Linking spatial patterns of leaf litterfall and soil nutrients in a tropical forest: a neighborhood approach.

Leaf litter represents an important link between tree community composition, forest productivity and biomass, and ecosystem processes. In forests, the spatial distribution of trees and species-specific differences in leaf litter production and quality are likely to cause spatial heterogeneity in nutrient returns to the forest floor and, therefore, in the redistribution of soil nutrients. Using mapped trees and leaf litter data for 12 tree species in a subtropical forest with a well-documented history of land use, we: (1) parameterized spatially explicit models of leaf litter biomass and nutrient deposition; (2) assessed variation in leaf litter inputs across forest areas with different land use legacies; and (3) determined the degree to which the quantity and quality of leaf litter inputs and soil physical characteristics are associated with spatial heterogeneity in soil nutrient ratios (C:N and N:P). The models captured the effects of tree size and location on spatial variation in leaf litterfall (R² = 0.31-0.79). For all 12 focal species, most of the leaf litter fell less than 5 m away from the source trees, generating fine- scale spatial heterogeneity in leaf litter inputs. Secondary forest species, which dominate areas in earlier successional stages, had lower leaf litter C:N ratios and produced less litter biomass than old-growth specialists. In contrast, P content and N:P ratios did not vary consistently among successional groups. Interspecific variation in leaf litter quality translated into differences in the quantity and quality (C:N) of total leaf litter biomass inputs and among areas with different land use histories. Spatial variation in leaf litter C:N inputs was the major factor associated with heterogeneity in soil C:N ratios relative to soil physical characteristics. In contrast, spatial variation soil N:P was more strongly associated with spatial variation in topography than heterogeneity in leaf litter inputs. The modeling approach presented here can be used to generate prediction surfaces for leaf litter deposition and quality onto the forest floor, a useful tool for understanding soil-vegetation feedbacks. A better understanding of the role of leaf litter inputs from secondary vegetation in restoring soil nutrient stocks will also assist in managing expanding secondary forests in tropical regions.