Assessing the performance of the CFSR by an ensemble of... Li Zhang , Wesley Ebisuzaki and Arun Kumar

Assessing the performance of the CFSR by an ensemble of analysis
Li Zhang1,2, Wesley Ebisuzaki1 and Arun Kumar1
1NOAA Climate
Prediction Center, Camp Springs, Maryland
2Wyle Science Technology and Engineering, McLean, Virginia
Abstract
This study compares the Climate System Forecast Reanalysis (CFSR) tropospheric analyses to two ensembles of
analyses. The first ensemble consists of 12 hour analyses from various operational analyses for the year 2007. This
ensemble show how well the CFSR analyses can capture the daily variability. The second ensemble consists of
monthly means from the available reanalyses from the years 1979 to 2009 which is used to examine the trends.
Data
High-frequency ensemble members: twice daily (0000 and 12000 UTC) for 2007
• CFSR: Reanalysis from CFS version 2
• CMC: Operational analyses from Canadian Meteorological Centre
• EC: Operational analyses from European Centre for Medium Range Weather Forecasts
• FNO:NOGAPS analysis from Fleet Numerical Ocean and Meteorological Center
• UK: Operational analysis from the UK MetOffice
• MERRA: Reanalysis from National Aeronautics and Space Administration (NASA)
• ERA-40: Reanalyses from European Centre for Medium Range Weather Forecasts (ECMWF)
Low-frequency ensemble members: monthly mean from 1979 to 2009
• CFSR: Reanalysis from Climate Forecast System version 2
• ERA-40: Reanalyses from European Centre for Medium Range Weather Forecasts (ECMWF)
• JRA: Japanese 25-year Reanalysis Project
• MERRA: Reanalysis from National Aeronautics and Space Administration (NASA)
• R1: NCEP/NCAR reanalysis
• R2: NCEP/DOE reanalysis
Results
Zonally averaged 200 hPa height (m) at the equator and Singapore
for different reanalyses and the observed 200 hPa height (thick black
line) for Singapore. .The CFSR showed a strong warming trend that
is not in the observations and not found in the other reanalyses.
For 200 hPa zonal winds, most of the ensemble
members are doing a good job. The only major
problem was with the equatorial winds at the eastern
Pacific ocean which shows uncertainty in the
analyses.
The warming trend can be seen at
45N although amplitude is very
small.
The black line shows the difference between
CFSR and observations and shows a
warming trend before year 2000.
Corresponding results for the ensemble mean
of the 5-member AMIP simulations are
shown by the blue line. The red line is the
AMIP results with a constant factor added.
The trends are very similar before the year
2000 and differ in the last decade in which
the CFSR reproduces the observed
Singapore Z200 (observation rich period).
The zonal 850 hPa wind over the
central Pacific (CPAC from 160E200E, 5S-5N) has the similar trend.
The sum of the sensible and latent
heat flux from the ocean to the
equatorial atmosphere (10S-10N).
Summary
The CFSR was very good representing the daily variability in the current environment. For various fields examined, the
CFSR did well and was comparable to the operational models circa 2007. Compared with the older R1 and R2, the
CFSR did better at capturing the daily variability. The biggest improvements were over the oceans which were the result
of improved satellite instruments and the improved data assimilation of the satellite data. Over the well observed land
areas, the improvements over R1 and R2 were smaller.
The standard deviations of the various ensemble members from the ensemble mean averaged over 2007 for the 500 hPa height
field (HGT) and 850 hPa relative humidity (RH) are shown . Using the ensemble mean as a proxy for the truth, the ERAI has
the best analysis with the EC and UK doing well. The CFSR compares nicely with the EC and much better than R1 and R2 (not
shown here). The operational analyses and the CFSR have a smaller RMS over the ocean than the older R1 and R2. This suggests
that the newer systems are making ore effective use of satellite observations.
The extra-annual variability is difficult for the reanalyses to capture because the amplitude is usually much smaller than
the daily variability and is often masked by changes in the observation platforms. The first sign of a problem shows that
the CFSR has the strongest Z200 trend of all the reanalyses. Rawinsonde data at Singapore and Ascension Island
don’t show any large trends unlike the CFSR. The CFSR’s Z200 trend corresponds to a decadal warming of the average
tropospheric temperature which has affects the circulation both on and off the equator and can be seen in the Northern
Hemisphere mid-latitudes too.