and their standard deviations also 
computed. 
Were 
In order to determine to what extent the two 
type of current measurements might be 
compared, the decorrelation scale for one of the 
LCD: :3178 - series’ was determined. The 
autocorrelation series were computed for the 
two component series after each series was 
detrended. 
CURRENT METER DATA 
The three current meter (CM) moorings used in 
COROAS were located at about 25.0°S, 
45.5°W; 25.4°S, 45.1? W; and 25.5°S, 44.9°W, 
in 100m (C100), 200m (C200) and 1000m 
(C1000) water depths, respectively (Fig. 1). 
While each mooring contained current meters 
set at different depths, only data from the 
shallowest current meter (32m) at each mooring 
was used. Data time series from the three 
current meters were limited to the same time 
intervals used for the drifter trajectories. For 
this study the analysis was limited to the current 
meter at 32m depth (CM32) at the C200 
mooring. 
Measurements of current speed and direction 
were made at 15 minute intervals and the data 
were averaged to obtain hourly current 
velocities. In order to obtain a visual indicatiion 
of the similarity between the drifter and current 
meter data, the velocity values for LCD 3178 
were converted in the usual manner to a 
progressive vector diagram (PVD). The origin 
for this PVD was adjusted to be the position of 
the C200 mooring. 
Mean velocities were computed for each of the 
CM series. Because the currents moved to the 
north in some series and to the south in other 
series, the series were separated into two groups 
and the means and their standard deviations for 
the two groups were determined. | 
656 
RESULTS 
The autocorrelation time series for LCD 317g 
are shown in Figure 2. 
  
    
—— Latitude ——- Longitude 
rü4 
  
  
CORRELATION (LATITUDE) 
CORRELATION (LONGITUDE) 
  
T T T 
ü 20 40 50 so 
LAG IN DAYS 
Figure 2. Decorrelation of the drifter with time 
is shown by the decrease in coefficients with 
increasing time lags. After about 23 days there 
is essentially no correlation between the earlier 
and later parts of the series. 
The correlation coefficients decrease from a 
maximum of 1 at zero time lag to à correlation 
of zero after about 23 days. We consider then 
the decorrelation scale to be about 23 days in 
the locale of the drifter trajectory. Most drifter 
and current meter time series used in this report 
were less than than 15 days. Such series are 
expected to show reasonable similarity. 
The similarity between the LCD 3178 trajectory 
and the current meter PVD is seen in Figure ? 
Both trajectories begin with a current direction 
toward the east, followed by a change towards 
the southwest. The longer PVD indicates that 
the current speed measured by the current meter 
was greater than the equivalent speed 
determined by the drifter. : 
Mean speeds and current directions for LCDs 
3178-3182 and (CM32) current meter time 
series at C200 were computed and are shown in 
Table 1. Analysis of the drifter trajectories and 
current meter velocities show a complicated 
surface circulation, with eddies frequently 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996