Multiple drivers of aboveground biomass in a human-modified landscape of the Caatinga dry forest

Aboveground biomass is an important predictor of net primary productivity and provision of ecosystem services such as carbon storage and food supply. However, biomass can be a complex and multi-driven ecosystem feature, particularly in tropical forests experiencing human disturbances. Here we examine the potential effects of forest successional stage, species richness, plant assemblage functional composition, rainfall, grazing and soil fertility as drivers of biomass in a human-modified landscape of the Caatinga dry forest. We recorded 8911 stems (DSH >= 3) across 35 0.-ha plots and calculated the aboveground biomass adopting allometric model developed specifically for Caatinga dry forest. Caatinga flora was dominated by species bearing moderate to heavy wood, with aboveground biomass concentrated into a few dominating species and narrow stems. Biomass averaged 28.48 +/- 23.32 Mg ha(-1) with a high cross-stand variation as it resulted from a complex interaction of age of forest stand, rainfall and species richness; i.e. old-growth forest stands supported twice as much biomass (38.81 +/- 25.08 Mg ha(-1)) than successional stands (14.68 +/- 10.52 Mg ha(-1)). Drivers of plot-level biomass also affected biomass at species level as well as the abundance of the species which contributed most for forest biomass. Our results confirm that slash-and-burn agriculture supports the emergence of biomass mosaics associated to the presence of old-growth and successional forest stands of varying age. In this context, Caattinga supports reduced forest biomass, which responds to both natural and human-driven forces and their complex interactions. Moreover, biomass persistence in human-modified landscapes requires not only long fallow periods, but also the protection of species-rich forest assemblages.