Antenna design tradeoffs for dense distributed radar network for weather sensing

The design of a dense distributed network of radars that work together to detect, track and predict localized weather phenomena in the lower atmosphere layer (0-3 km) in X-band represents a new approach proposed by Collaborative Adaptive Sensing of the Atmosphere (CASA). These collaborative systems require the ability to observe multiple non-contiguous regions. To achieve this capability, CASA proposes the use of electronically steered antenna. X-band provides fine special resolution with smaller antennas, but is affected more by precipitation than lower frequencies. The antenna's performance plays an important roll in the system's ability to overcome these effects. Polarimetric radar system is considered of great utility to improve the quantities estimation precipitation. Accurate polarization and sidelobe performance are required to minimize the induced errors in the system. Antennas of small sizes, low power, high scanning performance and low cost are principal requirements of the system required by CASA. This paper discuss the trade-offs of key design parameters of electronically scanned phased array antennas, sector configuration and topology for a dense radar network environment.

The principal objective of this paper is to define the optimum system parameters (network topology, sector and antenna configuration) for a dense network according to CASA's requirements. The approach consists in evaluating the performance of azimuth resolution (èaz), minimum radar sensitivity (Ze), differential reflectivity bias (Zbdr), integrated cross polarization (ICPR) and cost as a function of the number of sectors per node, scan loss, beamwidth broadening effect and cross-polarization performance. Three sector-configurations over a triangular and square network topology were considered to obtain an equivalent hemispherical coverage. Configuration A represents 3 sectors of 120 deg (±60 deg); configuration B represents 4 sectors of 90 deg (±45 deg) and configuration C represents 6 sectors of 60 deg (±30 deg).

Simulation results of èaz, Ze, Zbdr and ICPR for a single node and network environment were evaluated for the three sector-configurations and two grid topologies. An electronic scan of ±6 degrees in elevation at a 6 degree fixed tilt was used to achieve 100% coverage at 3 km of altitude. Mismatch beam pattern and cross-polarization as a function of angle are evaluated to determine the induced errors in the two polarimetric parameters Zbdr and ICPR. Results of a network environment show the polarization performance and cost as primary drivers to define the system configuration according with CASA requirements. Preliminary requirements for electronically scanned antennas are presented. For a single node and network environment, results show that configuration C (6x60 deg) is the best candidate in terms of radar performance. Nevertheless it represents the most expensive alternative in comparison with configurations A and B. The radar RF system cost of C is ~50% greater than B and almost 100% greater than A.