The use of molecular chemistry (pyrolysis-GC/MS) in the environmental interpretation of peat

  The molecular composition of organic matter in peatlands reflects local conditions and stores information about botanical composition (plant source) as well as the degree of and conditions during decomposition. A reliable hydrological (and hence palaeoclimatological) interpretation of source and decomposition proxies in peatlands requires the understanding of the interactions between decomposition and botanical composition and the reaction of both to changes in the water table. Only few studies combined vegetation and decomposition characteristics to investigate peat OM dynamics and reconstruct environmental changes, and these included only a limited number of samples. In order to study the three-way relationship between botanical composition, decomposition and environmental conditions, pyrolysis-GC/MS of high-resolution sampled peat cores was used on two contrasting ombrotrophic peatlands. The two sites comprise a Sphagnum-dominated peatland in Tierra del Fuego (Argentina) and a graminoid-dominated peatland in Galicia (Spain). The purpose of this study was to examine the use of analytical pyrolysis in the environmental interpretation of peat profiles and to gain an improved understanding of peat OM dynamics and thereby develop parameters that reliably reflect past environmental conditions. The molecular composition of the Sphagnum-dominated Harberton peatland was first explored for a selection of samples. Several peat fractions have been analysed, including NaOH-extract, NaOH-insoluble residue and bulk sample. A large number of pyrolysis products was identified and quantified for all peat samples; additionally total C, N and ash content were measured. Molecular parameters were identified by a systematic methodological approach. Peatland plants have been analysed in search for specific markers. Factor analysis was used to reduce the number of pyrolysis products for quantification and to assess differences in peat chemistry and relate these to environmental factors. This resulted in a number of molecular parameters that reflect botanical changes, (an)aerobic decomposition and fire incidence. The results of the different peat fractions and plant analysis were applied to a large number of bulk samples (67), and used to reconstruct the 12,000 years of vegetation history of the Harberton core, which was then interpreted in terms of past hydrological conditions. The C:N ratio showed a perfect agreement with the botanical composition according to pyrolytic plant markers. It appeared that the n-alkane distribution showed remarkable changes with respect to shifts in the abundance of Sphagnum. Consequently, the detailed vegetation reconstruction was used to discuss the application of the n-alkane distribution in ombrotrophic peat. The simultaneous effects of botanical shifts (source material) and decomposition may cause conflicting hydrological interpretations for non-specific plant markers such as n-alkanes. The results indicated a considerable effect of aerobic decomposition on the distribution of n-alkanes. Although pyrolysis is not the regular method to establish the n-alkane distribution, comparison of n-alkanes in the NaOH-extract and residue peat samples supported the findings. The results emphasise the importance of combining vegetation and decomposition characteristics. For the graminoid-dominated Penido Vello peatland, all 101 bulk peat samples of the 3 m thick peat deposit have been analysed. In addition to the search for specific compounds of peatland plants, the same pyrolysis products were quantified for plant pyrolysates and peat. Depth records of plant markers agreed well with their preferential habitat, with the degree of decomposition and with the transition from minerotrophic to ombrotrophic peat. Factor analysis was used to reconstruct past hydrology. The results showed the importance of high-resolution sampling through the higher correlations between molecular parameters in the part sampled each 2 cm compared to 5 cm. Furthermore, good correlations of pyrolysis results were found with mineral content, solid-state ¹³C CPMAS NMR data and total N. The hydrological reconstruction based on depth records of plant markers and factor scores obtained by factor analysis of all quantified pyrolysis products, and the low contribution of Sphagnumto the graminoid-dominated Penido Vello peat allowed studying the effects of source and decay on the lignin composition. A large number of lignin pyrolysis products were quantified for peatland plants and for the three different peat fractions for a selection of 15 samples. Lignin composition of woody and graminoid plant species were compared as well as the lignin composition in different peat fractions, and their relation to source or decay was established. To determine whether the lignin parameters, derived from analysis of plants and peat fractions, can be used to reconstruct peat environment, they were applied to all 51 bulk peat samples of the upper meter of the Penido Vello core. The results show a strong effect of vegetation type and anaerobicity on generally-used lignin decomposition proxies, which indicates that such proxies are not reliable without information on the context. To examine the interaction between vegetation type and decomposition processes, the results of Sphagnum-dominated peat and graminoid-dominated peat were combined. In addition to the well-known pyrolytic marker from Sphagnum (4-isopropenylphenol), depth records of pyrolysis products specific for lichens and graminoids functioned well in both peatlands and are considered reliable markers for application in other peatlands. A number of other identified markers only functioned well within one of both peatlands. The different functioning of these markers in the two peatlands emphasises the importance of plant analysis prior to the use of pyrolytic biomarkers with low specificity, but also shows that supposedly non-specific pyrolysis products can be specific within a certain peatland ecosystem. The use of different peat fractions (NaOH-extractable and non-extractable peat) provided information on the degree of decomposition and allowed comparison with studies on peat humic acids and humin, and the generally applied colorimetric method to determine peat humification. Furthermore, the comparison of pyrolysates of extracts and residues for a selection of peat samples, in combination with factor analysis, allowed separation of the effects of source, and several stages of aerobic and anaerobic decay. Degradation of lignin-cellulose (vascular plants) and polyphenol-cellulose (Sphagnum) is discussed using the abundance of their markers (lignin moieties, levoglucosan and 4-isopropenylphenol) in pyrolysates of the different peat fractions, and gives insight into degradation mechanisms. The results indicated that, in Sphagnum litter, polyphenols are more easily degraded than polysaccharides. Comparison of pyrolysis results and C:N ratio showed that differences in litter quality between Sphagnum and vascular plants is a major factor that determines the variance of C:N with depth. Large part of the variation in C:N in Sphagnum-dominated peat is caused by decomposition rather than small increases of vascular plants upon drier conditions, while in graminoid-dominated peat this is not true and several processes may disturb the effect of mass loss on the C:N ratio. The C:N ratio can thus be a consistent decomposition proxy in Sphagnum-dominated peat but not in graminoid-dominated peat. Pyrolysis-GC/MS in combination with the applied research design provided detailed chemical information that gave insight in both vegetation and decomposition characteristics. This methodology thus has a high potential for the reconstruction of past environmental conditions