RESUMO
Odonates (dragonflies and damselflies) can indicate the ecological health of aquatic biota within the rich but vulnerable biodiversity of tropical forests. The reaction of odonates to deforestation can be measured by changes in coarse taxonomic ratios. Suborder Zygoptera are thermal conformers susceptible to overheating, having the affinity with shaded, intact sites. Anisoptera have exothermic regulation and better dispersal capacities, suggesting their association with more altered, open environments. Similarly, with an increasing degradation, the proportion of Anisoptera species in assemblages should increase. However, based on the data from different continents, the Zygoptera/Anisoptera ratio may be too simple, strongly biased, and not applicable at the global scale. The main reason is that the most diverse, abundant, and cosmopolitan families, Coenagrionidae (Zygoptera) and Libellulidae (Anisoptera), comprise a great proportion of habitat generalists with high migratory capacity and affinity with open habitats. In this study, we sampled odonates from three bioregions (Indomalaya, Afrotropics, and Neotropics) over the gradient of tropical forest degradation with a comparable sampling effort to assess the suitability of species richness and suborder-based (Zygoptera/Anisoptera) and family-based (Libellulidae/other Anisoptera and Coenagrionidae/other Zygoptera) ratios and their abundance-weighted versions for monitoring tropical forest degradation. Our results show that simple Odonata as well as Zygoptera and Anisoptera richness are poor indicators of the forest biota alteration. Family-level indices weighted by relative abundance, especially those involving Coenagrionidae, were more sensitive to changes in forest conditions compared to suborder-level indices. Collectively, our results suggest that for biomonitoring, where financial resources and time are commonly critical, family-level ratio metrics may be effective tools to indicate even slight alterations of aquatic biota resulting from forest degradation. Although these indices have the potential for broader application, their effectiveness across tropical bioregions warrants further validation.