Modelling epiphytic bryophyte vegetation in an urban landscape

Abstract

Air pollution in the urban environment of Halle, Germany, has significantly decreased over the last decade. We investigated the distribution and recolonization of the epiphytic bryoflora. We tested the hypothesis that after a period of nearly complete absence of epiphytic mosses and liverworts, recolonization would occur first in the most favourable habitats and proceed from there. The aims of the study were to analyse these habitat conditions, to generate habitat models for the most abundant species and to identity favourability of habitats. At the community level, mixed linear models revealed the factors aspect of plots, inclination of trees and landscape features to be of particular relevance to bryophyte diversity and coverage. At the species level, generalized linear mixed models with logit link functions identified aspect, tree inclination and landscape type to be the key habitat factors for most taxa, while identity of the host tree species and stand structure were of minor importance. In general, epiphytic bryophyte vegetation did not attain high coverage and many species occurred only rarely. Thus, significant habitat models were obtained only for frequent species (e.g. Orthotrichum diaphanum) and only with low coefficients of determination. Nevertheless, the results clearly show that species-sorting is an effective mechanism of community assembly even in this early stage of recolonization.