A simple but novel approach to calculate seabird light fallout: A case study from Kauaʻi

Abstract

Injury and mortality of night flying seabirds due to lights (fallout) affects their long-term viability. However, the cryptic nature of these species hinders direct estimates and projections of fallout. The Kauaʻi Seabird Habitat Conservation Plan (KSHCP) characterizes the current and projected fallout for Newell's shearwaters (Puffinus auricularis newelli), Hawaiian petrels (Pterodroma sandwichensis) and the Hawaiʻi distinct population segment of the band-rumped storm-petrels (Oceanodroma castro) on the island of Kauaʻi across eight permittees. In the KSHCP, each permittee was required to itemize the lighting infrastructure related to their facilities (type, wattage, number of lights, etc.) as well as supply an overall expected seabird fallout number with an estimate of detection probability or ability to find the fallout birds. Light intensity is thought to be correlated to the degree of attraction to a light source. Here we identify seabird fallout using indirect estimates derived from the KSHCP and present a reproducible method to consistently project and calculate fallout between entities that need to reduce seabird mortality due to their light sources. Though the permittees did not provide an estimate of light intensity at their sites, we extrapolated luminous intensity (light brightness) from their itemized light information (i.e., lumens/watt). Using luminous intensity, we then derived the degree of fallout per lumen and its variation across the permittees. As a proof of concept, we then projected the fallout per lumen estimate onto streetlights owned by the Kauaʻi Island Utility Cooperative, as applied over a range of detectability correction factors. Though fallout varies significantly based on the detectability correction factor used, when considering the site differences on Kauaʻi, a previous fallout estimate, and the overlapping characteristics of the projected versus actual luminous intensities, the utilities projected fallout appears applicable. We feel that the methodology can be used as a baseline to infer fallout associated with various entities.