U. Callies et al.: Surface drifters in the inner German Bight 
815 
www.ocean-sci.net/13/799/2017/ 
Ocean Sci., 13, 799-827, 2017 
cific behaviour of drifter no. 7 during days 7-11, for instance 
(Figs. 4d, 7 and 8). This could suggest that some relevant as 
pects of near-shore transports are not properly represented in 
both models. Surprisingly small effects of resolutions in both 
space and time on the metrics for Lagrangian predictability 
were also reported by Huntley et al. (2011). 
Drifters will separate even if they are released from about 
the same location. Ohlmann et al. (2012) start with O (5- 
10 m) initial separations to resolve initial non-local disper 
sion with exponential growth of the mean square pair sepa 
ration, driven by eddies larger than the distance between the 
two drifters. In the present held experiment, simultaneous de 
ployments of drifter nos. 2 and 3 were originally intended 
to study an example of drifter dispersion. The two drifters, 
both tracked over 3.7 days, stayed very close together for 
some time until they abruptly started to separate. However, 
this separation might have been triggered by an unobserved 
interaction with the research vessel. Due to such concerns, 
drifter nos. 2 and 3 were excluded from the present analysis. 
Fortunately, drifter nos. 6 and 8 offered another opportu 
nity to estimate predictability of drift trajectories. The min 
imum distance of only 800 m qualified the two drifters as a 
“chance pair” (e.g. Doos et al., 2011). Note, however, that 
drifter nos. 6 and 8 were of different types (see Table 1) so 
that relative dispersion measured may not necessarily reflect 
diffusivity of the how. On the other hand, the two drifters 
travelling jointly for about 10 days in a sense justihes the 
assumption that consequences of different designs were not 
essential. Also Fig. 6b and c provide no evidence for system 
atic differences in observed drift speeds during the period of 
interest. 
From the perspective of a model with either 900 m (BSHc- 
mod) or 1.6 km (TRIM) grid resolution, the locations of 
drifter nos. 6 and 8 almost coincided for about 10 days. The 
subsequent separation rate of about 3 km day -1 (according 
to visual inspection of Fig. 5) indicates a lower bound of pre 
diction uncertainty under these specihc conditions. An in 
dependent second estimate can be obtained considering the 
period when the two drifters converged (days 8-11). Assume 
that modelling was undertaken to determine where an item 
collected on day 11 came from. Looking 4 days back in time, 
the two drifters (nos. 6 and 8) have separated by about 20 km, 
so that the uncertainty estimate (about 5 km day -1 ) even ex 
ceeds the above value. However, the separation rate is still 
much lower than that reported by Huntley et al. (2011, their 
Fig. 3) under open ocean conditions near the Kuroshio cur 
rent, considering a similar constellation with two drifters that 
separate after staying close for a couple of days. A wide spec 
trum of typical separation rates in different regions world 
wide provided by Barron et al. (2007) also shows systemati 
cally larger values. 
Error bounds estimated from drifter conver 
gence/divergence will combine with model deficiencies 
that at least theoretically could be eliminated by model im 
provement or calibration. However, the above error estimates 
roughly ht into the general range of simulation errors found 
in this study (Fig. 11a and b). Ohlmann et al. (2012) tried 
to reproduce observed drifter trajectories with a Lagrangian 
stochastic model based on Eulerian background velocities 
derived from high-frequency (HF) radar observations inter 
polated to a regular 2x2 km 2 grid. Substantial discrepancies 
exceeding the expected level of HF radar measurement 
errors were found in occasional periods. On average, the 
separation between corresponding centres of gravity was 
found to be about 5 km after 24 h, a value that compares well 
with estimations from the present experiment. It remains as 
an open question whether the quality of predictions would 
be better with HF radar observations replacing output from 
numerical models. Ullman et al. (2006) found skills in 
predictions based on currents from either a circulation model 
or HF radar comparable. Both Ullman et al. (2006) and 
Ohlmann et al. (2012) used hourly average velocities from 
HF radar observations, i.e. the same temporal resolution as in 
the present study. Higher-resolution (e.g. 20 min; Horstmann 
et al., 2017) measurements of currents could possibly better 
capture short-term fluctuations and enhance variability in 
drift simulations. 
According to Koszalka et al. (2009) and Dôôs et al. (2011), 
“chance pairs” should possibly be distinguished from pairs 
of drifters intentionally launched together, because their be 
haviour may depend on specihc hydrodynamic conditions. 
An interesting question is what characterizes the 10-day pe 
riod when drifter nos. 6 and 8 stayed close together. The 
drifter convergence (days 7-10) coincided with the transi 
tion from a cyclonic to an anticyclonic residual current cir 
culation (Fig. 3). The anticyclonic regime forced by winds 
from mainly the north-west dominated days (11-20), except 
for a short episode (days 14-16) with very low winds and 
a circulation returning to the cyclonic orientation for about 
1 day. Drifter nos. 6 and 8 started separating again when 
residual currents gradually returned to an either indifferent or 
cyclonic circulation, a process probably best represented in 
the time series of PCi in Fig. 3. Thus, it seems that both con 
vergence and divergence of the two drifters coincided with 
reorientations of the hydrodynamic regime. 
The present data are insufficient for a discussion of to 
which extent the drifters’ observed responses to changing 
winds and residual currents depend on drifter location. Based 
on model simulations, however, there are promising tech 
niques to better describe regions within which separation for 
drifters can be expected. Identification of Lagrangian coher 
ent structures (LCSs) is a held that developed recently (e.g. 
Shadden et al., 2009). Huhn et al. (2012) applied the method 
to identify transport barriers for drifters in an estuary; Pea 
cock and Haller (2013) discuss how such techniques could be 
used for optimizing drifter deployment in the sense of max 
imizing their dispersion. Olascoaga et al. (2013) employed 
LCSs to illustrate how mesoscale circulation shapes near 
surface transports in the Gulf of Mexico.