814 
U. Callies et al.: Surface drifters in the inner German Bight 
Ocean Sci., 13, 799-827, 2017 
www.ocean-sci.net/13/799/2017/ 
(a) Model error: distance (b) Model error: distance 
0 5 10 1 5 2 0 2 5 30 0 5 1 0 15 20 2 5 30 
km km 
(c) Model error: angle (d) Model error: angle 
On m 
-180 -90 
90 0 
Degrees 
Figure 11. Distribution of model errors in 25 h drifter simulations. Histograms are based on 164 simulations in total for drifter nos. 5, 6, 8 
and no. 9. Referring to drift simulations based on BSHcmod +W and TRIM, respectively, panels (a, b) evaluate spatial separations shown 
in Fig. 10a. For the same set of 164 simulations, panels (c, d) evaluate directional errors from Fig. lOd. Red lines indicate median values (4.6 
and 5.4km in a, b; —15 and 7° in c, d). 
the inner German Bight can therefore not always be fixed by 
simply adding windage or Stokes drift. 
In both BSHcmod + W and TRIM simulations, drifter dis 
placements were often rotated to the left of their observed 
counterparts, e.g. during days 13-23 (see Figs. lOd or 8d and 
h). A parametrization of wind-induced Ekman drift (Rohrs 
and Christensen, 2015) might be explored as a means to rem 
edy such model deficiencies including lacking representa 
tion of the Coriolis-Stokes drift (Hasselmann, 1970; Polton 
et al., 2005) driven by ocean surface waves. Fig. 11 shows 
error distributions that combine all data from Fig. 10a and d, 
respectively. Median errors of drifter displacements are of 
the order of 5 km for both BSHcmod (4.6 km) and TRIM 
(5.4km). BSHcmod+ W tends to have negative directional 
errors (median value of about 15° to the left of observations), 
while the median directional error for TRIM is about 7° to 
the right. Negative deflections of BSHcmod + W simulations 
happen to coincide with what one would expect from a sim 
ple parametrization of windage (or Stokes drift) that neglects 
effects of Coriolis force. However, distributions in Fig. 11 
combine simulations under very different wind conditions, 
and directional biases are not permanent. In many cases (e.g. 
day 18 in Fig. 8d and h), directional errors of the two simula 
tions resemble each other. One must therefore be very careful 
to interpret shifted median values in terms of specific model 
deficiencies. Differences between Fig. 11c and d are proba 
bly not statistically significant, so we refrained from trying 
to incorporate and tune additional effects of Coriolis force. 
Drifter nos. 5, 6 and 8 played a central role in this study 
because their trajectories overlapped for 40 days, enabling 
tentative conclusions regarding spatial scales that affected 
long- and short-term drifter displacements. Wind fields re 
solved in numerical models (and also corresponding fields 
of Stokes drift) tend to vary smoothly on a regional scale. 
A substantial impact of winds on surface currents may be 
one of the reasons why simulated trajectories resemble each 
other more than corresponding observations. However, also 
observed drifter paths show similarities that point to the im 
pact of large-scale forcing. 
Due to bathymetric constraints and different scales of rele 
vant processes, spatial variability of marine currents tends to 
be higher than that of wind fields (Rohrs et al., 2012). How 
ever, our study did not show clear effects of the higher resolu 
tion in BSHcmod regarding either space (900 m compared to 
1.6 km in TRIM) or time (15 min compared to 1 h in TRIM). 
Both TRIM and BSHcmod are unable to reproduce the spe