Asymmetric effects of environmental filtering on the assembly of tropical bird communities along a moisture gradient

Abstract

The species-sorting hypothesis (SSH) states that environmental factors influence local community assembly of metacommunities by selecting for species that are well adapted to the specific conditions of each site. Along environmental gradients, the strength of selection against individuals that are marginally adapted to the local conditions increases towards the extremes of the environment where the climate becomes harsh. In rainfall gradients, the strength of selection by the environment has been proposed to decrease with rainfall. Under this scenario SSH would predict that immigration of individuals from the metacommunity should be restricted into the dry end of the gradient creating a positive relationship between immigration and rainfall. However, if the selection is strong in both ends of the gradient, the restriction should be expected to be in both directions such that the ends behave as independent metacommunities even in the absence of geographical barriers. In this study we used models based on neutral theory to evaluate if SSH can explain the distribution of bird species along a steep rainfall gradient in Colombia. We found a strong positive relationship between immigration rates and precipitation suggesting that the dry forests impose stronger challenges for marginally adapted bird populations. However, a two-metacommunity model separating dry and wet forests was a better fit to the observed data, suggesting that both extremes impose strong selection against immigrants. The switch from the dry forest to the wet forest metacommunities occurred abruptly over a short geographic distance in the absence of any apparent geographic barrier; this apparent threshold occurs where the forest becomes mostly evergreen. The relative number of rare species in dry forest was lower than in wet forests suggesting that the selection against marginally adapted populations is stronger in the dry forests. Overall, our analyses are consistent with SSH at the regional scales, but the rarity analysis suggests that the mechanisms at the local scales are substantially different. Based on these results, we hypothesize that abiotic (climatic) factors limit immigration into dry forest communities and whereas biotic factors such as competition and predation may limit immigration into bird communities in the wet forest.