Discussion
These data confirm the multifaceted changes in environmental conditions along Himalayan rivers over their large altitudinal range (Manel et al. 2001), in turn accompanied by pronounced variations in the community composition and trait character of river birds. Functional distances between co-existing species decreased with increasing elevation after controlling for species richness such that only a subset of traits persisted (Fig. 2). Communities at higher altitudes shifted towards species with smaller bodies, shorter tarsi, smaller bills and a greater tendency to feed as insectivores in the ripariang zone or on aquatic prey. These patterns are consistent with the hypothesis that altitudinal trends affect these communities through environmental filtering (Dehling et al. 2014; Vollstädt et al. 2017; Hanz et al. 2019). They also echo similar filtering effects for example, on ground beetles along a land disturbance gradient (Ribera et al. 2001), plants along a salinity gradient (Pavoine et al. 2011), ants along a complexity gradient (Weischer et al. 2012), bats across a gradient of forest fragmentation (Farneda et al. 2015) and birds with urbanization (Evans et al. 2018). Beyond these filtering effects, however, phylogenetic dispersion also declined with increasing altitude, illustrating for the first time that bird species composition along Himalayan rivers is constrained by phylogenetic origins: passerines, and specifically muscicapids or their near-relatives, dominated higher altitude rivers.
A range of caveats affect interpretation in studies like ours where survey data are used to test hypotheses. Above all, the evolutionary phenomena implied in our analyses occur over temporal and spatial scales that preclude straightforward experimentation. Large-scale surveys of this type provide one of the few pragmatic methods of capturing large-scale phenomena, but need appropriate design to eliminate potential confounds as well as data that corroborate the ecological or evolutionary processes inferred from correlations (Manel et al. 2000). In support of our approach, our design involved surveys that were replicated across regions, and observations that we took as robust representations of past evolutionary processes – such as phylogenetic relatedness or trait expression. Nevertheless, there are well-known challenges in understanding how trait data or phylogenies reflect ecological processes (Cadotte et al. 2019). Furthermore, functional approaches fail to account for within-species variations across populations while the phylogenetic approach can inflate signals related to certain traits (Zhao et al. 2020). At a more empirical level, some parts of our analysis would have been improved by more detailed data. Feeding traits, for example, were represented only crudely by categorisations of prey use, yet river birds can make precise selection for different prey types when foraging. This includes targeting prey of specific size, elemental composition, accessibility and ease of handling (Ormerod and Tyler 1991a). Potentially even more important in the context of niche use and limiting similarity is the extent to which riparian and river birds in the Himalayan mountains use subtly different components of the available prey base. This is apparent among the common insectivores whose diet appears to be partitioned along dimensions of prey size, taxonomic composition and capture method (aerial vs terrestrial vs aquatic) (Buckton and Ormerod 2008). However, incomplete dietary information from several of the species in our study precluded more detailed dietary assessment. In similar vein, measurements of the availability or abundance of prey used by any of the species were beyond the capabilities of this study even though prey abundance is known to influence the density of river birds (Ormerod 1985; Ormerod and Tyler 1991b). A further limitation is that both our field surveys and data analysis focussed on the breeding season, yet several of the species in our study are altitudinal migrants that descend to lower elevations in winter. As a consequence, our investigation is likely to have reflected evolutionary effects during the breeding period when resources demands and selection pressures are likely to be large (Verhulst and Nilsson 2008).
Notwithstanding these caveats, our study revealed clear relationships among river character, species traits and community composition of river birds in the Himalayan mountains aided by the increasingly used RLQ analysis (Ribera et al. 2001). Here, over the largest altitudinal range on Earth, species composition and trait representation changed dramatically as several species of kingfishers, River Lapwing, Common Sandpiper, Blue Whistling-thrush and Spotted Forktail gave way at higher altitude to a generally smaller, insectivorous and functionally clustered array of species such as Plumbeous Water Redstart, White-capped Water Redstart, Brown Dipper and Little Forktail. The latter group of passerines that breed along high elevation river reaches come from several genera (Fig. 1(c)) and are morphologically adapted for different foraging techniques such as fly catching, ground gleaning and aquatic foraging in aquatic, bankside and riparian habitats. Besides tracking the trends in habitat structure and vegetation pattern assessed here, these community changes also reflect well known altitudinal trends in temperature, nutrient status, oxygen concentrations, discharge patterns and sediment regimes that have major effects on fish densities, invertebrate abundances and other factors influencing prey availability (Ormerod et al. 1994). The resulting heterogeneity in habitat character and productivity in this region has given rise to the greatest diversity of specialist river birds on Earth in which selective habitat use, foraging methods and niche partitioning are consistent with resource segregation (Buckton and Ormerod, 2002; 2008). These established patterns add to the support from our data for the hypothesis that river habitat templates have influenced trait distributions within river bird communities through the evolutionary history of the species involved.
In addition to major altitudinal trends in community composition and trait expression among Himalayan river birds, we found that species were assembled non-randomly along the elevation gradient into communities with distinct phylogenetic origins and functional character. Although reflecting patterns among a small group of species, the strength of this phylogenetic signal implies that historical contingency has influenced trait-environment relationships and river bird communities in the Himalaya, particularly at high altitude. It is particularly noteworthy that three of the four species most abundant at high altitudes were Muscicapidae - Plumbeous Water Redstart, White-capped Water Redstart and Little Forktail – an Old-World family with large richness across the Himalayan region in general (Sinha 2021). When communities of organisms are shaped predominantly by environmental conditions, their composition is typically aggregated by similar trait compositions in similar habitats, irrespective of evolutionary history or phylogenetic relatedness (Southwood 1977, Poff 1997). This contrasts with our case where a strong phylogenetic signal in composition and functional traits reflected circumstances where related species co-occurred because of shared environmental requirements, similar general morphology and behavioural character (Webb et al. 2002). Assuming that trait values (body mass, bill length and tarsus length) reflected niche occupancy, the strong phylogenetic signals in our data suggest that functional traits and niche occupancy were constrained by phylogeny, at least at higher elevation. Interestingly, however, some aspects of trait expression departed from the expectations of phylogenetic effects more than others: phylogeny was reflected in body mass, bill size and tarsus length more than in body size, tail length and breeding traits (Table 1). We suggest that community assembly in high altitude river birds must therefore reflect a blend of phylogenetic constraint and habitat filtering coupled with some proximate niche-based selection of trait character for specialization (Reif et al. 2015; Morelli et al.2019). This effect is particularly well illustrated in the forktails (Enicurus spp.), in which tail length in the Little Forktail is substantially reduced in comparison to its congeners and potentially linked to its highly specialised foraging niche around the splash zone of large boulders in highly turbulent flows (Buckton and Ormerod 2008). Similarly, White-capped Water Redstart and Plumbeous Redstart contrast in body size, with smaller size in the latter potentially facilitating energy efficiency in its extensive use of aerial foraging. Further detailed assessments of trait expression and function among Himalayan river birds would prove interesting.