Nutrient limitation of woody debris decomposition in a tropical forest : contrasting effects of N and P addition

Abstract

Tropical forests represent a major terrestrial store of carbon (C), a large proportion of which is contained in the soil and decaying organic matter. Woody debris plays a key role in forest C dynamics because it contains a sizeable proportion of total forest C. Understanding the factors controlling the decomposition of organic matter in general, and woody debris in particular, is hence critical to assessing changes in tropical C storage. We conducted a factorial fertilization experiment in a tropical forest in South China to investigate the influence of nitrogen (N) and phosphorus (P) availability on woody debris decomposition using branch segments (5 cm diameter) of four species (Acacia auriculaeformis, Aphanamixis polystachya, Schefflera octophylla, and Carallia brachiata) in plots fertilized with +N, +P, or +NP, and controls. Fertilization with +P and +NP increased decomposition rates by 5–53%, and the magnitude was species specific. Contrary to expectations, we observed no negative effect of +N addition on decay rates or mass loss of woody debris in any of the four study species. Decomposition rates of woody debris were higher in species with lower C : P ratios regardless of treatment. We observed significant accumulation of P in the woody debris of all species in plots fertilized with +P and +NP during the early stages of decomposition. N release from woody debris of Acacia (N-fixing) was greater in the +P plots towards the end of the study, whereas fertilization with +N had no impact on the patterns of nutrient release during decomposition. Synthesis: Our results indicate that decomposition of woody debris is primarily constrained by P availability in this tropical forest. However, contrary to expectations, +N addition did not exacerbate P limitation. It is conceivable that decay rates of woody debris in tropical forests can be predicted by C : P or lignin : P ratios, but additional work with more tree species is needed to determine whether the patterns we observed are more generally applicable.