The influence of model type and incubation time on the estimation of stable organic carbon in organic materials

There is a renewed interest in the stability and turnover of organic carbon (C) in organic materials added to soil, partly driven by concerns about global changes. We used mineralization data from laboratory incubations to determine the amount of stable organic C present. Four organic wastes were incubated with soil under controlled conditions and C mineralization was monitored, and four additional C mineralization experiments were selected from the literature. All data were rescaled to a common temperature, and we fitted five different models with biological significance to the C mineralization data, namely a first-order model (M1), a parallel first-order model (M11), a parallel first-zero-order model (M10), a second-order model (M2) and a Monod kinetic model (Monod). All models could describe the mineralization data satisfactorily. However, this does not automatically imply that the individual parameter values are realistic estimates of the underlying biological processes. All models were then fitted for progressively shorter incubation times and were used to extrapolate C mineralization to predict the amount of (measured) stable organic C. The M11 and Monod models performed best in estimating stable organic C, but they did not fit the short-term incubation data (< 100 days) for some materials. The M2 model allowed stable organic C to be estimated within less than 3% of the true value for all materials based on an incubation time of less than 60 days. The M1 and M10 models systematically and strongly under- or overestimated stable organic C as the incubation time was shortened, and should not be used for extrapolation from short-term data.