The quest for accurate and functional methods to calibrate weather radars continues to be a difficult endeavor after a half-century of experimentation with dozens of methods. Many of the methods were summarized in the AMS workshop in Albuquerque in 2001 (Joe and Smith 2001), including the use of metal calibration spheres supported on balloons to solar emissions at radar frequencies. Atlas (2002) described additional noteworthy techniques. In this article, we add to the list by describing the external calibration method used for the NOAA/ETL scanning X-band and Ka-band radars.

The ETL method uses a trihedral corner reflector and procedures similar to those described for a dihedral reflector by Rinehart at the Albuqueque workshop. However, the physical configuration we employ is perhaps unique and is designed to minimize problems commonly encountered with using a reflector target or metal spheres. A 30.5-cm aluminum trihedral reflector concentrates the reflected signal directly backward toward the transmitting antenna. It is mounted atop an unusually tall wooden pole in order to position the reflector above major contamination from nearby ground clutter. The 25-m pole is made wood to reduce the pole's contribution to the returned signal. It is located at the western end of the Erie-1 site in Colorado, which is home base for ETL's scanning radars. The distance between the radar antenna and the pole is independently measured (typically 0.475 km). At this distance, the reflector is in the radar's far field approximately 3-degrees (a few beam widths) above ground and approximates a point target, subtending only 0.04 degrees of arc. There are no intervening trees or structures across this flat tract of prairie land.

The result is a convenient, stationary, elevated target of precisely known backscatter cross section and distance, mounted high enough on a minimally reflective support to produce a very strong return signal that is more than 25 dB above that of all clutter in the immediate vicinity. The radar is scanned very slowly across the reflector region in 0.1-degree azimuth and elevation angle steps to produce a three-dimensional map of returned power. Data can be collected in range increments as fine as 7.5 m. At each range gate the scan data produce a two-dimesional map of returned power that serves as a reasonable facsimile of the antenna power pattern near the central beam axis. Application of a standard equation for returned signal from a trihedral reflector and the point-target radar equation to the measured power maximum allows the radar system gain and radar constant to be determined. Example calculations, plots, and error analysis from an actual calibration will be presented.