324
Eugene E. Clothiaux, Thomas P. Ackerman, David M. Babb
useful instruments for studying the vertical distribution of clouds are active remote sensors,
such as radar and lidar. Radar is the only device that can map out the vertical distribution
of most clouds under all weather conditions, whereas lidar is sensitive to tenuous upper level
clouds in otherwise clear sky conditions. In this article we will discuss the role of radar in the
remote sensing of clouds and their properties.
Radar remote sensing of cloud properties can be viewed as a two step process. First, the actual
power levels that are eventually recorded in one form or another within the radar unit must be
interpreted; that is, they must be related to physical quantities that are of interest and use in
inferring the properties of clouds. Second, the methods by which the radar-derived quantities
are used to infer actual cloud properties must be developed. The first step is, in many ways, a
solved problem. There are standard, accurate signal processing techniques for calculating the
reflectivity, mean Doppler velocity and mean Doppler width from the receiver signal voltages
that result when a meteorological target scatters transmitted electromagnetic radiation back to
the radar receiver. The second step is by no means a solved problem and much current research
is being devoted to the development of cloud property retrieval algorithms using radar-derived
quantities as input. The discussion that follows is organized around these two steps. We first
discuss the issues of radar signal processing and then we describe the results of some of the
research that attempts to infer various cloud properties using the radar-derived quantities as
inputs.
Fortunately, the signal processing techniques applied to cloud-sensing millimeter-wave radars
are identical to the techniques developed since the late 1940s for the radar remote sensing
of rain and other interesting weather conditions (Atlas, 1990). There are several published
books that describe to various levels of detail the methods of radar-signal processing when
observing meteorological targets. As an introduction to radar-remote sensing, Battan (1973)
and Sauvageot (1992) are good starting points. Doviak and Zrnic (1984; 1993) and Ulaby et
al. (1982) provide a much more thorough description about each aspect of a radar and its use
in observing meteorological targets; however, the increased levels of rigor make the discussions
more difficult to understand without a prior knowledge of the subject. The work of Doviak and
Zrnic (1993) is especially relevant to the information that we are presenting here. Therefore, we
have adopted the approach of trying to tie our discussion as closely as possible to their work,
while attempting to provide enough discussion and figures to clarify issues that we initially
found difficult to understand.
Retrieval of cloud properties using radar and related observations is a relatively new subject.
Our treatment here summarizes several recently published methods by outlining the basic
physics and essential mathematical relationships of each. Because this field is still in its infancy,
the methods have not been extensively compared or verified. Thus we are unable to present an
evaluation of the relative accuracy of the methods.
13.2 The Radar Signal
To keep the following discussion of radar-signal processing steps as concrete as possible, we will
describe each step in the context of the 94-GHz cloud radar developed at the Pennsylvania State
University (Figures 13.1 and 13.2). Our examples are largely drawn from this system because
we are intimately familiar with its operation (Clothiaux et al., 1995). As Figure 13.2 illustrates,
there are essentially two distinct components to the radar hardware: the transmitter and the
receiver. Therefore, our discussion is divided into two parts. First, we describe the transmission
of the electromagnetic wave from the extended interactive oscillator out to the meteorological
target and finally back to the receiver antenna. Since most descriptions of radar remote sensing
Eugene E. Clothiaux, Thomas P. Ackerman, David M. Babb
useful instruments for studying the vertical distribution of clouds are active remote sensors,
such as radar and lidar. Radar is the only device that can map out the vertical distribution
of most clouds under all weather conditions, whereas lidar is sensitive to tenuous upper level
clouds in otherwise clear sky conditions. In this article we will discuss the role of radar in the
remote sensing of clouds and their properties.
Radar remote sensing of cloud properties can be viewed as a two step process. First, the actual
power levels that are eventually recorded in one form or another within the radar unit must be
interpreted; that is, they must be related to physical quantities that are of interest and use in
inferring the properties of clouds. Second, the methods by which the radar-derived quantities
are used to infer actual cloud properties must be developed. The first step is, in many ways, a
solved problem. There are standard, accurate signal processing techniques for calculating the
reflectivity, mean Doppler velocity and mean Doppler width from the receiver signal voltages
that result when a meteorological target scatters transmitted electromagnetic radiation back to
the radar receiver. The second step is by no means a solved problem and much current research
is being devoted to the development of cloud property retrieval algorithms using radar-derived
quantities as input. The discussion that follows is organized around these two steps. We first
discuss the issues of radar signal processing and then we describe the results of some of the
research that attempts to infer various cloud properties using the radar-derived quantities as
inputs.
Fortunately, the signal processing techniques applied to cloud-sensing millimeter-wave radars
are identical to the techniques developed since the late 1940s for the radar remote sensing
of rain and other interesting weather conditions (Atlas, 1990). There are several published
books that describe to various levels of detail the methods of radar-signal processing when
observing meteorological targets. As an introduction to radar-remote sensing, Battan (1973)
and Sauvageot (1992) are good starting points. Doviak and Zrnic (1984; 1993) and Ulaby et
al. (1982) provide a much more thorough description about each aspect of a radar and its use
in observing meteorological targets; however, the increased levels of rigor make the discussions
more difficult to understand without a prior knowledge of the subject. The work of Doviak and
Zrnic (1993) is especially relevant to the information that we are presenting here. Therefore, we
have adopted the approach of trying to tie our discussion as closely as possible to their work,
while attempting to provide enough discussion and figures to clarify issues that we initially
found difficult to understand.
Retrieval of cloud properties using radar and related observations is a relatively new subject.
Our treatment here summarizes several recently published methods by outlining the basic
physics and essential mathematical relationships of each. Because this field is still in its infancy,
the methods have not been extensively compared or verified. Thus we are unable to present an
evaluation of the relative accuracy of the methods.
13.2 The Radar Signal
To keep the following discussion of radar-signal processing steps as concrete as possible, we will
describe each step in the context of the 94-GHz cloud radar developed at the Pennsylvania State
University (Figures 13.1 and 13.2). Our examples are largely drawn from this system because
we are intimately familiar with its operation (Clothiaux et al., 1995). As Figure 13.2 illustrates,
there are essentially two distinct components to the radar hardware: the transmitter and the
receiver. Therefore, our discussion is divided into two parts. First, we describe the transmission
of the electromagnetic wave from the extended interactive oscillator out to the meteorological
target and finally back to the receiver antenna. Since most descriptions of radar remote sensing