COS 87-1 - Power laws are better than exponential decay models for representing litter and soil organic matter decomposition

Thursday, August 11, 2011: 8:00 AM
Ballroom F, Austin Convention Center
Gervasio Piñeiro, IFEVA, Facultad de Agronomía, Universidad de Buenos Aires, CONICET, Buenos Aires, Argentina, Stefano Manzoni, Departments of Crop Production Ecology and Ecology, Swedish University of Agricultural Sciences, Sweden, John H. Kim, Program of Ecology, Duke University, Durham, NC, Esteban G. Jobbagy, Grupo de Estudios Ambientales, Universidad Nacional de San Luis, San Luis, Argentina, Margaret S. Torn, Energy and Resources Group, University of California, Berkeley, CA, William J. Riley, Earth and Environmental Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, Amilcare Porporato, Department of Civil and Environmental Engineering, Duke University, Durham, NC and Robert B., Jackson, Nicholas School of the Environment and Dept of Biology, Duke University, Durham, NC

Litter and soil organic matter (SOM) decomposition models are often based on first-order exponential decay functions of the type (Ct= Co.e-kt). The decomposition or decay factor k is assumed constant over time, implying that the same proportion of the litter or SOM pool is decomposed per unit time throughout the decomposition processes. These models can be parameterized for one or multiple pools, and are usually adjusted to observations from litterbag experiments, SOM incubations and SOM labeling experiments. We evaluated the assumptions behind exponential decay models (EDM) and compared them with four different power law models (PLM). We used the LIDET experiment (>200 litter decomposition series over 10 years) and compiled published datasets and our own experiments (>100 SOM incubations and labeling series). 


Our results show that for 71% of the litter studies and almost all SOM studies adjusted k values decreased with increasing time span and number of data points. Furthermore, k also decreased with time for different litter and SOM fractions, suggesting that even for fractions assumed to be relatively homogeneous decomposition rates declined over time. Our analysis showed that the turnover time (1-k) of SOM estimated in contrasting climatic regions and soils was strongly related with the length of the experiment (r2=0.57, p<0.01, n=58), suggesting that turnover has been misestimated and EDM erroneously applied. In contrast, PLM with two parameters had unbiased estimates of decomposition and were more accurate than two pool EDM with three parameters. Our results further change our conceptual view of the decomposition process by (i) showing that the decomposition rate of any separable organic matter fraction decreases over time (a smaller proportion is decomposed in each time step), challenging the characterization of organic matter pools by its turnover, and (ii) PLMs assume that organic matter is a heterogeneous pool and thus allows different mean residence time (life expectancy) and mean age (averaged ages), which were assumed to be identical in EDM. This conceptual model has strong theoretical basis and reconciles previous studies that found large differences in mean residence time and mean age of SOM pools, suggesting that litter and SOM are, in an analogy to human population, characterized by a high infant morality (recently added C is rapidly respired). This work changes our current understanding of the decomposition processes and opens new possibilities for precise modeling of carbon dynamics especially in long-term time scales.

Copyright © . All rights reserved.
Banner photo by Flickr user greg westfall.