Tropical metacommunities along elevational gradients: effects of forest type and other environmental factors

Elevational gradients provide a natural experiment for assessing the extent to which the structure of animal metacommunities is molded by biotic and abiotic characteristics that change gradually, or is molded by aspects of plant community composition and physiognomy that change in a more discrete fashion. We used a metacommunity framework to integrate species‐specific responses to environmental gradients as an approach to detect emergent patterns at the mesoscale in the Luquillo Mountains of Puerto Rico. Elements of metacommunity structure (coherence, species turnover and range boundary clumping) formed the basis for distinguishing among random, checkerboard, Gleasonian, Clementsian, evenly spaced and nested patterns. Paired elevational transects (300–1000 m a.s.l.) were sampled at 50 m intervals to decouple underlying environmental mechanisms: a mixed forest transect reflected changes in abiotic and biotic conditions, including forest type (i.e. tabonuco, palo colorado and elfin forests), whereas another transect reflected changes in environmental conditions but not forest type, as its constituent plots were located within palm forest. Based on distributional data (presence versus absence of species), the mixed forest transect exhibited Clementsian structure, whereas the palm forest transect exhibited quasi‐Gleasonian structure. In contrast, the distribution of modes in species abundance was random with respect to the latent environmental gradient in the mixed forest transect and clumped with respect to the latent environmental gradient in the palm forest transect. Such contrasts suggest that the environmental factors affecting abundance differed in form or type from those affecting distributional boundaries. Variation among elevational strata with respect to the first axis of correspondence from reciprocal averaging was highly correlated with elevation along each transect, even though axis scores were not correlated between mixed forest and palm forest transects. This suggests that the identity of the environmental characteristics, or the form of response by the fauna to those characteristics, differed between the two elevational transects. Despite the proximity of the transects, the patchy configuration of palm forest, and the pervasive distribution of the dominant palm species, the relative importance of abiotic variables and habitat in structuring gastropod metacommunities differed between transects, which is remarkable and attests to the sensitivity of metacommunity structure to environmental variation.