Effects of the loss of forest cover on odonate communities in eastern Amazonia

In recent decades, the conversion of native forests into anthropogenic landscapes has been impacting progressively the physical habitat of the aquatic systems of Brazilian Amazonia. Given this, we evaluated the effects of the loss of natural vegetation cover on the diversity and morphology of Amazonian odonates (Zygoptera and Anisoptera). To do so, we sampled adult odonates at 50 streams in eastern Brazilian Amazonia. We evaluated land use patterns at each site together with two morphological metrics (length of the thorax and the abdomen) of the odonates. We found a positive relationship between natural vegetation cover and both odonate species richness, in general, and the proportion of zygopteran species richness. However, the relationship was negative for all the anisopteran diversity metrics. We also found only negative relationships in the anisopteran abdomen when we investigated the variation in morphological trait change points. We identified 12 species that were associated with natural forest cover, six (all zygopterans) with more forested environments, and six species (three anisopterans and three zygopterans) that were associated with environments with reduced vegetation cover. The results of the present study indicate that preserving continuous natural vegetation cover together with the riparian vegetation will be crucial to the maintenance of the physical characteristics of streams and the structure of the odonate community. The presence of adult odonates was also associated with the quality of both the aquatic and the terrestrial environments, given that their effects on odonate diversity were detected within a 500 m buffer around the streams. Implications for insect conservation The analysis of landscape metrics may increase the potential of the associations observed between the characteristics of natural communities and the loss of natural vegetation cover, in particular in the case of the forest specialists. The application of analyses based on these metrics should better support the development of more effective mitigation measures.