Geoscience Reference
In-Depth Information
0:7
ı
,areused
After 2001 ERA-Interim reanalysis fields, which have a resolution of
in conjunction with CCMP winds.
A comparison of the WCRA13 and WCRA31 analyses during the overlapping
period 1999-2011 will reveal the impact of the resolution of the
prior
surface
forcing fields on the circulation estimates.
14.4.3
WCRA Observations
The observations assimilated into the model during each WCRA analysis were
collected by various platforms, and are summarized in Table
14.1
, along with the
combined measurement error and error of representativeness that is assumed for the
diagonal entries of
R
.
All of the in situ hydrographic profiles of
were taken from the quality
controlled EN3 data archive maintained by the UK Met Office as part of the
European Union ENSEMBLES project (
Ingleby and Huddleston 2007
). The version
of EN3 used here is version 2a which includes the XBT and MBT temperature error
corrections of
Levitus et al.
(
2009
).
The in situ observations from the EN3 archive are available from a variety of
different observing platforms including: expendable bathythermographs (XBTs),
mechanical bathythermographs (MBTs), conductivity temperature depth devices
(CTDs), free drifting Argo profiling floats, and autonomous pinniped bathythermo-
graphs (APBs) in the form of tagged marine mammals. Figure
14.2
shows a time
series of log
T
and
S
of the total number of super observations from EN3 and each satellite
platform that fall within the ROMS CCS model domain during each month of the
year spanning the full period of WCRA31.
The SSH observations assimilated into the model are in the form of 1 day gridded
composites of the mean dynamic topography from Aviso. Before assimilation, the
mean dynamic topography of the Aviso data averaged over the ROMS CCS domain
was corrected to match that of the model. This data is used rather than the raw
along-track data because at the present time there is no temporal correlation included
in the
prior
error covariance matrices
D
or
R
. The result is that information from
individual along track observations is lost quite quickly and becomes ineffective
for constraining the model unless some additional effort is made to persist the
along-track observations over time. This problem is alleviated by using the gridded
products, although we appreciate that this is not an ideal solution because of
the limitations of the objective mapping technique used to map the altimeter
observations onto a regular grid. In addition, satellite SSH observations near the
coast are known to be unreliable (
Saraceno et al. 2008
) so only observations that are
more than 50 km from the coast are assimilated into ROMS.
As indicated in Table
14.1
, SST observations are available from several different
platforms. The along track data from each platform are used, and when multiple
platforms are concurrently available, the data from each platform are combined to
form super observations. Only the number of super observations for each individual
platform are shown in Fig.
14.2
.
10
Search WWH ::
Custom Search