Contribution of species abundance and frequency to aboveground forest biomass along an Andean elevation gradient

Aims: To determine whether species that contribute most to a plot's biomass are the most abundant (high local abundance at plot scale) or the most frequent (occur the most across plots at landscape scale), or both. In the tropical Andes, these patterns may change with elevation. This study assesses the contribution to plot's above-ground biomass (AGB) of the plant community abundance pattern –the prevalence of within-plot dominant species– and the over-occurrence of regionally frequent species, in an elevation gradient. Methods: We considered all trees ≥2.5 cm DBH from 446 0.1 ha plots in an Amazonia-Andes 260–4350 m elevation cline in N Bolivia. Plot AGB was calculated as the sum of AGBs for all stems contained. We grouped plots into four bins segregated by elevation and ran a bootstrap analysis over subsets of 58 random plots per bin with 100 iterations. Simpson evenness index (E_D) for all species in each plot was used as a measure for its species abundance. Values for each plot's species frequency was calculated as the mean of all species’ in the plot mean frequencies across the bin (i.e. the fraction of plots where each species occurs). We used linear models to correlate plot AGB with (1) elevation and mean annual precipitation (MAP), and (2) E_D, plot species frequency and elevation. We performed all analyses at the species, genus and family levels. Results: Plot AGB was related negatively with elevation, and thus positively with MAP, and also negatively with plot E_D and plot species frequency, all significant. Plot species abundance therefore contributes positively to explain the relationship with AGB along elevational gradients, while plot species frequency does so negatively (i.e. less frequent species contribute more to a plot's AGB across elevation). AGB, for both generic and familial levels was also significantly and negatively correlated with E_D, but not related with plot species frequency biomass at these taxonomic levels. Conclusions: Plot AGB was mainly associated with elevation and floristic composition where species, genera and families tended to be abundant at the local (plot) scale. Species that were less frequent at the regional scale contributed with more AGB regionally, while frequency at generic and familial scales did little to explain AGB patterns. This association seems stronger at lower elevations for all taxonomic levels while decreases toward higher elevation. Our study reveals a relationship between plot structural features like C stocks –influenced by species local abundances– and the distribution of taxa across the landscape.