Temporal differentiation in environmental niche modeling of Nearctic narrow-winged damselflies (Odonata: Coenagrionidae)

Narrow-winged damselflies (Odonata: Coenagrionidae) can be observed in a variety of habitats, by both professional collectors and amateur odonatologists. Their abundance and ease of recognition has resulted in a large amount of occurrence data, which can be used to establish species distribution maps through environmental niche modeling. Distributional models often aim to maximize the quantity of occurrence points and environmental variables to relate to the distribution, neglecting both the quality and overlap of these two datasets when generating the models. In order to examine the effects of temporal data and environmental variables influencing change in species distributions, we used occurrence data for twelve species of Coenagrionidae damselflies to generate niche models separated by time periods of specimen collection. Our study examines environmental niche models generated for four time periods for each of these coenagrionid species: Amphiagrion abbreivatum (Selys,1876), Enallagma civile (Hagen,1861), Chromagrion conditum (Hagen in Selys, 1876), Nehalennia gracilis Morse, 1895, Enallagma hageni (Walsh, 1863), Hesperagrion heterodoxum (Selys, 1868), Nehalennia irene (Hagen, 1861), Argia moesta (Hagen, 1861), Ischnura ramburii (Selys, 1850), Argia tibialis (Rambur, 1842), Argia translata Hagen in Selys, 1865, and Argia vivida Hagen in Selys, 1865. The best supported models in each analysis were generated with occurrences of specimens collected from the 1970s to 2000s, and we used occurrence data outside of this range, from the 1800s to 2017, to compare the consistency of model predictions based on specimens of different time periods. In this approach, combining traditional environmental niche modeling and analysis of the specimen records themselves, we have found that ranges for narrow-winged damselflies expand over time, with increase in distributional coverage and decrease in model strength without temporal overlap between occurrences and environmental variables.