843 
Table 1. Settlement detectability in five SIR-A strips of the U.S. environment 
±1, Mo. strip, 
iity detectable 
Strip 
Description 
Region 
No. 
detected 
Detectability 
m 
Look direction 
azimuth 
1 
St. Joseph, Mo. 
Interior Plains 
14 
87 
6°30' 
2 
Mobile, Al. 
Gulf Atlantic 
Coastal Plain 
10 
63 
32° 
3 
Tallulah, La. 
-ditto- 
9 
60 
21° 
4 
Louisville, Ky. 
Appalachian Plateau 
14 
82 
19° 
5 
Sterling, Co. 
Great Plains 
7 
100 
1° 
Figure 5. An extract from the Tallulah, La. strip. 
Note the poor contrast between settlements and the 
alluvial plain of the Mississippi River. 
Figure 7. An extract from the Stirling, Co. strip. 
Note the large, regular fields and the center pivot 
irrigated circular plots. 
e, Al. strip, 
to delineate, 
îe industriell 
r fields. 
es environment 
diversity and 
man settlements 
ble 1). 
distinguished 
ckscatter which 
rner reflector 
has long been 
of the radar 
features being 
f the microwave 
he grey tone of 
1979; Hardaway 
hat when linear 
ar to the radar 
ailed cardinal 
uta (1930) also 
ctability was 
zimuth angle on 
Figure 6. An extract from the Louisville, Ky. strip. 
Note the cardinal effect. 
aircraft imagery in which the settlement was 
oriented parallel to the flight line. In the case 
of satellite imagery such as the Seasat SAR, the 
effect was also noticeable. As for the SIR-A 
imagery which was flown at much lower altitude 
than that of Seasat (259 km versus 800 km), one 
should expect the applicability of the same 
principle. Indeed, from Table 1, it is obvious 
that when the look direction of the SIR-A antenna 
is oriented more towards the North direction, a 
higher degree of detectability of settlements 
occurs. In other words, as the settlements are 
oriented more orthogonally to the radar antenna 
look direction, strong radar, echoes result, thus 
confirming the previous observations of Henderson 
and Anuta (1980). Some cardinal effects can 
actually be observed (Fig. 6). 
An important observation of the present research 
is that the detection of large settlements 
(population over 100,000) such as Mobile, Al. 
(Fig. 4), Louisville, Ky., and Monroe, La. was 
much more difficult than the small ones. This is 
because the large settlement spreads itself out 
over a much larger area and is less compact. The 
orientation of buildings showed a greater degree 
of irregularity, thus the strength of radar 
backscatter was not uniform, as exemplified by the 
images of the three large cities mentioned above 
(Fig. 4). This observation suggests that 
detectability of settlements in radar images is 
probably not as much affected by the size factor 
as in other types of remotely sensed Imagery. It