Climatology and interannual variability of dynamic variables in multiple reanalyses evaluated by the SPARC Reanalysis Intercomparison Project (S-RIP)

Two of the most basic parameters generated from
a reanalysis are temperature and winds. Temperatures in the
reanalyses are derived from conventional (surface and balloon),
aircraft, and satellite observations. Winds are observed
by conventional systems, cloud tracked, and derived from
height fields, which are in turn derived from the vertical
temperature structure. In this paper we evaluate as part of
the SPARC Reanalysis Intercomparison Project (S-RIP) the
temperature and wind structure of all the recent and past
reanalyses. This evaluation is mainly among the reanalyses
themselves, but comparisons against independent observations,
such as HIRDLS and COSMIC temperatures, are
also presented. This evaluation uses monthly mean and 2.5◦
zonal mean data sets and spans the satellite era from 1979–
2014. There is very good agreement in temperature seasonally
and latitudinally among the more recent reanalyses
(CFSR, MERRA, ERA-Interim, JRA-55, and MERRA-2)
between the surface and 10 hPa. At lower pressures there
is increased variance among these reanalyses that changes
with season and latitude. This variance also changes during
the time span of these reanalyses with greater variance during
the TOVS period (1979–1998) and less variance afterward
in the ATOVS period (1999–2014). There is a distinct
change in the temperature structure in the middle and upper
stratosphere during this transition from TOVS to ATOVS
systems. Zonal winds are in greater agreement than temperatures
and this agreement extends to lower pressures than the
temperatures. Older reanalyses (NCEP/NCAR, NCEP/DOE,
ERA-40, JRA-25) have larger temperature and zonal wind
disagreement from the more recent reanalyses. All reanalyses
to date have issues analysing the quasi-biennial oscillation
(QBO) winds. Comparisons with Singapore QBO winds
show disagreement in the amplitude of the westerly and easterly
anomalies. The disagreement with Singapore winds improves
with the transition from TOVS to ATOVS observations.
Temperature bias characteristics determined via comparisons
with a reanalysis ensemble mean (MERRA, ERAInterim,
JRA-55) are similarly observed when compared
with Aura HIRDLS and Aura MLS observations. There is
good agreement among the NOAA TLS, SSU1, and SSU2
Climate Data Records and layer mean temperatures from
the more recent reanalyses. Caution is advised for using reanalysis
temperatures for trend detection and anomalies from
a long climatology period as the quality and character of reanalyses
may have changed over time