Beam occultation impacts on quantitative precipitation estimates from the Purdue University X-band Teaching and Research Radar (XTRRA)

Weather radars are sometimes installed in areas suffering from beam blockage owing to nearby buildings, trees, and terrain. All of these types of obstructions surround the XTRRA at Purdue University, impacting the quality of its data. XTRRA is situated on top of a five story building, but still suffers significant beam blockage in multiple sectors. We used high-resolution lidar data collected over Purdue’s campus to estimate the percentage occultation of XTRRA at various elevation and azimuth angles, modeling XTRRA’s 1-degree beam as a spotlight. By applying this occultation information to mask derived quantitative precipitation estimates (QPEs), we sought to improve root mean squared error (RMSE) statistics relative to nearby tipping bucket rain gauges. As expected, the blockage severely compromises QPE accuracy at the lowest elevation angles (<=1.3°), rendering QPEs over more than half of XTRRA’s coverage area unreliable. QPE RMSE improves over some portions of the sector at higher elevation angles, but this improvement is range-limited by the steep elevation angle. This “spotlight” occultation estimation technique can be applied to radars whose coverage areas have been mapped in detail using airborne lidar. One limitation is that the date of lidar data collection must be precisely known, because blockage from deciduous trees varies substantially by season, and because new building construction can alter the digital terrain model.