Inconsistent shifts in warming and temperature variability negatively impact avian fitness

Climate change can create challenging conditions for reproduction by creating mismatches between breeding activity, resource availability, and favorable weather. Phenological advancement may mitigate some of these effects, but it can also result in exposing developing offspring to greater temperature variability during reproduction. We combined >300,000 community science breeding records with data on historical temperature to understand how exposure to extreme climatic events (ECEs) has changed for 24 North American songbirds during breeding. More specifically, we wanted to understand if changes in average spring temperature were associated with when the latest cold-snap or earliest heat-wave occurred across the breeding range and the relationship to individual fitness. While average spring temperature increased universally across North America since 1950, the change in timing of ECEs was inconsistent in both direction and magnitude. Moreover, increase in average temperature did not predict the amount of change in average cold-snap and heat-wave timing. Thus, the current relationships between temperature and temperature variability differ from those experienced historically during breeding. After accounting for covariates, we found that 16 of 24 species had reduced reproductive success when a cold-snap occurred during the breeding attempt and 11 of 24 had reduced success when a heat-wave occurred. Four of the five aerial insectivores included showed marked sensitivity to ECEs, as did all the sallying insectivores. For three widespread species, we found evidence for latitudinal differences in susceptibility to temperature extremes. However, the pattern of sensitivity differed for each species, suggesting that life history details might play an important role in species responses to ECEs. Large-scale patterns and consequences of climate change on breeding birds ultimately result from the accumulated effects of individual susceptibility. Our results add to evidence demonstrating that understanding individual and species level sensitivity and linking them to population level processes is critical for predicting vulnerability to changing climates.