S 
0 
N 
AD 
no 
AS 
5 
„J 
S 
„© 
NEE 
AD 
N 
AA 
5 
{I 
Ocean Dynamics 
Canuto et al.(2010) 
+ realisability 
Canuto et al. 
(2002) 
Canuto et al. 
(2010) 
vertical diffusivity [m?2/s’ 
d.0 0.3 0.6 I9 
vertical diffusivity [m?/s) 
D.0 3.3 0.6 0.° 
vertical diffusivity [m?/s) 
D.0 0.3 0.6 0.9 
\ 
N 
D 
.) 
€ 
= 
So 
Q 
U 
7 
\ 
20 
20 
20 
\ 
40 
-0.70.0 0.7 1.4 2. 
- T—— 
—0.70.0 0.7 1.4 2.1 
_ 40 
10 
—-0.70.0 0.7 1.4 2.1 
0.3 0.6 
0.0 0.3 0.6 
I.9 
3 
0.0 0.3 
_ 
€ 
0.6 
0.3 
D 
0.9 
J 
7 
X 
Lr 
Lr 
£ 
kl 
= 
Anl 
Q 
D 
5 
20 
20 
- 
20 
1 ; 
40 
‚40 
40 
-0.70.0 0.7 1.4 2.1 —-0.70.0 0.7 1.4 2.1 —0.70.0 0.7 1.4 2.1 
egastward velocity [m/s] eastward velocity [m/s] eastward velocity [m/s] 
Fig. 3 Profiles of both eastward current component and eddy diffusivity at position 6.55° E/55.58° N before and during the peak of storm event on 5 
December 2013 
5.1 Drift of an object 
current data sets, occurred at the beginning of the drift calcula- 
tion, this is certainly an expected result. In the further course, the 
distance of the calculated positions decreases again, whereby in 
his calculation, it was always between 5 and 8 km (Fig. 6). If 
ane also takes into account, the uncertainty that a drift calcula- 
tion entails anyway, this means that the differences (the error 
resp.) caused by the missing realizability checks, could mean a 
question of life and death in a search and rescue operation. 
The calculated drift paths (Fig. 5) show significant deviations 
especially at the beginning of the calculation. Thus, the distance 
of the calculated positions after only a few hours is approximate- 
ly 10 km (Fig. 6). Since the instabilities in the underlying cur- 
rents, which were calculated without an explicit realizability 
check, and therefore the greatest difference between the two 
Canuto et al. 
(2002) 
Canuto et al. Canuto et al. 
(2010) (2010) + realisab. 
8°E 
7° 
8°" 
g°r 
9°F 
7°E 
8°= 
9° 
SS 
; x 
ed 
SS 
S 
u 
N 
a) 
MN 
— 
© 
N 
: © 
& 
X 
55°N 
54°N 
> 
7 
—0.6 —0.4 —0.2 0.0 0.2 
eastward velocity difference [m/s] 
Fig. 4 £’s of the eastward current component at the surface between two model runs, one with O2- and one with O3-optimized code, at the peak of the 
storm on 5 December 2013 at 12 UTC after 12-h simulation 
a Springer