95 
features much lower current magnitudes than the Cape Split image (here the range is from 
0 m/s to 1.6 m/s) and thus this image is closer to the phase noise floor of the radar, which 
is 10-20 cm/s for this type of target. There was also very little wind on the day of the 
experiment, which meant that the water was a very dark radar target, and thus in many 
areas (the ones where no velocity vectors are shown) the backscattered signal from the 
water was not strong enough to give any meaningful phase information. The fact that the 
channels between the islands were relatively narrow meant that there was an additional 
complication caused by shadowing from the land blocking some parts of the water. 
Despite all of this, the vectors shown in the figure do make sense for a flood tide as water 
passes from the Bay of Fundy through Letete Passage and Little Letete Passage into 
Passamaquoddy Bay. It is particularly interesting to see how the current coming through 
Letete Passage in the upper right of the figure breaks into a large eddy to the N and a 
second portion to the SW which follows the coastline of the central island. 
4. DETECTION OF UNDERWATER PHENOMENA 
On June 2, 1994 the InS AR flew a long flight line approximately two hours after low tide. 
Since the water depths were still relatively low while the currents had already become 
large, this turned out to be almost an ideal time to detect underwater phenomena with the 
radar. A large number of objects were observed in both the magnitude and velocity images 
which could be related to known subsurface features. One of the most interesting of these 
was the Scots Bay dune field. This dune field lies SW of Cape Split and is composed 
mainly of asymmetric dunes, with wavelengths ranging from 50 m to 500 m. The 
modulations in bottom topography produced modulations in the current flowing over the 
dune field and this gave rise to modulations in both the radar backscatter (and thus in the 
SAR image magnitude) and the surface current velocity (and thus the interferogram 
phase). Figure 4 shows the radar magnitude image of this dune field with three transects 
through the dune field indicated by the arrows. The intensity modulations are known to be 
caused by the dune field because this image was geocoded to UTM coordinates and 
compared with the geocoded digital terrain model (DTM) of the bottom topography 
produced by the sonar survey. The results (shown in Figure 5) are striking. In each of 
Figure 5(a), (b) and (c), three quantities are plotted as a function of distance along the 
corresponding transect; A, B or C. These quantities are the depth of the bottom below the 
water surface (obtained from the sonar DTM and corrected for the tidal height at the time 
of the radar overflight), the modulation of the radar image intensity and the radial velocity 
component of the surface current measured from the radar interferogram. Since this data 
was obtained from a single radar pass over the area, only one component of the surface 
velocity was available. Despite this, there is a close correspondence between the 
modulations in each of the three measured quantities. Depth changes of as little as 2-3 m 
produce significant variations in the velocity and backscatter, with the intensity of these 
modulations increasing as the absolute depth decreases or the depth modulation increases. 
It is also expected that the modulations would increase if the absolute current flowing over 
the dune field were larger, but that can not be proven from this experiment. Since only one 
component of the velocity was measured, it is not possible to directly relate the velocity to