Effects of Beam Occultation on Radar-Based Precipitation Estimates From the Purdue University X-band Teaching and Research Radar (XTRRA)

The X-band Teaching and Research Radar (XTRRA) at Purdue University is a weather radar located in West Lafayette, IN, 210 meters above sea level. Due to its location, the radar is not tall enough for its beam to clear tall buildings, trees, and terrain at its lowest elevation angles. These obstructions can cause errors in estimated precipitation values when compared to actual ground level measurements. To map the obstructions, we used detailed lidar data collected over Tippecanoe County, Indiana to estimate the percentage of occultation at each elevation, azimuth, and distance, modeling XTRRA’s 1° beam as a spotlight. By masking highly occulted areas of the radar and the respective quantitative precipitation estimates (QPEs), we expected the root mean squared error (RMSE) statistics to improve in comparison to nearby tipping bucket rain gauges. We found that the blockages do indeed cause noticeable QPE inaccuracy at the lowest elevation angles (<= 1.3°) when unmasked. Over some portions of the sector at higher elevation angles, QPE RMSE improves, but the improvement at these high elevation angles is limited.

The spotlight occultation model described can be used on other radars whose coverage areas have high-quality lidar maps, but it is important to know the date of the lidar data collection. One limitation is that blockage from deciduous trees varies by season and new building construction can alter the occultation values in previously unocculted locations.