Environmental drivers of tree community turnover in western Amazonian forests

摘要

A more comprehensive understanding of the factors governing tropical tree community turnover at different spatial scales is needed to support land-management and biodiversity conservation. We used new forest inventory data from 263 permanent plots in the Carnegie Biodiversity-Biomass Forest Plot Network spanning the eastern Andes to the western Amazonian lowlands of Peru to examine environmental factors driving genus-level canopy tree compositional variation at regional and landscape scales. Across the full plot network, constrained ordination analysis indicated that all environmental variables together explained 23.8% of the variation in community composition, while soil, topographic, and climatic variables each explained 15.2, 10.9, and 17.0%, respectively. A satellite-derived metric of cloudiness was the single strongest predictor of community turnover, and constrained ordination revealed a primary gradient of environmentally-driven community turnover spanning from cloudy, high elevation sites to warm, wet, lowland sites. For three focal landscapes within the region, local environmental variation explained 13.4-30.8% of compositional variation. Community turnover at the landscape scale was strongly driven by topo-edaphic factors in the two lowland landscapes examined and strongly driven by potential insolation and topography in the montane landscape. At the regional scale, we found that the portion of compositional variation that was uniquely explained by spatial variation was relatively small (2.7%), and was effectively zero within the three focal landscapes. Overall, our results show strong canopy tree compositional turnover in response to environmental gradients at both regional and landscape scales, though the most important environmental drivers differed between scales and among landscapes. Our results also highlight the usefulness of key satellite-derived environmental covariates that should be considered when conducting biodiversity analyses in tropical forests.