Limitations of individual, long-range radars are that at distant ranges, the beam overshoots the lower altitudes because of the earth’s curvature, and the crossbeam resolution degrades progressively with increasing range.
A denser network of radars would provide solutions to many of these problems. Overshoot reduction and resolution enhancements are the obvious benefits. If the radars were close enough to provide coverage above their nearest neighbors, then the cone-of-silence voids would be eliminated. Because of the directional dependence of anomalous propagation, a dense network of radars might also provide some mitigation of these data corruption problems, provided that there is a data-level integration of the base data throughout the radar network. In addition to filling data voids, multiple viewing angles in overlap regions could provide multiple-Doppler resolution of the wind field. It is extremely important to investigate cost-effective ways to achieve a sufficiently dense radar network.
Thus, there are many reasons to consider a future radar system comprised of a network of long-range advanced radars operating at S-band and subnetworks of smaller radars operating at higher frequencies such as X-band. The latter would be used, for example, to cover the lower elevations below the radar horizon of the longer-range S-band radars, particularly in tornado-prone areas (where tornadoes have frequently been found to form at lower levels) (Trapp et al., 1999) and in mountainous regions. In addition, low-level boundaries such as coastal, seabreeze, gust, and some synoptic-scale fronts may also be below the radar horizon. These boundaries are often the sites of significant weather events. Such short-range radars would also serve a variety of other functions. In particular, polarimetric X-band systems can possibly measure hydrologically and climatologically important light to moderate rainfall more accurately than is possible with S-band systems (Martner et al., 2001). These radars may also be useful in detecting conditions in clouds prior to the onset of lightning where electric fields align crystals vertically, leading to regions of negative differential phase shift that are readily detected by polarimetric radars (Caylor and Chandrasekar, 1996; Carey and Rutledge, 1998).
These X-based radars would be smaller and much less expensive than the larger S-band systems and would operate unattended with control from a central command office. Spread spectrum coding of the individual radar waveforms would allow them to operate in the same band with minimal interference. A relatively small fixed phased array, perhaps cylindrical or lampshade-shaped, could be used in order to allow coverage in all azimuths for a limited sector of elevations. Furthermore, line-of-sight attenuation measurements between the individual sensors over ranges of 20–50 km may provide additional information on precipitation rates or water vapor content using differential absorption techniques.
Anticipated advances in networking technology in the next decade will bring the cost and bandwidth of wide-area networking to levels that would make a