Threshold responses in bird mortality driven by extreme wind events

Abstract

Changing environmental conditions can have non-linear effects from populations to ecosystems, which could increase following the strengthening of human activities. Investigating the potential role of threshold responses to increasing extreme events could help elucidate the mechanisms conferring resilience to populations and ecosystems, since extreme events could have a higher impact than mean changes in ecological systems. Among the potential impacts of extreme climatic events, the assessment of mortality deserves particular attention for long-lived species. In the NE Atlantic, the vulnerability of birds was emphatically demonstrated in the winter of 2013/2014 when at least 54,982 individuals were reported dead in cumulative mass-mortality events. The aim of the present work is to uncover the link between extreme wind events and bird mortality for the 2004–2014 period in southern European waters to detect potential threshold responses. As extreme wind events can be estimated in different ways, we employed a multi threshold wind speed approach to assess which criteria yielded the strongest predictions. By focusing on a highly vulnerable species, the common guillemot Uria aalge, we hypothesized that bird mortality would be related to a cumulative effect of extreme wind events. At the weekly scale, cumulative patterns of extreme wind events and bird mortality were correlated following similar sigmoidal curves. The period of maximum mortality fell within the time frame of maximum extreme wind events during winter (beginning of December-end of February). We identified a threshold response of bird mortality in relation to extreme wind events above which birds experienced consistent increased mortality, which was especially evident in the heavier winter (e.g. 2013/2014). Our methodological approach provides insights on the definition of the number of extreme wind events that should serve as a warning signal to activate the coordination of local administrations and volunteers to anticipate the occurrence of mass-mortality events under future climatic scenarios.