1 |
North Atlantic low-frequency changes of volume and |
2 |
heat transports in the ECCO state estimates |
3 |
|
4 |
K. Lorbacher, A. Koehl and D. Stammer, SIO |
5 |
|
6 |
From a solution of the global ECCO state estimation we |
7 |
diagnose |
8 |
the low-frequency (seasonal to interannual) variability of meridional |
9 |
volume and heat transports between 20N and 60N in the North Atlantic |
10 |
Ocean. The model transports are compared with estimates across the |
11 |
nominal "48N"-section (WOCE/A2) -- reaching from the English Channel |
12 |
to the Grand Banks -- where a coherent hydrographic data set was |
13 |
obtained during seven cruises from 1993 to 2000. |
14 |
|
15 |
The standard deviation of the monthly-mean anomalies of the simulated |
16 |
overturning rate at "48N" amounts to 2 Sv and of the heat transport to |
17 |
0.08 PW, respectively; the year-to-year variability is half of those |
18 |
numbers. In contrast to the model simulations, the observed amplitudes |
19 |
lie 50% above the model transport variations on interannual timescale. |
20 |
The percentage of the different components contributing to the total |
21 |
heat transport in the model is consistent with the estimates from the |
22 |
observations: In the time-mean, the baroclinic component of the heat |
23 |
transport contributes 80% to the total integral and the Ekman component |
24 |
10%. However, the latter explains 90% of the diagnosed low-frequency |
25 |
variability. In the observations the baroclinic component is mainly |
26 |
responsible for interannual variations and thereby changes in the sign |
27 |
of the horizontal component appears to be dynamically relevant. From |
28 |
the model output we compute error bars on the observed transport |
29 |
estimates to test their representativeness of low-frequency variability. |