Monitoring data over the period 1994-2007 were analysed for three streams (Cottage Hill Sike, CHS; Rough Sike, RS; Trout Beck, TB) draining blanket peat underlain by glacial clay and limestone-rich sub-strata at Moor House (Northern England). Dissolved organic carbon concentration, [DOC], showed complex relationships with both discharge and calcium concentration, [Ca]. A model based on [Ca] was constructed to simulate stream [DOC] by mixing dissolved organic matter (DOM) from shallow peat, quantified by measured [DOC](15-30 mg l -1 ) in peat porewater, with DOM assumed to be present at a constant concentration (c. 5 mg l -1 ) in groundwater. A temperature-based adjustment to the measured porewater [DOC] was required to account for relatively low streamwater [DOC] during winter and spring. The fitted model reproduced short-term variation in streamwater [DOC] satisfactorily, in particular variability in RS and TB due to groundwater contributions. Streamwater DOM is largely derived from surface peat, which accounts for more than 96% of the total DOC flux in both RS and TB, and 100% in CHS. Model outputs were combined with streamwater and porewater DO 14 C data to estimate the 14 C contents, and thereby the ages, of DOM from peat and groundwater. The peat-derived DOM is 5 years old on average, with most of it very recently formed. The derived age of groundwater DOM (8,500 years) is comparable to the 4,000-7,000 years estimated from the DO 14 C of water extracts of clay underlying the peat, suggesting that the clay is the source of groundwater DOM.

Lake Vanda, a perennially ice-covered Antarctic lake has a highly stratified structure with a pronounced density gradient from 45 m and is anoxic from 68 m down. In order to gain a greater understanding of the carbon cycling in the lake we attempted to measure sulphate reduction in the summer of 1994, but rates were below detectable limits of 1.2 × 106 mol sulphate yr−1. Therefore sulphate reduction was estimated by calculating the rate of sulphide diffusion from the anoxic zone. Sulphate reduction rates were estimated to be 17.7 × 103 mol sulphate yr−1, accounting for 30% of planktonic primary production over the whole lake, but over 100% of primary productivity in the deep chlorophyll maximum. Radiocarbon dating of organic matter in the sediment (565 ± 50 yr bp), suggests that little new organic carbon reaches the sediment from the upper water column. Oxygen diffusion into the anoxic zone could account for only 44% of sulphide reoxidation; therefore, alternative oxidizing potential presumably accounted for the remainder. Sulphate concentration may be controlled by the solubility limit for a mineral phase, perhaps gypsum, as it shows a positive correlation with conductivity. The sulphur species in the anoxic zone are highly enriched in 34S (sulphate δ34S = + 42 to + 46%0, sulphide δ34S = + 13.9%0) which may be explained by either selective removal of sulphide at the oxic/anoxic interface by precipitation of metal sulphides or previous loss ofH2S to the atmosphere. While sulphate reduction is an important part of the carbon cycle within the lake, the oligotrophic nature of Lake Vanda means that all microbiological process rates are very low. The data presented suggest a dynamic and complex history for Lake Vanda and are entirely consistent with present theories of the lake's history.

One of the difficulties in reporting accurate radiocarbon results from compoundspecific radiocarbon analysis (CSRA) is the lack of suitable process standard materials to correct for the amount and 14 C content of carbon added during extensive sample processing. We evaluated the use of n-alkanes extracted from modern grass material (1.224 ±0.006 fraction modern) as process standards for CSRA. The n-alkanes were isolated using preparative capillary gas chromatography (PCGC) from two independent chemical extraction methods applied to the grass. Since this was our first assessment of the 14 C content of the grass n-alkanes, we corrected for extraneous carbon derived from PCGC isolation using commercially available single compounds of modern and 14 C-free content. Results were consistent across the two extraction methods showing that the C 29 n-alkane has a fraction modern value that is within 1σ of the bulk value of the grass while C 31 n-alkane and less abundant nalkanes have values within 2σ of the bulk value of the grass. C 29 and C 31 n-alkanes were the most abundant n-alkanes in the grass and, as such, the more feasible for collection of sufficient amounts of carbon for accelerator mass spectrometry (AMS) analysis. Our results suggest that choosing a grass n-alkane with an elution time closest to that of the unknowns may be advisable due to possibly greater effect from GC column bleed ( 14 C-free) at later elution times. We conclude that C 29 and C 31 nalkanes in modern grass of known 14 C content can be used as in-house standards to correct for the addition of 14 C-free carbon during sample preparation for 14 C analysis of n-alkanes. 2