Decomposition vectors: a new approach to estimating woody detritus decomposition dynamics

A chronosequence of three species of logs (Pinus sylvestris L., Picea abies (L.) Karst, and Betula pendula Roth.) from northwestern Russia was resampl ed to develop a new method to estimate rates of biomass, volume, and densit y loss. We call this resampling of a chronosequence the decomposition-vecto r method, and it represents a hybrid between the chronosequence and time-se ries approaches. The decomposition-vector method with a 3-year resampling i nterval gave decomposition rates statistically similar to those of the one- time chronosequence method. This indicated that, for most cases, a negative exponential pattern of biomass, volume, and density loss occurred. In the case of biomass loss of P. sylvestris, however, polynomial regression indic ated decomposition rates were initially low, then increased, and then decre ased as biomass was lost. This strongly suggests three distinct phases: the first when decomposers colonized the woody detritus, a second period of ra pid exponential mass loss, and a third period of slow decomposition. The co nsequences for this complex pattern of decomposition were explored at the e cosystem level using a simple model. We found that a single rate constant c an be used if inputs vary within a factor of 10, but that this approach is problematical if inputs are more variable.