Future work and caveats
Our study is unique in its spatial and taxonomic extent, yet the majority of the species used are in Odonata and Lepidoptera. This is not surprising, since these clades have relatively large body sizes and are easy to see and photograph. In addition to the taxonomic biases in our dataset, the 25 x 25-km grid cells that were included in our analyses were biased geographically, with most cells occurring in metropolitan areas in the Eastern Temperate Forests ecoregion and along the California Coast. There are noticeable sampling gaps in the Great Plains, North American deserts, and the complex and diverse ecoregions of southern Mexico. Finally, we note that increasing observer effort, measured as the number of observations, did show a significant, albeit relatively weak effect on phenology estimates. Although our modeling framework attempted to control for these effects, care must be taken when fitting estimations using community science datasets. As the amount of available incidental data continues to grow, we expect better spatial, and phylogenetic resolution, albeit with biases still likely, unless concerted efforts are made by community scientists to survey more rural locations (Shirey et al., 2021).
Even if more species-specific phenology estimates can be generated for insects, our full ability to understand drivers of insect phenology will be hampered by the lack of available trait data. We had enough incidental intra-annual observations to generate phenological estimates for 154 species with at least five year-by-cell combinations but could only use 101 species because of missing trait data. This is particularly problematic because our results highlight the importance of life history traits in predicting all aspects of adult insect phenology. Therefore, to better understand how insects will respond to climate change and urbanization, it is imperative that continued effort goes into generating, compiling, and archiving openly available insect life history information (e.g., Middleton-Welling et al., 2020). Improving insect trait knowledge and access will allow researchers to better understand ecological processes using the accelerating accumulation of occurrence and other natural history records. This research showcases a framework to gather and use these resources to answer fundamental questions about the duration of adult activity across broad spatial and phylogenetic scales.