River Barriers and Spatial Constraints: Unveiling the Factors Shaping Species Distribution in the Amazonian Rainforest

ABSTRACT Aim We aimed to quantify the relative contributions of geographic distance, environmental variation, and riverine barriers to shaping species distributions across the Amazon forest. Location Amazon forest, Brazil. Time Period Present. Major Taxa Studied Class Arachnida; Order: Opiliones (harvestmen). Methods To disentangle the influence of geography, environment, and barriers directly on individual species, we developed a Bayesian mechanistic model that estimates the effects of these variables on species niches and dispersal. We then demonstrate how the distribution of individual species along these gradients and barriers gives rise to the commonly observed decay in species similarity with distance. We applied the method to understand the distribution of 71 harvestmen species in 105 plots distributed across the Amazon River. Results We identified 71 harvestmen species from 105 plots, with a higher species richness south of the river (55 species) compared to the north (27 species). Only 15% of species were found on both sides of the river. Our analysis estimated that, on average, a species at the river margin has a 43% probability of crossing the Amazon River, and this barrier effect is equivalent to dispersing 172 km through the forest. A substantial portion of species (45%) in the southern region has naturally narrow distributions, with a near‐zero probability of occurrence at the river margins, suggesting their absence from the opposite bank is not directly due to the river barrier. When all species were considered, the river had a stronger effect on species composition than geographic distance. However, after excluding the five most common species (7% of the total), the effect of geographic distance became stronger than the river's effect on species composition. Our model, which integrates these factors, had a higher explanatory power ( R 2 = 0.527) for the observed Sørensen similarity index than a direct multiple regression ( R 2 = 0.409). Main Conclusions Our findings suggest that the results of many previous studies have been biased due to an overemphasis on a minority of species with broad distributions and a limited geographic scope near large rivers. The vast distances and the heterogeneous Amazonian environment have a profound, yet underappreciated, impact on the distribution of individual species and entire communities.