Page 133 - Fundamentals of Radar Signal Processing

also exhibit spatial correlation: the reflectivity samples from adjacent resolution

cells may be correlated. Two excellent general references on land and sea
clutter phenomenology are Ulaby and Dotson (1989) and Long (2001). A good
brief introduction is Currie (2010).
Clutter echoes differ from target echoes in that they will typically exhibit
different PDFs, temporal and spatial correlation properties, Doppler
characteristics, and power levels. These differences can be exploited to

separate target and clutter signals. Means to do so are the principal concern of
Chaps. 5 and 9. Clutter differs from noise in two major ways: its power
spectrum is not white (i.e., it is correlated interference), and, since it is an echo
of the transmitted signal, the received clutter power is affected by such radar
and scenario parameters as the antenna gain, transmitted power, and the range
from the radar to the terrain. In contrast, noise power is affected by none of
those factors, but is affected by the radar receiver noise figure and bandwidth.

2.3.1 Behavior of σ 0
Area clutter (land and sea surface) reflectivity is characterized by its mean or
0
median value of radar cross section, σ (dimensionless), the probability density
function of the reflectivity variations, and their correlation in space and time.

0
Many of the same PDFs described in Sec. 2.2.5 are applied to modeling σ as
well. Popular examples include the exponential, lognormal, and Weibull
distributions.
0
The area reflectivity σ of terrain observed by the radar is a strong function
of terrain type and condition (e.g., surface roughness and moisture), weather
(wind speed and direction, precipitation), engagement geometry (especially
grazing angle), and radar parameters (wavelength, polarization). Consequently,

selection of a PDF is not sufficient to model clutter. It is also necessary to
0
model the dependence of σ on these parameters. Consider land clutter. Values
0
of σ commonly range from –60 to –10 dB. Extensive measurement programs
over the years have collected statistics of land clutter under various conditions

and resulted in many tabulations of σ for various terrain types and conditions,
0
0
as well as models for the variation of σ . Figure 2.20 shows one set of
representative data for the area reflectivity of desert terrain versus radar
0
frequency and grazing angle. Note that σ generally increases with radar
frequency, and decreases at shallower grazing angles. For a given frequency, the

variation with grazing angle over the range shown is 20 to 25 dB. For a given
grazing angle, the variation across frequency in this example is about 10 dB.
0
Figure 2.21 is one example of the variation in σ versus grazing angle for
different terrain types at a fixed frequency, in this case S band. Generally,
reflectivity increase with terrain roughness, from the presumably smoother

desert terrain to the complex, rough urban terrain.