internal waves generated by ships and imaged by SAR were obtained during 
the 1983 Georgia Strait Experiment. One such example is shown in 
Figure 11. Another example of an apparent ship-generated internal wave in 
Georgia Strait is shown for both X- and L-band in Figure 12. The dif- 
ference between the appearance of these two wakes is due to the difference 
in stratification conditions, the former being characterized by a maximum 
Brunt-Vaisala frequency of 115 cycles per hour at a depth of 6.5 meters, 
and the latter having a maximum B-V frequency of 60 cycles per hour at a 
depth of 9.5 meters. 
A sketch indicating the displacements caused by ship-generated interna] 
waves is shown in Figure 13. These displacements imply the existence of 
surface currents which interact with small-scale surface waves to produce 
changes in surface roughness detectable by a SAR. The X- and L-band 
imagery shown in Figure 12 were co-registered and the locations of the 
SAR-observed wake features were compared. The results of this study indi- 
cated that the wave pattern in each of the two bands is very nearly spa- 
tially coincident. That is, the alternating bright and dark wake-related 
features occur in the same location in each image, in contrast with the 
bright lines resulting from the Bragg dispersion mechanism discussed above 
which are strongly wavelength-dependent. A detailed description of these 
interactions is beyond the scope of this paper. However, a great deal of 
effort has been expended in constructing and testing models for these 
interactions. In general, existing models underpredict the observed sur- 
face effects associated with internal waves, particularly at X-band wave- 
lengths, and further effort is required in both the theoretical and experi- 
mental aspects of these investigations. 
It has recent been observed by Swanson (1986) that there appears to be 
a correlation between the presence of ship-generated internal waves and the 
absence of turbulent wake features in SAR imagery. His explanation for 
this observation is that the wake vortex energy is transferred to the 
internal waves, which weakens the vortex-related currents. This reduces 
their effect on ambient waves, and hence, radar backscatter. A cursory 
examination of historic SAR data sets archived at ERIM appears to support 
Swanson's observation. His explanation could be tested by examining two 
data sets with similar parameters but different stratification characteris- 
tics. Hopefully, this would allow us to isolate the effects of internal 
wave formation. 
5. SUMMARY 
A summary of the various ship wake phenomena observed in SAR images, 
along with an estimate of their dependence on environmental conditions and 
SAR parameters, is presented in Table 1. The narrow wakes produced by the 
Bragg wave dispersion mechanism are observed only under very low wind con- 
ditions, under any stratification conditions, and are most strongly ob- 
served at L-band. Examples exist for all SAR look directions, although a 
look direction perpendicular to the ship track may result in stronger sig- 
natures. Classical Kelvin wakes, on the other hand, are observed under 
moderate wind conditions since they are visible through the modulation of 
an existing field of ambient Bragg waves. These wakes are observed at both 
L-band and X-band, and with all look directions, although certain look 
directions may optimize the visibility of various parts of the Kelvin wake. 
For example, the Kelvin envelope is most easily observed when it is aligned 
in the azimuth direction, and the individual cusp waves are better resolved 
when they are range-travelling (i.e., for a look direction of 35° with 
422 
  
Fi