[Mitgcm-support] Baroclinic Kalman filter (kf017) and control (c20000630)

[mitgcm-support at dev.mitgcm.org]

HyperNews notification of new message.  See: 
  http://escher.JPL.NASA.GOV:2000/HyperNews/get/forums/assim/106/2/2.html

Comparison of TAO data with baroclinic Kalman filter results (kf017) and
control integration (c20000630):

This[1] is a map of TAO mooring locations superimposed on thermocline
depth (i.e., depth of maximum temperature gradient) and temperature
gradient at the thermocline from TAO moorings, c20000630 control
integration, and kf017 baroclinic Kalman filter results. On average,
model and filter estimates display a sharper, shallower thermocline than
the data. Some raw temperature time series are here[2] .

The next figure[3] shows mean profiles of temperature, standard
deviation, and standard deviation difference and correlation
coefficients between model and data. On average, TAO data standard
deviation is stronger than that of the simulation by about 0.8 deg C at
the thermocline. Kalman filter estimates are improved relative to
simulation: standard deviation is strengthened at the thermocline,
standard deviation difference with TAO data is decreased, and
correlation coefficient with the data is increased. Horizontal maps of
standard deviation and model/data standard deviation difference are
displayed here[4] . Horizontal maps of correlation coefficient have also
been computed for 50-m, 100-m[5] , 150-m, and 250-m[6] depths. Most
pronounced improvement after assimilation is in Eastern Equatorial
Pacific at 200-m depth and in Western Equatorial Pacific at 100-m depth.
Statistics for individual moorings are here[7] .

During 93-97 assimilation period, data from only three TAO current meter
moorings are available. Raw time series are here[8] . Zonal velocity is
dominated by seasonal cycle which is somewhat similar to that of the
model, but to the eye there is little apparent improvement after
assimilation. Meridional velocity is dominated by high-frequency
fluctuations (less than 20 days) in the data, which are mostly lacking
from the model or the assimilation results. Summary statistics are here[9]
. Statistics for individual moorings ( 110 W[10] , 140 W[11] , and 165 E[12]
) are also displayed. Overall, assimilation does not appear to improve
current estimates at the equator, and possibly degrades zonal velocity
estimates somewhat relative to TAO moorings.
-----------------------------------------------------------------------
[1] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats/THERM1.ps
[2] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TimeSeries
[3] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats/global.ps
[4] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats/THERM5.ps
[5] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats/Corr1.ps
[6] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats/Corr2.ps
[7] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/TempStats
[8] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/Currents
[9] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/Currents/CURR.ps
[10] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/Currents/curr_110w.ps
[11] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/Currents/curr_140w.ps
[12] http://escher:2000/hosts/triton/dm1/dimitri/data/tao/matlab/FIGS/Currents/curr_165e.ps