Journal Article

PreJuSER-8985

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Information content of incubation experiments for inverse estimation of pools in the Rothamsted carbon model: a Bayesian perspective

Scharnagl, B.FZJ* ; Vrugt, J. A. ; Vereecken, H.FZJ* ; Herbst, M.FZJ*

2010
Copernicus Katlenburg-Lindau [u.a.]

This record in other databases:  

Please use a persistent id in citations: http://hdl.handle.net/2128/20956  doi:10.5194/bg-7-763-2010

Abstract: A major drawback of current soil organic carbon (SOC) models is that their conceptually defined pools do not necessarily correspond to measurable SOC fractions in real practice. This not only impairs our ability to rigorously evaluate SOC models but also makes it difficult to derive accurate initial states of the individual carbon pools. In this study, we tested the feasibility of inverse modelling for estimating pools in the Rothamsted carbon model (ROTHC) using mineralization rates observed during incubation experiments. This inverse approach may provide an alternative to existing SOC fractionation methods. To illustrate our approach, we used a time series of synthetically generated mineralization rates using the ROTHC model. We adopted a Bayesian approach using the recently developed DiffeRential Evolution Adaptive Metropolis (DREAM) algorithm to infer probability density functions of the various carbon pools at the start of incubation. The Kullback-Leibler divergence was used to quantify the information content of the mineralization rate data. Our results indicate that measured mineralization rates generally provided sufficient information to reliably estimate all carbon pools in the ROTHC model. The incubation time necessary to appropriately constrain all pools was about 900 days. The use of prior information on microbial biomass carbon significantly reduced the uncertainty of the initial carbon pools, decreasing the required incubation time to about 600 days. Simultaneous estimation of initial carbon pools and decomposition rate constants significantly increased the uncertainty of the carbon pools. This effect was most pronounced for the intermediate and slow pools. Altogether, our results demonstrate that it is particularly difficult to derive reasonable estimates of the humified organic matter pool and the inert organic matter pool from inverse modelling of mineralization rates observed during incubation experiments.

Keyword(s): J

Note: We thank Sirgit Kummer and Wolfgang Tappe for providing the experimental data used to estimate the precision of mineralization rate measurements. The first, third and fourth author gratefully acknowledge financial support by the TERENO project and by the SFB/TR 32 "Pattern in Soil-Vegetation-Atmosphere Systems: Monitoring, Modeling, and Data Assimilation" funded by the Deutsche Forschungsgemeinschaft (DFG). The work of the second author was sponsored by a J. Robert Oppenheimer Fellowship from the LANL Postdoctoral Program. We thank the two anonymous referees for their insightful comments on the discussion paper.

---

Research Program(s):

1. Terrestrische Umwelt (P24)

Appears in the scientific report 2010

---

Database coverage:
;

---