Wind Project Siting & Environ. Compliance 2018

Sessions Speakers Buzz

POSTER: Machine Vision Field Survey Data and Implications for Risk Modeling

20 Mar 18
5:00 PM - 6:00 PM

Tracks: Eagle Policy, Biology, and Permits, Poster Presentation

Machine vision has been shown to detect more eagles than are observed by typical human point-count protocols. Bayesian eagle fatality risk modeling is driven by empirical data that uses measured eagle presence by human point-counts, measured eagle fatality occurrences, and from these an empirical eagle fatality collision risk probability distribution is derived. If human point-counts underestimate eagle presence, this necessarily implies the collision risk per unit of exposure is less than previously considered. Thus, machine vision data and point-count data are apples and oranges. This paper examines this conundrum and suggests how higher quality and quantity machine vision data can be incorporated into a revised Bayesian modeling protocol.