Free-shape 14C age-depth modelling of an intensively dated modern peat profile

The paper presents a new algorithm for finding the best-fit age-depth model of a peat profile dated with C-14. This algorithm of free-shape modelling is useful when the age-depth curve is wiggly and difficult to approximate by a simple mathematical function. The performance of the free-shape algorithm is illustrated with the example of the MA2 peat profile from Mauntschas mire, using 29 radiocarbon dates spanning the last 1300a, and where the age-depth Curve revealed a relatively complicated shape. The best-fit age-depth model searched for is a Curve maximising the 'Fit quality', defined as a combination of fit of C-14 dates to the radiocarbon calibration Curve, general smoothness of the age-depth line, and similarity of relative changes in the modelled sediment accumulation rates to those Suggested by independent data. The algorithm of free-shape age-depth modelling searched for a reasonable equilibrium or compromise between the conditions above. Uncertainty of the free-shape model was assessed with a Monte Carlo Markov chain built from large number of age-depth lines accepted with probabilities dependent on the 'fit quality' defined as above. Parallel fluctuations in concentrations of most pollen taxa in MA2 strongly Suggest a complicated shape of the age-depth curve, which is Supported in the upper part of the profile by the record of spheroidal carbonaceous particles, and in some levels also by the C-14 dates. A modification of total pollen concentrations was therefore used as a quantitative indicator of relative changes of the peat accumulation rate. The use of pollen concentrations improved the fit of the C-14 dates in several cases. Application of the final age-depth model to the pollen data resulted in a diagram of pollen accumulation which allows interpretation in ecological terms, whereas age-depth models that did not use pollen concentrations led to problematic and Confusing results. The obtained age-depth model is therefore accepted as a good approximation of the actual conditions. Copyright (C) 2009 John Wiley & Sons, Ltd.