Initial white rot type dominance of wood decomposition and its functional consequences in a regenerating tropical dry forest

Efforts to model woody debris dynamics are limited by our empirical understanding of the patterns and drivers of decomposition. This knowledge gap is significant in tropical forests, particularly in the dry tropics where research has been minimal and where forest regeneration is a management priority. Here, we coupled trait-function relationships in decomposing logs with indices of microbial and insect activity in a regenerating Costa Rican dry forest. We cut and placed logs (～18 cm dia) of eight tree species in ground contact at two sites. We assessed density loss and element dynamics in sapwood and heartwood twice annually over two years. At time 0 and year 2, we measured lignin, nitrogen, structural carbohydrates, extractives, insect galleries, and two residue ‘signatures’ of fungal rot type: dilute alkali solubility (DAS; higher for brown rot) and lignin:glucan loss (higher for white rot). After two years, sapwood mean density losses ranged from 11.6 to 44.4% among tree species (excluding one thoroughly-degraded species). The best predictor of sapwood density loss was initial pH, but this correlation was negative, contrasting positive correlations proposed for temperate forests. Mean heartwood density losses were consistently less than those in sapwood, and although heartwood extractives contents were as high as 16.4%, trait correlations were insignificant. Insect galleries contributed little to density loss (<3%), and DAS and lignin loss patterns indicated dominance by white rot fungi. This was often matched by dense fungal zone line patterns (spalting), outlining many small territories. Perhaps as a consequence, element patterns were spatially variable, with overall trends roughly similar to those from temperate studies (e.g., Ca gain, P, K loss). Estimated CO₂ fluxes from logs ranged from ～25 to 75% percent of annual fluxes from litter fall. This collectively implies an important role for wood decomposition in dry forest carbon cycling, and in our case, it shows an interesting pattern suggesting high decomposer spatial complexity but low functional diversity.