The enzymic basis of plant litter decomposition: emergence of an ecological process

Abstract

Plant litter decomposition is an integral process in the macronutrient cycles of ecosystems. The absence of fine-scale models for this process hampers attempts to simulate ecosystem responses to disturbance. Two recent studies have suggested that mass loss from decomposing plant litter can be directly related to the temporally-integrated activity of extracellular lignocellulose-degrading enzymes. To evaluate the generality and potential application of such relationships, we surveyed the literature and found eight studies that included mass loss data and some enzymic measure of lignocellulose degradation potential. Although the number of suitable studies was small, they encompassed a broad range of ecosystem and litter types. For all studies, there were strong linear relationships between temporally integrated enzyme activity, expressed as cumulative activity-days, and mass loss. No single enzyme gave the best fit in all cases; multiple linear regressions that included all enzymes measured within a particular study generally yielded better goodness of fit statistics than single enzyme models. Where methodological compatibility permitted, direct comparisons of apparent enzymatic efficiency (relative mass loss/activity-day) were made between studies; values for particular enzymes varied by a factor of ten and were strongly correlated with mean exposure temperature ( r 2 = 0.88). Activation energy for enzymatic decomposition was estimated at 58 kJ mol −1. Principal components analysis (PCA) was used to generate a composite lignocellulase variable from each study, providing a common format for comparison. The results suggested that the three microcosm studies differed from the field investigations: enzymatic efficiency was approximately half that estimated for field studies and the data were more stochastic. We attributed these differences to disruptions in microbial succession caused by the lack of exogenous sources of colonists, nutrients and grazers. PCA also permitted the calculation of a global regression model for mass loss as a function of cumulative lignocellulase activity. For the five field studies, this model had an r 2 value of 0.73 in linear form and 0.76 in logarithmic form. Our analyses suggest that enzymatic decomposition models retain predictive value even when viewed from the ecosystem perspective. This hierarchal penetrance suggests useful applications: as a monitoring tool for the estimation of litter turnover in the field and as a basis for the simulation of decomposition processes at the microbial community level.