Soil texture and functional traits of trees structure communities of epiphytic mosses in a tropical dry forest

Abstract

The question which guided our study was: How environmental variables and trees affect the parameters of diversity, composition and abundance of epiphytic moss species at a local scale? Our hypothesis is that these moss communities are shaped by tree functional traits, which in turn are determined according to soil texture. Therefore, we collected moss from 25 permanent plots distributed along six transects (three in each texture of soil) in a Seasonally dry tropical forest (SDTF) in Northeastern Brazil. We assessed the distribution and abundance of epiphytic moss and tree species separately by performing an ordination analysis for each plot and tested if the spatial pattern of mosses was correlated with that of trees by performing a Procrustes analysis. Structural, functional and physico-chemical variables of trees were measured and subjected to uni- and multivariate analyses as potential abiotic filters. We then described the relationship between the response (richness and abundance) and predictor variables (structural, functional and physico-chemical variables) of the trees. As a result, we found that deciduous tree species were more frequent in sandy soils with less water retention. Leaf shedding by deciduous species reduces the total evapotranspiration surface during the dry season compared to evergreens. In contrast, evergreen trees were more frequent in clay soils, which have a higher water retention capacity. These differences in tree functional groups between soil types promoted different moss epiphytic communities in environments with different soil textures. Despite generalist and sun moss species dominating the study area, we found two groups with different environmental requirements based on the abundance of life forms: tuft and weft in drier areas; and mat and dendroid in wetter areas. Epiphytic moss community composition reflects the environmental conditions for each of the formed groups through the pattern of the trees’ functional traits predicted by different soil textures.