Near-term ecological forecasting for dynamic aeroconservation of migratory birds.

Kyle G Horton,Daniel Sheldon,Heidi J Albers,Andrew Farnsworth,Benjamin M Van Doren

doi:10.1111/cobi.13740

Abstract

Near‐term ecological forecasting has the potential to mitigate negative impacts of human modifications on wildlife by directing efficient action through relevant and timely predictions. We used the U.S. avian migration system to highlight ecological forecasting applications for aeroconservation. We used millions of observations from 143 weather surveillance radars to construct and evaluate a migration forecasting system for nocturnal bird migration over the contiguous United States. We identified the number of nights of mitigation required to reduce the risk of aerial hazards to 50% of avian migrants passing a given area in spring and autumn based on dynamic forecasts of migration activity. We also investigated an alternative approach, that is, employing a fixed conservation strategy based on time windows that historically capture 50% of migratory passage. In practice, during both spring and autumn, dynamic forecasts required fewer action nights compared with fixed window selection at all locations (spring: mean of 7.3 more alert days; fall: mean of 12.8 more alert days). This pattern resulted in part from the pulsed nature of bird migration captured in the radar data, where the majority (54.3%) of birds move on 10% of a migration season's nights. Our results highlight the benefits of near‐term ecological forecasting and the potential advantages of dynamic mitigation strategies over static ones, especially in the face of increasing risks to migrating birds from light pollution, wind energy infrastructure, and collisions with structures.

Highlights

Knowing when to direct action to protect species and habitats is essential for successful conservation (Knight et al, 2010; Wilson et al, 2005), and there are many examples of such campaigns (e.g., Burgess et al, 2019; Liberati et al, 2019; Luther et al, 2016; Wilson et al, 2019)
Under the dynamic selection scenario, we identified the minimum number of nights of conservation action needed to capture 50% of seasonal activity
At all locations we examined, forecasting resulted in fewer action nights than static, fixed window approaches that captured comparable aerial passage and alerted protective actions

Summary

Introduction

Knowing when to direct action to protect species and habitats is essential for successful conservation (Knight et al, 2010; Wilson et al, 2005), and there are many examples of such campaigns (e.g., Burgess et al, 2019; Liberati et al, 2019; Luther et al, 2016; Wilson et al, 2019). Safeguarding highly dynamic ecological processes, such as movement and migration, poses a greater challenge (Reynolds et al, 2017). The spatial process of migration creates an opportunity to reduce the amount of time during which conservation measures are necessary at any particular location. Ecological forecasting of animal movements at relevant spatial and temporal scales may provide a pathway toward real-time conservation (Dietze et al, 2018; Van Doren & Horton, 2018). Conservation actions relevant to migrating species may include the temporary removal of terrestrial or aquatic barriers (e.g., fences, dams), aerial obstacles (e.g., wind turbines, aircraft), or point-source pollutants (e.g., light pollution, chemical pollution) (Marschall et al, 2011; Naidoo et al, 2012; Van Doren et al, 2017)

Objectives

Methods

Results

Conclusion