The use of plant-specific pyrolysis products as biomarkers in peat deposits

Schellekens, J.; Bradley, J.A.; Kuyper, T.W.; Fraga, I.; Pontevedra Pombal, X.; Buurman, P.


Peatlands are archives of environmental change that can be driven by climate and human activity. Proxies for peatland vegetation composition provide records of (local) environmental conditions that can be linked to both autogenic and allogenic factors. Analytical pyrolysis offers a molecular fingerprint of peat, and thereby a suite of environmental proxies. Here we investigate analytical pyrolysis as a method for biomarker analysis. Pyrolysates of 48 peatland plant species were compared, comprising seventeen lichens, three Sphagnum species, four non-Sphagnum mosses, eleven graminoids (Cyperaceae, Juncaceae, Poaceae), five Ericaceae and six species from other families. This resulted in twenty-one potential biomarkers, including new markers for lichens (3-methoxy-5-methylphenol) and graminoids (ferulic acid methyl ester). The potential of the identified biomarkers to reconstruct vegetation composition is discussed according to their depth records in cores from six peatlands from boreal, temperate and tropical biomes. The occurrence of markers for Sphagnum, graminoids and lichens in all six studied peat deposits indicates that they persist in peat of thousands of years old, in different vegetation types and under different conditions. In order to facilitate the quantification of biomarkers from pyrolysates, typically expressed as proportion (%) of the total quantified pyrolysis products, an internal standard (5-a-androstane) was introduced. Depth records of the Sphagnum marker 4-isopropenylphenol from the upper 3 m of a Sphagnum-dominated peat, from samples analysed with and without internal standard showed a strong positive correlation (r2 = 0.72, P <0.0005, n = 12). This indicates that application of an internal standard is a reliable method to assess biomarker depth records, which enormously facilitates the use of analytical pyrolysis in biomarker research by avoiding quantification of a high number of products.