Aim Large trees (d.b.h.≥70 cm) store large amounts of biomass. Several studies suggest that large trees may be vulnerable to changing climate, potentially leading to declining forest biomass storage. Here we determine the importance of large trees for tropical forest biomass storage and explore which intrinsic (species trait) and extrinsic (environment) variables are associated with the density of large trees and forest biomass at continental and pan-tropical scales. Location Pan-tropical. Methods Aboveground biomass (AGB) was calculated for 120 intact lowland moist forest locations. Linear regression was used to calculate variation in AGB explained by the density of large trees. Akaike information criterion weights (AICc-wi) were used to calculate averaged correlation coefficients for all possible multiple regression models between AGB/density of large trees and environmental and species trait variables correcting for spatial autocorrelation. Results Density of large trees explained c. 70% of the variation in pan-tropical AGB and was also responsible for significantly lower AGB in Neotropical [287.8 (mean)±105.0 (SD) Mg ha -1 versus Palaeotropical forests (Africa 418.3±91.8 Mg ha-1; Asia 393.3±109.3 Mg ha-1). Pan-tropical variation in density of large trees and AGB was associated with soil coarseness (negative), soil fertility (positive), community wood density (positive) and dominance of wind dispersed species (positive), temperature in the coldest month (negative), temperature in the warmest month (negative) and rainfall in the wettest month (positive), but results were not always consistent among continents. Main conclusions Density of large trees and AGB were significantly associated with climatic variables, indicating that climate change will affect tropical forest biomass storage. Species trait composition will interact with these future biomass changes as they are also affected by a warmer climate. Given the importance of large trees for variation in AGB across the tropics, and their sensitivity to climate change, we emphasize the need for in-depth analyses of the community dynamics of large trees. (Résumé d'auteur

SignificanceIdentifying and explaining regional differences in tropical forest dynamics, structure, diversity, and composition are critical for anticipating region-specific responses to global environmental change. Floristic classifications are of fundamental importance for these efforts. Here we provide a global tropical forest classification that is explicitly based on community evolutionary similarity, resulting in identification of five major tropical forest regions and their relationships: (i) Indo-Pacific, (ii) Subtropical, (iii) African, (iv) American, and (v) Dry forests. African and American forests are grouped, reflecting their former western Gondwanan connection, while Indo-Pacific forests range from eastern Africa and Madagascar to Australia and the Pacific. The connection between northern-hemisphere Asian and American forests is confirmed, while Dry forests are identified as a single tropical biome.

The marked biogeographic difference between western (Malay Peninsula and Sumatra) and eastern (Borneo) Sundaland is surprising given the long time that these areas have formed a single landmass. A dispersal barrier in the form of a dry savanna corridor during glacial maxima has been proposed to explain this disparity. However, the short duration of these dry savanna conditions make it an unlikely sole cause for the biogeographic pattern. An additional explanation might be related to the coarse sandy soils of central Sundaland. To test these two nonexclusive hypotheses, we performed a floristic cluster analysis based on 111 tree inventories from Peninsular Malaysia, Sumatra, and Borneo. We then identified the indicator genera for clusters that crossed the central Sundaland biogeographic boundary and those that did not cross and tested whether drought and coarse-soil tolerance of the indicator genera differed between them. We found 11 terminal floristic clusters, 10 occurring in Borneo, 5 in Sumatra, and 3 in Peninsular Malaysia. Indicator taxa of clusters that occurred across Sundaland had significantly higher coarse-soil tolerance than did those from clusters that occurred east or west of central Sundaland. For drought tolerance, no such pattern was detected. These results strongly suggest that exposed sandy sea-bed soils acted as a dispersal barrier in central Sundaland. However, we could not confirm the presence of a savanna corridor. This finding makes it clear that proposed biogeographic explanations for plant and animal distributions within Sundaland, including possible migration routes for early humans, need to be reevaluated.climate change | human migration | plant distribution | sea-level change

This is an open access article under the terms of the Creat ive Commo ns Attri bution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

No abstract