| Summary: | Tropical forests play a vital role in the global carbon cycle and international policy, such as the United Nations Collaborative Programme on Reducing Emissions from Deforestation or Degradation (REDD+), but the amount of carbon they contain and its spatial distribution remain uncertain. Allometric equations used to estimate tree mass are a key source of this uncertainty, because large-scale variation in tree allometry and fundamental differences between functional groups are not accurately represented in pantropical biomass equations. This research tests the effects of accounting for sources of variation not currently explained in treemodels (i.e., crown size and structure) and recognising important distinctions between functional groups (monocots vs. dicots) at both the level of individuals and across the landscape. Southwestern Amazonian forests are politically and ecologically important, but biomass estimates may be articularly uncertain in this region. Specifically, tree biomass estimates vary greatly among published models, but these models do not account for crown structure nor have their predictions been tested against directly measured data in the southwestern Amazon. Palms are also abundant in western Amazonia but theirmass has been widely misrepresented: using models developed for dicotyledonous trees is likely inaccurate because these two groups have very different structures. To test these ideas, 51 trees and 136 arborescent palms were harvested and weighed in Peru, including the heaviest tropical tree on record. Existing pantropical equations that included height underestimated tree biomass by 11-14%because large crowns partially compensate for lower stature. Including crown parameters in new allometric models greatly improved performance and reduced error, especially for the largest trees. Palm biomass was often underestimated by dicot models because they can be much taller at small diameters, and stem height was the most important variable in new equations. These results were confirmed on a larger scale. Based on a network of 53 forest plots, biomass carbon in trees and palms in the southwestern Amazon is 9%greater than estimated by the recommended pantropical biomass equation. Original total aboveground carbon stocks over the entire 746,653 km2 ecoregion is estimated at 11.5 Pg C. Nearly one third of forests in this region are at imminent risk of deforestation and forest degradation, which would result in emissions up to 4.4 Pg C. These results significantly advance allometricmodelling and reduce uncertainty in forest biomass estimates, especially in southwestern Amazonia, which should help to underpin effective forest management and provide better forest biomass estimates for REDD+ and other carbon-based conservation projects.
|
|---|
