[Mitgcm-support] 1-year adjoint assimilation completed

[mitgcm-support at dev.mitgcm.org]

HyperNews notification of new message.  See: 
  http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/assim/110/1.html

Adjoint assimilation for the whole year of 1997 has completed 15
iterations of minimization. This figure[1] shows the change of cost
function and its various components with iteration. The change has slow
down substantially from iteration 10 to 15, indicating a near
convergence. The deviation of model sea level anomaly from TOPEX data is
reduced by approximately 80%. This figure[2] shows the spatial
distribution of the rms deviation of model SLA from TP data, in cm,
before and after assimilation (upper 2 panels). The lower 2 panels are
the similar plots but normalized by the prior data constraint error of
SLA. For simulation, the deviations of model SLA from TP data in the
entire tropics as well as NE and SE Pacific are much larger than unity
(much larger than the data constraint error there). After assimilation,
they are pretty much less than unity, i.e. consistent with data to
within the levels of assumed data constraint error. Comparisons of the
model state with the independent TAO data are shown below: Comparison of
temperature[3] (averaged over all 75 moorings that have complete records
in 1997). Comparison of zonal velocity[4] (averaged over all 3 moorings
(165e, 140w, 110w) that have more than 3-month records in 1997).
Comparison of meridional velocity[5] (averaged over all 3 moorings
(165e, 140w, 110w) that have more than 3-month records in 1997).
Systematic improvement of temperature and zonal velocity are pronounced.
For meridional velocity, the rms deviation from TAO data is basically
unchanged, but the correlation is improved. In terms of comparisons with
the TAO data, iteration 15 is better than 10 (which is better than 5).
But the improvement from 10 to 15 is smaller than that from 5 to 10, so
I don't think it's worth the computing time to continue the optimization
.

-----------------------------------------------------------------------
[1] http://escher.jpl.nasa.gov:2000/hosts/escher/escher4/medea/tlee/FIG_FORUM/cost_1yr.jpg
[2] http://escher.jpl.nasa.gov:2000/hosts/escher/escher4/medea/tlee/FIG_FORUM/h_anom_stat.jpg
[3] http://escher.jpl.nasa.gov:2000/hosts/escher/escher4/medea/tlee/FIG_FORUM/plot_statt_ave.jpg
[4] http://escher.jpl.nasa.gov:2000/hosts/escher/escher4/medea/tlee/FIG_FORUM/plot_statu_ave.jpg
[5] http://escher.jpl.nasa.gov:2000/hosts/escher/escher4/medea/tlee/FIG_FORUM/plot_statv_ave.jpg