Annual Global Mean Temperature explains reproductive success in a marine vertebrate from 1955-2010

5 years ago

Annual Global Mean Temperature explains reproductive success in a marine vertebrate from 1955-2010

Donald C. Dearborn, Robert A. Mauck, Charles E. Huntington

The salient feature of anthropogenic climate change over the last century has been the rise in global mean temperature. However, global mean temperature is not used as an explanatory variable in studies of population-level response to climate change, perhaps because the signal to noise ratio of this gross measure makes its effect difficult to detect in any but the longest of datasets. Using a population of Leach's storm-petrels breeding in the Bay of Fundy, we tested whether local, regional, or global temperature measures are the best index of reproductive success in the face of climate change in species that travel widely between and within seasons. With a 56-year dataset, we found that Annual Global Mean Temperature (AGMT) was the single most important predictor of hatching success, more so than regional sea surface temperatures (breeding season or winter) and local air temperatures at the nesting colony. Storm-petrel reproductive success showed a quadratic response to rising temperatures, in that hatching success increased up to some critical temperature, then declined when AGMT exceeded that temperature. The year at which AGMT began to consistently exceed that critical temperature was 1988. Importantly, in this population of known-age individuals, the impact of changing climate was greatest on inexperienced breeders: reproductive success of inexperienced birds increased more rapidly as temperatures rose and declined more rapidly after the tipping point than did reproductive success of experienced individuals. The generality of our finding that AGMT is the best predictor of reproductive success in this system may hinge on two things. First, an integrative global measure may be best for species in which individuals move across an enormous spatial range, especially within seasons. Second, the length of our dataset and our capacity to account for individual- and age-based variation in reproductive success increase our ability to detect a noisy signal. This article is protected by copyright. All rights reserved.

Publisher URL: http://onlinelibrary.wiley.com/resolve/doi

DOI: 10.1111/gcb.13982