Marine Ecosystems Research Department - jamstec japan agency ...
Marine Ecosystems Research Department - jamstec japan agency ...
Marine Ecosystems Research Department - jamstec japan agency ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
JAMSTEC 2002 Annual Report<br />
Frontier <strong>Research</strong> System for Global Change<br />
Our dataset shows good consistency with previous<br />
knowledge on the ocean state. As for water masses,<br />
North Pacific Intermediate Water (NPIW), Subtropical<br />
Mode Water, and South/North Pacific Tropical Water,<br />
which characterize subsurface water mass structure in<br />
the North Pacific, are well reproduced. In order to validate<br />
the net air-sea heat flux field in the North Pacific,<br />
we compared our result with recent products derived<br />
from the satellite data (J-OFURO, Kubota et al., ).<br />
Since the satellite products are only one-year measurement,<br />
the quantitative difference between the two<br />
mainly comes from large interannual variabilities.<br />
Nevertheless, its overall patterns/intensities in the midlatitude<br />
region (˚N(S)-˚N(S)) have good similarities<br />
(.Wm rmsd) with the observation.<br />
Further, we have compared our results with those<br />
obtained by the nudging method often used. As a result<br />
the NPIW is reproduced largely by the artificial supply<br />
of less saline water in the nudging case (not shown),<br />
which prevents us from investigating its formation and<br />
transformation processes. In contrast, it is not the case<br />
in our reanalysis dataset because artificial sources or<br />
sinks for temperature and salinity are never added in<br />
our D-VAR system.<br />
We have also performed a sensitivity experiment.<br />
This is the powerful advantage that facilitates the identification<br />
of the water mass pathway. The experiment<br />
reveals that the origin of NPIW can be traced back to<br />
the Okhotsk Sea and the Bering Sea in the subarctic<br />
region and to the subtropical Kuroshio region further<br />
south (Fig.). These results are in broad agreement<br />
with recent studies (e.g., Yasuda et al. ).<br />
d-. Preliminary Results of s Ocean Reanalysis<br />
Using the assimilation method described above, we<br />
have initiated a long-term ocean reanalysis experiment.<br />
The aims are to get a dynamically consistent ocean<br />
state in the s and to investigate the origin and the<br />
propagation of interannual phenomena like El Niño.<br />
The assimilation method is the D-VAR with preconditioned<br />
CG, using the ocean model MOM and its adjoint<br />
code. The period is from January to December<br />
, and a spinup is performed from to .<br />
Taking account of the typical time scales of interannual<br />
phenomena, assimilation windows are set to . years.<br />
We have already performed the preliminary experiment<br />
to check the ability of our method.<br />
adS*dS sens_1nn2 -01yrs -0 / 4 sigma 26.8<br />
adS*dS sens_1nn2 -05yrs -0 / 4 sigma 26.8<br />
65N<br />
65N<br />
60N<br />
60N<br />
55N<br />
55N<br />
50N<br />
50N<br />
45N<br />
45N<br />
40N<br />
40N<br />
35N<br />
35N<br />
30N<br />
30N<br />
25N<br />
25N<br />
20N<br />
20N<br />
15N<br />
15N<br />
10N<br />
10N<br />
5N<br />
110E 120E 130E 140E 150E 160E 170E 180 170W 160W 150W 140W 130W 120W 110W<br />
-5e-05-4.5e-05-4e-05-3.5e-05-3e-05-2.5e-05-2e-05-1.5e-05-1e-05 -5e-06<br />
-5e-05-4.5e-05-4e-05-3.5e-05-3e-05-2.5e-05-2e-05-1.5e-05-1e-05 -5e-06<br />
adS*dS sens_1nn2 -03yrs -0 / 4 sigma 26.8<br />
adS*dS sens_1nn2 -06yrs -0 / 4 sigma 26.8<br />
65N<br />
65N<br />
60N<br />
60N<br />
55N<br />
55N<br />
50N<br />
50N<br />
45N<br />
45N<br />
40N<br />
40N<br />
35N<br />
35N<br />
30N<br />
30N<br />
25N<br />
25N<br />
20N<br />
20N<br />
15N<br />
15N<br />
10N<br />
10N<br />
5N<br />
110E 120E 130E 140E 150E 160E 170E 180 170W 160W 150W 140W 130W 120W 110W<br />
5N<br />
110E 120E 130E 140E 150E 160E 170E 180 170W 160W 150W 140W 130W 120W 110W<br />
5N<br />
110E 120E 130E 140E 150E 160E 170E 180 170W 160W 150W 140W 130W 120W 110W<br />
-5e-05-4.5e-05-4e-05-3.5e-05-3e-05-2.5e-05-2e-05-1.5e-05-1e-05 -5e-06<br />
-5e-05-4.5e-05-4e-05-3.5e-05-3e-05-2.5e-05-2e-05-1.5e-05-1e-05 -5e-06<br />
Fig.22 Product of adjoint variable of S by its increment dS on 26.8 sigma in the case of 'artificial cost' input<br />
at 43N, 180E (400m-depth) with streamline on the same isopycnal surface; 1-year, 3-year, 5-year,<br />
and 6-year backward calculation.<br />
140