Environmental Suitability Reveals Under-Surveyed Refugia for Endemic Damselflies in an Afromontane Region

Abstract Freshwater biodiversity is declining globally, yet many endemic freshwater species remain poorly documented, particularly in tropical montane regions where survey effort is often limited. In the Eastern Highlands of Zimbabwe, several endemic Odonata species of conservation concern are known from only a handful of historical records, making it difficult to determine whether their apparent rarity reflects genuinely restricted distributions or inadequate sampling. In this study, we integrated ensemble and stacked species distribution modeling approaches to identify suitable habitats, potential refugia, and under-surveyed areas for three endemic Odonata of conservation concern ( Chlorolestes elegans, Platycypha inyangae , Pseudagrion vumbaense ) from the eastern highlands of Zimbabwe. Our comparison of predicted suitability with historical sampling data showed several areas of suitable habitat outside protected areas that have received little or no historical sampling, particularly for C. elegans in the northern region. Stacked model outputs further showed that areas of high predicted richness and endemism were concentrated in high-elevation upper catchments, with several falling beyond existing protected area boundaries. Environmental conditions at known localities of two critically endangered data-deficient species, Africallagma cuneistigma and Elattoneura lapidaria , were characterized by high elevation, high precipitation, and low temperature, conditions that overlap with those predicted as suitable for the three modeled species. This correspondence indicates that model outputs can inform targeted surveys even where species-specific modeling is not feasible. By combining species distribution models with historical sampling data, this study provides a framework for prioritizing future surveys of endemic Odonata and strengthening freshwater biodiversity assessments in the Eastern Highlands biodiversity hotspot. Implications for insect conservation Our results show that large areas of suitable habitat for threatened and endemic montane Odonata remain historically under-surveyed, and that apparent rarity may reflect sampling gaps rather than true absence. Integrating species distribution models with sampling effort has potential to improve conservation assessments and guide targeted surveys for these species and other data-limited freshwater insects across Afromontane biodiversity hotspots.