Atmospheric carbon dioxide records indicate that the land surface has acted as a strong global carbon sink over recent decades1, 2, with a substantial fraction of this sink probably located in the tropics3, particularly in the Amazon4. Nevertheless, it is unclear how the terrestrial carbon sink will evolve as climate and atmospheric composition continue to change. Here we analyse the historical evolution of the biomass dynamics of the Amazon rainforest over three decades using a distributed network of 321 plots. While this analysis confirms that Amazon forests have acted as a long-term net biomass sink, we find a long-term decreasing trend of carbon accumulation. Rates of net increase in above-ground biomass declined by one-third during the past decade compared to the 1990s. This is a consequence of growth rate increases levelling off recently, while biomass mortality persistently increased throughout, leading to a shortening of carbon residence times. Potential drivers for the mortality increase include greater climate variability, and feedbacks of faster growth on mortality, resulting in shortened tree longevity5. The observed decline of the Amazon sink diverges markedly from the recent increase in terrestrial carbon uptake at the global scale1, 2, and is contrary to expectations based on models. (Résumé d'auteur

Abstract. Forest structure and dynamics vary across the Amazon Basin in an east-west gradient coincident with variations in soil fertility and geology. This has resulted in the hypothesis that soil fertility may play an important role in explaining Basin-wide variations in forest biomass, growth and stem turnover rates. Soil samples were collected in a total of 59 different forest plots across the Amazon Basin and analysed for exchangeable cations, carbon, nitrogen and pH, with several phosphorus fractions of likely different plant availability also quantified. Physical properties were additionally examined and an index of soil physical quality developed. Bivariate relationships of soil and climatic properties with above-ground wood productivity, stand-level tree turnover rates, above-ground wood biomass and wood density were first examined with multivariate regression models then applied. Both forms of analysis were undertaken with and without considerations regarding the underlying spatial structure of the dataset. Despite the presence of autocorrelated spatial structures complicating many analyses, forest structure and dynamics were found to be strongly and quantitatively related to edaphic as well as climatic conditions. Basin-wide differences in stand-level turnover rates are mostly influenced by soil physical properties with variations in rates of coarse wood production mostly related to soil phosphorus status. Total soil P was a better predictor of wood production rates than any of the fractionated organic- or inorganic-P pools. This suggests that it is not only the immediately available P forms, but probably the entire soil phosphorus pool that is interacting with forest growth on longer timescales. A role for soil potassium in modulating Amazon forest dynamics through its effects on stand-level wood density was also detected. Taking this into account, otherwise enigmatic variations in stand-level biomass across the Basin were then accounted for through the interacting effects of soil physical and chemical properties with climate. A hypothesis of self-maintaining forest dynamic feedback mechanisms initiated by edaphic conditions is proposed. It is further suggested that this is a major factor determining endogenous disturbance levels, species composition, and forest productivity across the Amazon Basin.

Aim To examine the contribution of large‐diameter trees to biomass, stand structure, and species richness across forest biomes. Location Global. Time period Early 21st century. Major taxa studied Woody plants. Methods We examined the contribution of large trees to forest density, richness and biomass using a global network of 48 large (from 2 to 60 ha) forest plots representing 5,601,473 stems across 9,298 species and 210 plant families. This contribution was assessed using three metrics: the largest 1% of trees ≥ 1 cm diameter at breast height (DBH), all trees ≥ 60 cm DBH, and those rank‐ordered largest trees that cumulatively comprise 50% of forest biomass. Results Averaged across these 48 forest plots, the largest 1% of trees ≥ 1 cm DBH comprised 50% of aboveground live biomass, with hectare‐scale standard deviation of 26%. Trees ≥ 60 cm DBH comprised 41% of aboveground live tree biomass. The size of the largest trees correlated with total forest biomass (r2 = .62, p < .001). Large‐diameter trees in high biomass forests represented far fewer species relative to overall forest richness (r2 = .45, p < .001). Forests with more diverse large‐diameter tree communities were comprised of smaller trees (r2 = .33, p < .001). Lower large‐diameter richness was associated with large‐diameter trees being individuals of more common species (r2 = .17, p = .002). The concentration of biomass in the largest 1% of trees declined with increasing absolute latitude (r2 = .46, p < .001), as did forest density (r2 = .31, p < .001). Forest structural complexity increased with increasing absolute latitude (r2 = .26, p < .001). Main conclusions Because large‐diameter trees constitute roughly half of the mature forest biomass worldwide, their dynamics and sensitivities to environmental change represent potentially large controls on global forest carbon cycling. We recommend managing forests for conservation of existing large‐diameter trees or those that can soon reach large diameters as a simple way to conserve and potentially enhance ecosystem services.

Abstract. Forest structure and dynamics have been noted to vary across the Amazon Basin in an east-west gradient in a pattern which coincides with variations in soil fertility and geology. This has resulted in the hypothesis that soil fertility may play an important role in explaining Basin-wide variations in forest biomass, growth and stem turnover rates. To test this hypothesis and assess the importance of edaphic properties in affect forest structure and dynamics, soil and plant samples were collected in a total of 59 different forest plots across the Amazon Basin. Samples were analysed for exchangeable cations, C, N, pH with various P fractions also determined. Physical properties were also examined and an index of soil physical quality developed. Overall, forest structure and dynamics were found to be strongly and quantitatively related to edaphic conditions. Tree turnover rates emerged to be mostly influenced by soil physical properties whereas forest growth rates were mainly related to a measure of available soil phosphorus, although also dependent on rainfall amount and distribution. On the other hand, large scale variations in forest biomass could not be explained by any of the edaphic properties measured, nor by variation in climate. A new hypothesis of self-maintaining forest dynamic feedback mechanisms initiated by edaphic conditions is proposed. It is further suggested that this is a major factor determining forest disturbance levels, species composition and forest productivity on a Basin wide scale.

Summary Tree mortality is an important process in forest ecology. We explored the extent to which tropical tree death is a predictable outcome of taxon and individual level properties by means of mixed‐species logistic regression, for trees ≥ 10 cm in diameter. We worked in two lowland forest regions with markedly different floristic composition and dynamic regimes – the high wood density, low‐mortality northeastern (NE) Amazon (in eastern Venezuela), and the low wood density, high‐mortality northwestern (NW) Amazon (in northern Peru). Among those genera that are shared between regions there were no detectable regional differences in mortality rates. This suggests that floristic compositional differences are a major driver of the twofold regional contrast in stand‐level mortality. In NE forests, mortality risk of individual trees is best predicted by low taxon‐level wood density, slow relative growth, and large size, reflecting phylogenetically determined life‐history strategy, physiological stress and senescence. In NW forests, trees with low wood density and slow relative growth are also at most risk, but probability of death is independent of tree size, indicating that senescence is unimportant in this region. Synthesis. This study shows that despite fundamental floristic and dynamic differences between the two Amazonian regions, mortality risk can be predicted with mixed‐species, individual‐based statistical models and that the predictors are remarkably similar, such that tree growth and wood density both play important roles.