•  
  •  
 

Abstract

Following the August 13, 2011, Indiana State Fair stage collapse tragedy, caused by a wind gust from an approaching thunderstorm, Purdue University enforced a wind speed restriction of 30 mph (13 m s-1) for tents at outdoor events. During these events, volunteers stand outside with handheld anemometers, measuring and reporting when the wind speeds exceed this limit. In this study, we report testing of a new system to automate high-wind alerts based on observations from a Doppler radar, the X-band Teaching and Research Radar (XTRRA), near Purdue’s campus. XTRRA scans over campus at low elevations approximately every 5 minutes. Using XTRRA data collected over its first eight months of operation, we developed an algorithm that generates high-wind alerts whenever observed winds at altitudes below 240 m (the height of Ross-Ade Stadium) exceed the 13 m s-1 threshold. We describe how a combination of median filtering, clutter filtering, and statistical outlier removal mitigated false alarms caused by noise and ground clutter. The high-wind alerts are validated against wind gust observations from a nearby Automated Surface Observing System at Purdue University Airport, known as KLAF. Results indicate that the alerts work well in high-wind events associated with precipitation but less well in high-wind events not associated with precipitation (e.g., frontal passages). This is likely because XTRRA, which has a wavelength of 3 cm, is less sensitive to clear-air echoes than an operational WSR-88D. Following further testing, we envision that these automated high-wind alerts will be distributed to interested parties such as campus event coordinators and safety officials.

Share

COinS