Neutral‐based processes overrule niche‐based processes in shaping tropical montane orchid communities across spatial scales

Abstract

Tropical montane forests (TMF) are characterised by high endemism, species richness and turnover across elevations. A key question is how niche‐based processes, via adaptation of species to local environmental conditions, and neutral‐based processes from dispersal limitation shape community composition at different spatial scales across human‐modified landscapes. We expect that (1) communities are highly distinct even within the same habitat type and (2) niche‐based processes play the main role in compositional turnover. To address these expectations, we investigated the compositional turnover of orchids, one of the most prominent floristic elements of TMFs. We sampled orchids in 332 plots spanning over 270 km in the eastern Colombian Andes. Plots ranged between elevations of 1160–3750 m. We used two different spatial extents (whole gradient and two elevational bands), two grains of analysis for the first expectation (regional and local) and two spatial grains for the second expectation (broad and fine scales based on Moran's Eigenvector Maps [MEMs]). We found 331 orchid species in 171 (51%) plots. We found a strong pattern of high compositional turnover across scales (>72% of total beta diversity is given by species turnover), with 87.5% of the total species pool occurring in fewer than five plots, supporting our first expectation. Contrary to our second expectation, we found that community composition is mainly driven by geographical distance, while the relative influence of elevation, environmental variables and their combined fractions were negligible across habitats and spatial scales, rejecting niche‐based processes. Synthesis. High compositional turnover, even across habitats with the highest degree of human intervention, suggests that both forest‐dwelling and open‐habitat species do not easily disperse across habitats. Species dispersal is the major force of orchid community turnover and might be strongly dependent upon macroevolutionary processes and species life‐history traits over multiple scales. Dispersal limitation also draws attention to the importance of preserving habitat connectivity to halt biodiversity losses.

The origin and speciation of orchids

“…Contrary to the findings of Pérez‐Escobar et al . (2017), the northern Andes do not appear to host the fastest speciating orchid clades in the American tropics despite being one of the most species‐rich areas world‐wide (Pérez‐Escobar et al ., 2022; Parra‐Sánchez et al ., 2023). Instead, southern Mesoamerica, comprising the moist and seasonal forests of Costa Rica and Panama, has the highest orchid speciation rates per grid cell, matching its outstandingly high levels of species richness also of other groups (Myers et al ., 2000; Mittermeier et al ., 2011; Crain & Fernández, 2020; Fig.…”

“…Because of the overall topological congruence between the ITS-and matK-derived trees and the Angiosperms353 low-copy nuclear MSC trees, as well as a goodness-of-fit test (Balbuena et al, 2013;P erez-Escobar et al, 2016, see Methods S1; Notes S1), we proceeded to compute ML phylogenetic trees from: a supermatrix derived by concatenating the ITS and matK sequences; and a supermatrix derived from the concatenation of the Angiosperms353 loci, again using RAXML v.8.0 with the same settings previously specified, considering each of the two supermatrices as a single partition. Lastly, we inferred a consensus network using the 500 RAXML bootstrap replicates produced from the Angiosperms353 supermatrix and SPLITSTREE v.4.0 (Huson & Bryant, 2006), with a mean edge weight and a threshold value of 0.75 to filter out any splits not found in at least 75% of the bootstrap trees.…”