Powdermill tunnel scores 2010-2011
Description
As many as a billion birds die each year from collisions with glass, in the U.S. alone. To reduce mortality, it is essential to have a way to test the effectiveness of proposed solutions. Ideally, with enough money and time, glass killing large numbers of birds would be monitored for several years, the glass would be replaced with one incorporating a signal intended to deter birds, and monitoring would take place again. However, this is much too slow a process, even if it were practical. We used a binomial choice protocol, a 10-meter-long ‘tunnel’ installed at the Powdermill Avian Research Center of Carnegie University, which has had a banding station for over 50 years. The tunnel is operated during peak migration, essentially the months of May, September and October. Birds meeting the criteria for testing are brought, after banding, to the tunnel and placed into its dark interior. At the far end, birds see daylight and two possible exits. A net at the end of the tunnel stops birds from hitting glass, which is mounted about 0.5 meters past the end of the tunnel itself. Glass samples measure 0.5 x 1.0 meters. If two clear glass samples are used, fifty percent of the birds fly to the left and fifty percent to the right. We use this test to ensure that the tunnel apparatus is not biased. We can test materials like glass with patterns created with ceramic frit, glass with patterns created with decals, window film, tape, paint and more. Test panels are switched with control glass on a regular basis, and each material is tested at different times of day. Each bird is videotaped while in the tunnel. If a bird fails to fly for 30 seconds, it is removed and released. For each flight, we record species code, band number, date and time, test panel identification, location of test panel (left or right side), whether flight was towards glass on left or right – or if bird landed on ceiling, floor or wall, in the data spreadsheet. The more birds fly towards the control, as opposed to the test panel, the more we presume they consider the test panel to be an obstacle. We compile data and define a ‘tunnel score’ as (#flights to control/total left and right flights) *100. Flight videos are reviewed daily, to ensure accuracy.