Correction of Radar Reflectivity Due to Partial Beam Blocking

Beam blocking by objects such as trees, posts or towers within 1 or 2km of the radar is a major problem when using radar to estimate rainfall. The magnitude of the size of the beam-blocking problem is immediately obvious when the reflectivity profile along each azimuth is accumulated over a month or so and the partially blocked rays are clearly visible as ‘spokes’ of low integrated reflectivity. Correction for operational radars usually involves calculating the degree of beam blockage from a horizon survey or a digital terrain map but this approach is usually limited to, at most, 70% beam blocking (5dB) and cannot respond to changes in blocking when, for examples, trees come into leaf or shed their leaves in the fall.

 We propose a new method for estimating the degree of beam blocking by comparing the reflectivity at each gate over short ranges at two different elevations on days when there is widespread rain and both beams are sampling rain below the melting layer. Clutter is carefully eliminated so the degree of blocking of the lower beam can then be used to correct the reflectivity in the rain at ranges beyond 25km when the upper unblocked ray is sampling the ice rather than the rain Tests with the UK Chenies C-band radar close to London on rainy days before and after trees close to the radar were pruned, demonstrate that changes in beam blocking of the lowest elevation scan of up to 15dB can be estimated to better than 1dB. The day-to-day beam-blocking changes are small, but at some azimuths an annual cycle of blocking is evident, with more blocking in the summer, presumably because the trees are in leaf. Beam blocking is only very rarely confined to a single azimuth, and frequently extends over many degrees. Attempts to find an optimum beam-blocking pattern that minimizes the azimuthal variability of the integrated reflectivity at larger ranges fail to converge, presumably because so many adjacent azimuths have varying degrees of blocking. However, once the beam blocking has been established by the new technique, the corrected values of reflectivity integrated over a month show little variability of less than 1dB from azimuth to azimuth, even for beams with 15dB of blocking, thus providing independent evidence of the accuracy of the technique. Future work will involve more analysis of the seasonal variation of blocking, and studies with rain gauges that are situated on azimuths with blocked beams.