From swamps to software: a comparison of field and modeled dragonfly richness across Southeast Australia (Odonata: Anisoptera) 00

Abstract

Biodiversity monitoring is crucial for detecting species declines and informing conservation efforts; however, traditional field-based surveys are constrained by time, resources, and geographic scale. Ecological Niche Modeling (ENM) provides an alternative using large citizen-science datasets but requires validation against empirical data. Here, we compare field-collected and ENM-derived estimates of species richness for dragonfly (Odonata: Anisoptera) species across southeastern Australia, assessing concordance and strength of each approach. We conducted field surveys at 42 localities across New South Wales and Victoria, collecting 476 individuals representing six families, 18 genera, and 33 species. Using occurrence records from GBIF, we generated a stacked richness map, derived from individual species ENMs. While modeled richness predominantly overestimated richness at field sites, we observed two instances of oversampling in field data, where richness exceeded model predictions. Field-based richness exhibited no significant relationships with elevation, latitude, or longitude, most likely due to limited sampling effort and spatial coverage. In contrast, ENM-based richness declined significantly with elevation and increased with latitude and longitude. However, rank-based correlations found significant associations between field-based and ENM-based richness estimations, suggesting broadscale patterns can still be estimated from field data. Our data highlights the complementary value of field surveys and ENM, in which broadscale richness gradients can be identified, while also capturing local-scale variation and validation of predictions.

Keywords: Dragonfly, elevation, field-sampling, latitude, linear models, longitude, niche modeling, species richness

Issue section: Original Article