Drivers of adult insect emergence, termination, and duration
In comparison to other aspects of insect phenology, the drivers of adult emergence are relatively well understood. The emergence of many insect species is controlled by temperature because the growth rate of immature stages increases at warmer temperatures (Gilbert and Raworth, 1996; Hodgson et al., 2011). Life history traits, particularly diapause stage, are also known to be an important predictor of when insects emerge (Scott and Epstein, 1987; Altermatt, 2010b). Our results confirm earlier emergence in warmer areas and in species that diapause as adults.
Less is known about drivers of termination and duration of adult insects. For many insects, photoperiod is likely predictive of activity termination since it is a primary cue in many insects to initiate diapause (Tauber and Tauber, 1976; Denlinger, 2002). Although photoperiod is often the main driver, diapause induction has also been related to temperature, food availability, moisture, and chemical cues (Danks, 2007). Our results also suggest that many factors contribute to the termination of adult activity and indicate that these processes may be linked to temperature seasonality, timing of the first fall freezes, or resource depletion caused for example by drought. In multivoltine insects, we expected longer activity durations in warmer regions, due to the production of additional adult generations in areas with longer growing seasons (Altermatt, 2010a; Zeuss et al., 2017). It is less clear how temperature influences the duration of univoltine insects. Because obligate univoltine species cannot produce additional generations in warmer regions, we expected more consistent activity durations across temperature gradients, regardless of emergence dates. However, we did not find an interaction between temperature and voltinism, suggesting similar temperature-driven increases in duration regardless of voltinism. This surprising result could be explained by multiple mechanisms for extending adult activity, including lengthening timing of activity of cohorts of adults, more generations per season, and reduced adult synchrony of univoltine populations in warm areas. However, this result may be due to our simplistic trait coding system labeling semi-, parti-, and merovoltine species as “not univoltine”. While this does not impact our finding of longer activity periods for univoltine species in warmer regions, more finely scored voltinism states may help elucidate trait-mediated phenology responses to temperature. We also note that our understanding of life history traits, particularly voltinism and migratory behavior, is incomplete across broad temperature gradients. Species that are documented in the literature as non-migratory and univoltine may in fact have undescribed migratory patterns (Robinson et al., 2009) or potentially multiple generations in warmer regions that remain unreported.
One of the challenges with understanding phenological sensitivities is that regional contexts may impact not only the strength of sensitivity but also its direction. Li et al. (2019) showed, for example, that urbanization shifts from advancing flowering in cold regions to delaying in warmer ones. Here we also find strong evidence of these region-specific contextual effects. One of the clearest examples is that the influence of precipitation and life history traits on insect duration varies with regional temperature. Warm, wet areas demonstrated longer adult insect duration than cooler, wetter regions. This interaction is expected, considering that snow contributes to annual precipitation in cool regions and snowmelt date delays the beginning of activity more than the end (Stemkovski et al., 2020). These regional effects also included life history interactions with temperature. We found that detritivores terminated activity later and had longer durations than carnivores, potentially due to detritus being more available later into the fall compared to green leaf materials or prey items. This effect is especially strong in warm environments, where detritus can accumulate year round.
We also expected a complex interaction between temperature and urbanization, similar to the results of Diamond et al. (2014) for butterflies in Ohio, which documented earlier emergence in cold areas with high human population density and delayed emergence in warmer, more urbanized areas. While we found such an interaction, it was in the opposite direction than expected: insects generally emerged later in cool, urbanized areas and earlier in warm, urbanized areas, in comparison to corresponding rural areas. Due to our species selection protocol, our species list likely overrepresented exploitative species since these are commonly observed. Exploitative butterflies showed smaller delays in first appearance in warm and urbanized areas (Diamond et al., 2014), potentially, in part, explaining our differing results. We encourage future work exploring how urbanization impacts insect phenology and at what spatial scales urbanization influences are most apparent.