30% vs 60% There is confusion about which estimate to use in describing the effectiveness of the prediction system developed in a study of whether managing window lighting at the McCormick Center in Chicago. I use 30%.
Citation: Drivers of fatal bird collisions in an urban center Benjamin M. Van Doren ,PNAS 2021 Vol. 118 No. 24
The 60% figure appears in the abstract of the McCormick Center study published in 2021. It was used in the promotional material when the Lights Out progam was launched by the Cornell Bird Lab. It does not reflect how the Lights Out system works. The author makes clear that the reduction can reach 60% only if all lights are shuttered for the entire migration period. The first quote below descibes ou how the the Lights Out system works. It only issues alerts when its system predicts that there will be major migration event. The second quote below is from the McCormick Center study. It shows that if lights are shuttered only on nights when high migration is predicted, the Lights Out system saves only 27-30%, not the 60% in the abstract. Dr. Van Doren is very careful to explain the difference between the 60% and the 30% figure in the text.
—————————–
(Lights Out Describes How It Works https://birdcast.info/science-to-action/lights-out/alerts/
“We categorize forecasts as low, medium, or high based on historic data (weather variables and bird migration amounts from radar) for the selected location. A forecast triggers an alert whenever a prediction falls within the “high” category. “
Drivers of fatal bird collisions in an urban center Benjamin M. Van Doren ,PNAS 2021 Vol. 118 No. 24 On page
It may not be feasible to extinguish lighting every single night, so we quantified the predicted decrease in mortality if lighted window area had been reduced only on the nights with the largest 25% of migration events. In this scenario, we expect a decrease of 32% (95% CI [26, 38]) in collisions in spring and27% (95% CI [22, 31]) in fall. (Page 4)