Atmospheric Radar Research Center

Using radar backscattered signals from stationary targets (ground clutter), it is possible to extract information related to the moisture/temperature fields [Fabry et al., 1997; Fabry, 2004; Cheong et al., 2005b]. In a vacuum, an electromagnetic wave travels at the speed of light. In the atmosphere, however, the wave is slowed by a factor equivalent to the index of refraction of the air, denoted by n. Assuming a horizontally homogeneous atmosphere, the two-way time delay for a target at range r is given by td = 2r·n/c. For the necessary accuracy, signal phase is used as a proxy for this time delay which can then be processed to produce an estimate of n.
It is important to recognize that the phase depends on the path-integrated refractive index from the radar to the target. Typically, the first 10-40 km of a scanning radar’s signal is contaminated by ground clutter signals. The phase of these numerous targets can be used in concert to generate a map of refractive index, which is related to the moisture and temperate fields. This promising technique provides a previously unexplored (by radar) atmospheric parameter and is extremely important for propagation studies of electromagnetic waves (ducting, anomalous propagation, etc.), for example.

Over the past several months, the PI and his students have conducted preliminary experiments using PAR for refractivity measurements [Cheong et al., 2005b]. Following the procedure outlined by Fabry et al. [1997], it has been possible to implement the technique in order to derive refractivity change (N related to n) as shown in Figure 3. It should be noted that we have used the first several frames as the reference map. Therefore, the map depicts the change in the moisture/temperature fields over the experiment time. These preliminary results are encouraging but more research is needed to refine and validate the method.

Figure 3: First results from refractivity retrieval using the Phased Array Radar in Norman, Oklahoma. The experiment was conducted September 28, 2005, and is compared here to the Oklahoma Mesonet refractivity (right-most plot).

Although first refractivity retrieval results have been obtained from PAR, several important steps are necessary to bring the technique to a stage where it could be considered for operational use.

Generally, refractivity retrieval relies on signals reflected from ground clutter without the presence of precipitation or clear-air turbulence signals.
Mixed clutter-weather signals are difficult to interpret and the weather echoes cause a loss in phase coherence, effectively reducing the number of useful ground clutter targets for refractivity retrieval. It is proposed to explore the use of digital filters for the separation of ground clutter signals from weather signals, while retaining the information content of both. Of course, it is very common to use filters for the attenuation of ground clutter signals [e.g., Heiss et al., 1990; Torres and Zrnic, 1999].

Here, we will implement complementary filters producing a signal stream for both ground clutter and weather echoes. The ground clutter signals will be processed for refractivity retrieval while the weather signals will be used to produce the standard spectral moments. Limitations exist on the effectiveness of these filters, however, mostly due to dwell-time and the length of the filter impulse response. This is particularly important for phased array radars which do not produce a continuous stream of data but are rather steered from pulse-to-pulse for the case of multiple-application use, such as tracking and weather surveillance [Weber et al., 2005]. For this case, sophisticated methods have been developed which may prove important for high-clutter environments [Urkowitz and Owen, 1998]. These limitations will also be explored in this phase of the project.

Working closely with NSSL engineers, it is proposed to investigate the potential of real-time implementation of refractivity retrieval on PAR. Currently, full time-series data are stored during experiments for later processing. It may be possible to significantly reduce this storage/computational burden by relatively simple calculations in real-time. In particular, by calculating and storing only the average phase of the time-series data block, a significant reduction in storage would be realized. From this single real number (average phase) for each gate, it is possible to implement the refractivity retrieval algorithm. As a result, possible real-time implementation would be much closer to a reality.

In the process of reconstructing the refractivity field from phase measurements, phase wrapping typically occurs on the order of every 5-10 km in range for an S-band radar. For future gap-filling radars, which could eventually become part of the MPAR network, this phase wrapping becomes a larger concern since it is likely that shorter wavelengths (X-band) would be used for such radars. One possible solution for this dilemma is the use of a dynamic phase reference. Basically, by using a phase difference over a relatively short time, the phase wrapping will be inherently reduced. By tracking this residual phase, it should be possible to reconstruct the entire phase history allowing effective refractivity retrieval.