The sand sheet is a mirror reflector to the 
L-band radar by the Rayleigh roughness cri 
terion and therefore is dark on the image. 
Why do not the sand dunes have any return on 
SIR-A image? The reason is that the SIR-A 
radar beam only illuminates on the gentle 
slope of the dunes. These features of the 
dunes have been clearly indicated on the 
SIR-B imagery (Fig. 6b). The orientation of 
the dune slipface is S60°E and SIR-B illumi 
nating direction is N70 8 W. So the radar beam 
direction is approximately perpendicular to 
the dune slipface orientation. These dunes 
can be recognized from the analysis to SIR-A 
imagery in the last chapter. 
As mentioned above that the wavelength, 
polarization of SIR-A and SIR-B are the same. 
A little difference of the looking angle and 
resolution between the two systems is not an 
important influential factor for imaging the 
dunes. Therefore, it powerfully proved that 
the return will occur only when the orien 
tation of the slipface is toward to the radar 
beam and the gentle slope can not produce 
any return. The radar illuminating direction 
is a key factor to recognize the sand dunes. 
5 RADAR RESPONSE FROM SUBSURFACE ROCKS 
The Tiekuwula area to the east of Yamalik 
Dunes is a northeast valley of 10 kilometers 
long and 2 kilometers wide. Its north side 
consists of Presinian migmatite and 
south side Presinian gneiss and quartzite. 
The valley floor is widely scatterred with 
fine-selected thin sandy alluviate. 
The bedrock on both side of the valley 
gives a bright radar return, the alluvial 
valley reflects little return and shows a 
dark tone on the image. But the middle part 
is bright (In the frame of Fig. 2a). The op 
tical processed MSS image shows that the 
valley floor consists of single alluviate 
and no any trace of rock outcrops; the digi 
tal MSS composite image shows only weak re 
flection speckles on strong reflecting back 
ground (In the frame of Fig. 2b). 
The computation shows that there is the 
echo from 4 square kilometer area on SIR-A 
image. The SIR-B image with digital record 
method shows a more stronger echo (In the 
frame of Fig. 6b). It is impossible for the 
two small speckles presented on computer 
processed MSS image to produce such a large 
area return. It clearly comes from subsur 
face geological body. Two small speckles on 
MSS image is the representation of bedrock, 
the Cretaceous sand conglomerate covered by 
a thin layer of alluviate, and it is supposed 
that the large area return on SIR-A and SIR-B 
images comes from this sand conglomerate. The 
radar wave can penetrate the thin alluviate 
and detect the sand conglomerate under it. 
Another similar study area is Aertengaobao, 
100 kilometers to the west of Tiekuwula. 
Analyzed from the brightness and texture, 
the triangle-shaped echo represented clearly 
on the SIR-A image is the intrinsic bedrock 
scatterring characteristics in that area 
(Fig. 7a). The triangle is composed of three 
parts: the top and lower parts are bedrock 
and the middle is an aeolian sand zone. It 
is interesting to note that this sand zone 
is classified as aeolian sand zone on 
1:500,000 geological map of Ningxia Auto 
nomous region published in 1971. 
This means that the information provided by 
MSS coincides with that of geological map. 
What reason can explain the fact that the 
sand zone is clearly shown on MSS image and 
geological map and is shown the bedrock 
characteristics on the SIR-A image? 
It seems that radar wave has penetrated 
the windblown sand the echo come from bed 
rock buried beneath it. The geological data 
shows that the top and lower parts of Aer- 
tengaobao .triangle is: composed.of Presinian 
matomorphic rock and Mesozoic granite, res 
pectively. Other data shows that the rock 
beneath the two-meter-thick sand is gneiss. 
The temperature in this area is less than 
-20*C in winter and can reach as much as 
74° C in summer. Tne harsh climate causes 
severe weathering and erosion to the rock. 
The area is covered by sand blown from Ba- 
dain Jaran Desert and gives a smooth appea 
rance just like the one shown on the MSS 
images, but the underlain bedrock constitutes 
the source of strong scatterring because of 
its high roughness and dielectric constant. 
The penetration capability of the long 
wave radar has been verified by a lot of 
experiments. Recently, McCauley and blom 
have described the penetration of the space 
shuttle radar. Through the analysis to 
images of the SIR-A and Seasat SAR, tney 
have found the ancient river buried Deneath 
dry sand and dike covered up with alluviate. 
In October 1984, researchers from JPL buried 
radar responsor under desert area of Nevada 
in order to get synchronizm with SIR-B flight 
and it was found that the responsor has a 
sharp response to SIR-B pulse. 
The study has shown that the penetration 
of radar is confined to three conditions: 
Firstly, the fineness in granul of the sand; 
secondly, the monor thickness of covering- 
material, which can be met in Tiekuwula and 
Aertengaobao;-and thirdly,.the severe arid 
climate. The average annual rainfall is be 
low 150 mm in Alashan Plateau and even lower 
than 10 mm in part of the region, the aridity 
index is is 4-12 and reaches 7-12 in Badain 
Jaran Desert ajacent to nertengaobao and the 
evaperativity is as much as 31*5 times of 
rainfall. All of those indicate that it is 
a severe arid climate over these areas. 
It has been made clear from the above dis 
cussion that the space shuttle has penetrated 
the thin alluviate in Tiekuwula and sand in 
Aertengaobao and it posseses detectivity to 
underlain bedrock. 
6 THEORATI CAL CONSIDERATION TO TWO KINDS OF 
RETURNS FROM GROUND OBJECTS 
The above analysis to the image characteris 
tics and penetration of SIR-A and SIR-B in 
dicates that the response mechanism of dunes 
to radar wave is different to common ground 
objects such as bedrock and vegetation. In 
1980, we had measured the dielectric con 
stant (6) of igneous, sedimentory and meto- 
morphic rocks, and tne average value is 6.67* 
The average Q of dry sand is 3* There is no 
significant difference on surface roughness 
between bedrocks and dunes because of ripple- 
mark on the surface of dunes, this means that 
though the differences of dielectric constant 
and surface roughness between dunes and other 
objects have a significant impact on radar 
return, they are hot the most significant- 
factors. 
The authors suggest that the porosity of 
dune leads to much absorption of incident 
radar energy and should be considered an im 
portant factor in causing different returns, 
but the more important factor is the geome 
tric shape of dunes and their position rela 
tive to radar.