Aim: Tropical forests are highly diverse at many spatial scales. In these forests, epiphytic bryophyte communities can be species-rich already within a few cm², and their species numbers increase when expanding the sampling along the tree and the forest. Understanding how this diversity increase depends on scale and position within the tree is critical to evaluate the processes that maintain biodiversity. We, therefore, studied vertical zonation and alpha and beta diversity of epiphytic bryophytes across spatial scales ranging from 100-cm² quadrats to 24 trees up to 10 km apart.

Location: Tropical lowland forest in Yasuní National Park (Amazonian Ecuador).

Methods: We sampled epiphytic bryophytes in 100-cm² quadrats on 24 trees (15-22 sampling quadrats per tree), using a spatially hierarchical design. We applied ordination, indicator species analysis, linear mixed models, and beta partitioning to study alpha and beta diversity at different spatial grains and extents.

Results: At the scale of quadrats, tree crowns held more bryophyte species than trunks. Contrastingly, all crown quadrats together held fewer species than all trunk quadrats together (77 vs. 93 species), as species turnover among trunk quadrats exceeded that among crown quadrats. Across spatial extents, species turnover consistently dominated beta diversity over nestedness. Beta diversity was higher between crowns and trunks within trees than among crowns of different trees at all studied scales, while beta diversity among trunks exceeded vertical diversity when considering all 24 trees.

Main conclusions: These bryophyte communities experience two distinct vertical zones, trunks and crowns, which differ more among them than trees up to 10 km apart. Still, high species turnover makes their species composition highly unpredictable, especially on trunks. Chance processes appear to play an important role in shaping these communities, although analyses at even finer spatial grains may identify the importance of fine-scale habitat filters and species interactions.

Community species matrix derived from 438 samples of non-vascular epiphytes, 240 collected in the medium canopy and 198 along the vertical gradient drawn between the ground level and the canopy of 12 valley trees and 12 ridge trees. These 24 trees were located in clusters of 6 trees (3 valley trees and 3 ridge trees) at 4 distances (166, 500, 820, 5986 m) from the Tiputini river in the Yasuni National Park (Orellana, Ecuador), near the Yasuni Scientific Station of the Pontificia Universidad Catolica de Quito (PUCE).

We sampled non-vascular epiphytic communities established in the middle canopy (sensu Johansson, 1974) and on the tree trunk along the vertical gradient from the ground level to the first living branch of the crown, by using adapted climbing techniques (Perry, 1978). The sampling protocol consisted of collecting the non-vascular epiphytes found in 100-cm² quadrats randomly selected from a larger set of quadrats used to estimated epiphyte cover (ten out of 20 quadrats in the medium section of crown branches, and one out of four quadrats per distance from the tree base to the height of the first living canopy branch). Thus, our survey can be considered a representative sampling of the epiphytic bryophyte diversity of these tropical lowland forests (Gradstein et al., 2003). We sampled 24 trees, with 240 samples collected in the canopy layer (24 trees × 10 samples), and 204 samples collected in tree trunks (the number of samples collected per tree trunk varied between 5 to 12 depending on the height of the first canopy branch, which averaged 14 m).

Collected samples were air-dried and transported to different herbaria. 

Sample height (from the ground level), tree descriptors (DBH, total height, height of the first branch, average height of the canopy, taxonomic identity), and tree location (latitude, longitude, and elevation) are environmental descriptors. These were recorded at the tree and the quadrat level. Tree descriptors were valley/ridge position, distance from the Tiputini river (as a categorical variable identifying the 6-tree clusters), diameter at the breast height (DBH, cm), height (m) and average crown radius (m), a simple field measurement (Zhu, Kleinn, & Nölke, 2021) used as a proxy of crown size. The effect of tree species could not be analyzed, as most tree species were represented only once in our sample (Table S1). Quadrat descriptors were position on the trunk or in the crown and the height from the tree base (m). To estimate canopy cover, pictures were taken with a mobile phone and processed in Gap Light Analyzer (GLA Version 2.0; Frazer, Canham, & Lertzman, 1999) to estimate canopy openness, as an indicator of light availability.

Analysis

Ordination (NMDS), Indicator Species (ISA), Beta diversity partitioning, heatmaps based on Sörensen dissimilarity, and cluster analysis based on Euclidean distances.

R code compiled in an RMD script.