Assessment of the dark diversity's ability to predict the absence of Zygoptera (Odonata) species sensitive to anthropogenic disturbance in human-altered Amazonian ecosystems.

Estimates of dark diversity, species that belong to a given species pool but are not present locally, can help to understand how environmental conditions influence species distribution. However, it remains uncertain whether dark diversity can predict the absence of indicator species in preserved environments after environmental changes. We explored the sensitivity of dark diversity (the set of species absent from a particular area), in detecting the absence of Zygoptera (Odonata) indicative of preserved forest environments in altered habitats, and the influence of sample coverage on the detected patterns. We sampled 98 streams in the Amazon region, where the dark diversity of Zygoptera was estimated based on probabilistic species co-occurrence patterns using the Beals index, encompassing 16 species in the Santarém/Belterra region and 23 species in the Paragominas region. The mean total richness of observed Zygoptera species in Paragominas, 42.7 species, and Santarém/Belterra, 25.93 species, was higher than the estimated mean dark diversity for the two study sites, which were 12.32 and 12.20 species, respectively. The dark diversity was not effective in detecting the absence of forest-indicator Zygoptera in human-altered streams and exhibited a positive relationship between Zygoptera dark diversity and species common to different environments concerning environmental integrity only in Paragominas, but not in Santarém. We found that there is a possibility that observed values of dark diversity may occasionally be subject to sampling coverage biases. In this regard, we suggest considering sampling coverage alongside field-collected biological datasets to assess whether it is related to dark diversity.