-hourly and daily time scales the dominant contributor to the mean square errors is the correlation term. Also, it is found that the hourly CFSR data have discontinuities present due to the assimilation cycle, while the hourly MERRA data do not contain these jumps.

AB - Reanalysis products produced at the various centers around the globe are utilized formany different scientific endeavors, including forcing land surface models and creating surface flux estimates. Here, flux tower observations of temperature, wind speed, precipitation, downward shortwave radiation, net surface radiation, and latent and sensible heat fluxes are used to evaluate the performance of various reanalysis products [NCEP-NCAR reanalysis and Climate Forecast System Reanalysis (CFSR) from NCEP; 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) and ECMWF Interim Re-Analysis (ERA-Interim) from ECMWF; and Modern-Era Retrospective Analysis for Research and Applications (MERRA) and Global Land Data Assimilation System (GLDAS) from the Goddard Space Flight Center (GSFC)]. To combine the biases and standard deviation of errors from the separate stations, a ranking systemis utilized. It is found thatERA-Interimhas the lowest overall bias in 6-hourly air temperature, followed closely by MERRA and GLDAS. The variability in 6-hourly air temperature is again most accurate in ERA-Interim. ERA-40 is found to have the lowest overall bias in latent heat flux, followed closely by CFSR, while ERA-40 also has the lowest 6-hourly sensible heat bias. MERRA has the second lowest and is close to ERA-40. The variability in 6-hourly precipitation is best captured byGLDAS and ERA-Interim, and ERA-40 has the lowest precipitation bias. It is also found that at monthly time scales, the bias term in the reanalysis products are the dominant cause of the mean square errors, while at 6-hourly and daily time scales the dominant contributor to the mean square errors is the correlation term. Also, it is found that the hourly CFSR data have discontinuities present due to the assimilation cycle, while the hourly MERRA data do not contain these jumps.

KW - Atmosphere-land interaction

KW - Land surface model

KW - Model evaluation/performance

UR - http://www.scopus.com/inward/record.url?scp=84863338090&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863338090&partnerID=8YFLogxK

U2 - 10.1175/JCLI-D-11-00004.1

DO - 10.1175/JCLI-D-11-00004.1

M3 - Article

AN - SCOPUS:84863338090

VL - 25

SP - 1916

EP - 1944

JO - Journal of Climate

JF - Journal of Climate

SN - 0894-8755

IS - 6

ER -