Volume 51 (1999) Number 4 
467 
located at - or close to - the centre of the corre 
sponding model layers. 
BSH model 
ADCP Data 
layer 1: 
surface 8.0 m 
surface - ~3 m: 
bad data, side lobes 
05 - 07 m 
layer 2: 
8 m -12.0 m 
bin 13: 
09 m -11 m 
layer 3: 
12 m -16.0 m 
bin 11: 
13 m-15 m 
layer 4: 
16 m- 20.0 m 
bin 9 
17 m-19 m 
layer 5: 
20 m - 24.0 m 
bin 7 
21 m - 23 m 
layer 6: 
24 m - 30.0 m 
bin 5 
25 m- 27 m 
layer 7: 
30 m - 37.6 m 
bin 1 
no de 
ita: 
33 m - 35 m 
35 m - 38 m 
Table 1: Vertical spacing of model layers and ADCP bins 
Oh 
H 
-h 
Cm 
H 
- 
o 
Fig. 1 BSH circulation model scheme 
4 External Forcing: Wind data 
The meteorological data recorded by the auto 
nomous DWD station on board UFS DB represent 
an average over the 10 minutes preceding the 
hourly data transmission. The winddirection is giv 
en in 10° intervals, and the wind speed in 1 knot in 
tervals. Model wind data, which are stored every 
15 minutes, are not provided at fixed intervals. This 
may partly explain the small differences between 
both data sets which can be observed in Figure 2 
and Table 2. However, these differences are not 
significant. Figure 2 shows the time series of wind 
direction and speed for the model wind (red) and 
the local observations (black). In general, all mod 
elled winds (and also simulated currents) are taken 
from the 12-36 hour prediction interval (after the 
meteorological analysis). During the first 2.5 days 
of the experiment, no observational data are avail 
able, but generally there is good agreement be 
tween model predictions and observations. Table 2 
gives a percentage frequency distribution of wind 
speed at 3 m/s intervals, and wind direction at 30° 
intervals. With respect to the different sampling 
conditions, Table 2 confirms the good agreement 
between both data sets. There are no events with 
significant differences between predicted and lo 
cally observed winds, i.e. the external forcing onto 
the sea surface, which is essential for the surface 
currents, proved to be predicted correctly during 
this period. 
5 Currents 
5.1 Current conditions at UFS DB 
Due to moderate to weak winds in the summer 
of 1999, also the current velocities were quite small. 
Figure 4 shows a progressive vector diagram (here 
after PVD) of ADCP measurements (a) and model 
data (b) from June 15 to August 10, 1999. The start 
ing position is marked with a big cross, time marks 
(X) are given every 5 days. The numbers at the 
tracks give the sampling depth (bin centre, respec 
tively layer centre) in metres. PVD’s for some sam 
pling depths show only small differences between 
measurements and model predictions. However, 
stronger deviations occur at the surface and at 
about 27 m depth. In general, both in the model re 
sults and in the measurements, there is an eastward 
drift at the surface and a southward drift near the