[Mitgcm-support] Note: February 8, 2001
mitgcm-support at dev.mitgcm.org
mitgcm-support at dev.mitgcm.org
Wed Jul 9 15:56:07 EDT 2003
HyperNews notification of new message. See:
http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/wkmtg/82.html
=======================================================================
NOTES FROM THE ASSIMILATION MEETING : 8 February 2001
1) K-filter
------------
- A summary of runs calibrating process noise (Q) have been posted[1].
Process noise needs to be calibrated so that model errors are
consistent with actual model-data differences. Model errors must
taper to zero at open cell boundaries to account for
representation error.
However, comparisons with independent data suggest runs with
calibrated Q are actually worse than non-calibrated Q.
===> Comparisons with TAO and XBTs to be summarized and
posted [Dimitris].
===> Further calibration and/or adjustment is likely necessary
[Ichiro/Benyang].
- Assimilation (and control run) has been extended to the end of
the 3rd quarter in 2000. The particular assimilation is kf029,
and is a combined BC-BT filter using the calibrated Q of kf024i
(BC) and kf022e (BT).
- A 5-mode BC filter is being prepared. The code is being checked
for consistency.
===> The results of the 1-year adjoint below could be analyzed to
examine the significance of higher baroclinic modes in correcting
the model. [Ichiro/Benyang/Tony]
2) Adjoint
-----------
- The 1-year adjoint has converged to an acceptable limit as
reported here[2]. The Ames queue can now be used for;
* Repeat T/P and Levitus assimilation for 1998
* Process N.Smith's XBT data and assimilate that for 1997
===> Doing the '98 assimilation seems most rewarding [Tony/Benny].
- Changes in wind stress accounts for most of the correction
achieved by the 1-year adjoint. Plots analyzing the model and
wind correction is posted here[3].
The wind corrections show patterns of westward propagation.
Could these be due to air-sea coupling or is it an artifact of
the assimilation? W.White has found evidence for a westward
propagating wind signal at mid-latitudes that he attributes to
coupling.
===> Changes in wind could be validated wrt NSCAT and/or TAO [Tony].
- There appears to be a bug in Moto's package implemented in the
ECCO code (and also in the latest version of the MITcode).
Changes in surface temperature do not match the sum of terms
controlling temperature. Other depths are OK. It is anticipated
that the control run (c20000630 and Moto's analysis) were OK.
===> The gsgt package needs to be debugged [Dimitris].
- Request to purchase TAF has been forwarded to procurement.
3) Figure for Eric
-------------------
- Two figures were prepared for Eric's NASA Report, both showing
anomalies in November 1997. The first figure[4] compares T/P sea
level (top) and an assimilation estimate (lower) of heat content
and surface velocity. The second figure[5] compares assimilation
estimates of temperature and velocity at the surface (top) and
200m depth (bottom). The assimilation estimate is T/P K-filter
assimilation run kf024f. Anomalies are with respect to '93-'95
average seasonal cycle.
===> The figures (or a combined version) are to be put on our web-site
[Dimitris/Tony/Ichiro/Zhangfan].
- Further analysis of assimilation results point to other possibly
intriguing mechanisms of change. Evidence suggests tropical
temperature change near the surface (at least above 200m) being
mostly due to zonal advection. At depth (~1200m), things look
different; it shows Rossby wave-like patterns suggesting changes
being dominated by vertical heaving. There is a curious
coincidence between what appears to be a reflected Rossby wave of
the 97-98 ENSO reaching the western end in 1999 along 8N (~1200m)
and the emergence of the negative PDO pattern at the surface.
4) Other
---------
- Directories on hpss have been changed. All files have been moved
to
/proj/ocsatmeas/xing
etc. Now all group members have read access to other members'
files.
[Somehow I now have to turn off the "passive mode" for my ftp to
work with hpss.]
- Lee showed comparisons of ssh along 30N in the Pacific between
T/P and c20000630 from 1993 till end of 1999. There are
similarities in large-scale westward propagating signals. There
appears to be a sudden increase in amplitude in 1994, coincident
with a basin-wide wind event.
- There is a problem with the automatic Legato backup for guppy and
nautilus. Zhangfan and I-Lin are working on it. Make sure you
are getting the automatic e-mail of backup information and that
the backups are successful.
- There are several web-sites that provide images from
assimilation. They should be examined and used as a reference to
implement something similar for our project. These web-sites
include;
DIADEM: http://diadem.halo.is/diadem/
MERCATOR: http://www.mercator.com.fr/
NSIPP: http://pong.gsfc.nasa.gov/index.html
The NOAA OGP site has an animation of SST that runs in the
browser;
http://www.ogp.noaa.gov/enso/animation.html
=======================================================================
-----------------------------------------------------------------------
[1] http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/assim/112/1.html
[2] http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/assim/110/1.html
[3] http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/assim/110/2.html
[4] http://escher.JPL.NASA.GOV:2000/hosts/escher/escher2/medea/if/forum/Dimitris/ODA.jpg
[5] http://escher.JPL.NASA.GOV:2000/hosts/escher/escher2/medea/if/forum/Tony/globe2.ps
More information about the MITgcm-support
mailing list