Phenological indices of avian reproduction: Cryptic shifts and prediction across large spatial and temporal scales

Background/Question/Methods

Climate change-induced shifts in phenology have important demographic consequences, and are frequently used to assess species’ sensitivity to climate change. Therefore, developing accurate phenological predictions is an important step in producing mechanistic models of species' responses to climate change. The ability of such phenological models to predict effects at larger spatial and temporal scales has rarely been assessed. It is also unclear whether the most frequently used type of phenological index, namely the average date of a phenological event across a population, adequately captures phenological shifts arising from changes in the distribution of events across the season. We use the single-brooded long-tailed tit Aegithalos caudatus as a case study to explore these issues. We use an intensive 17-year local study to assess how climate influences (a) breeding initiation date, (b) termination date, (c) timing of re-nesting, and (d) breeding season length. These models take predation timing/intensity into account. We test whether the standard phenological index, i.e. population average lay date, detects phenological shifts revealed by the more detailed indices. Finally, we test the capacity of a local climatic model of timing of breeding to predict phenology at much larger spatial and temporal scales.

Results/Conclusions

The local population’s mean lay date and initiation date were primarily determined by March temperature and showed no temporal trend. In contrast, termination date was primarily determined by April temperature, probably due to an earlier decline in food availability in years with warmer Aprils. Termination date has advanced by 16 days over the past 17 years, leading to a 33% reduction in average breeding season length – a substantial loss of reproductive opportunity. The standard phenological index of mean population lay date did not detect these key phenological responses that could have important demographic impacts. There was no evidence for microevolution in phenological traits over the study, suggesting phenotypic plasticity is responsible. The local and national models of lay date in response to March temperature were statistically indistinguishable, and the locally-derived phenological model performed well in predicting previous mean lay dates at the national scale. We thereby show that intensive local studies capturing a broad range of phenological responses can provide useful inference at much larger spatial scales. Furthermore, the local model’s predictive capacity did not decrease further back in time or at higher temperatures, indicating that temporal extrapolation from relatively short-term studies (17 years in this case) is possible.