Isotopic composition of modern ice wedges and winter precipitation in the Laptev Sea region (Table 2)

The first step for the application of stable isotope analyses of ice wedges for the correct paleoclimatic reconstruction supposes the study of the isotopic composition of modern ice wedges and their relationship with the isotopic composition of modern precipitation. The purpose of this research is to present, to analyze and to discuss new data on isotopic composition (d18O, dD, 3H) of modern ice wedges obtained in the Laptev Sea region in 1998-99. Investigations were carried out at two sites: on Bykovsky Peninsula in 1998 and on Bol'shoy Lyakhovsky Island in 1999 and were based on the combined application of both tritium CH) and stable isotope (d18O, dD) analyses. Tritium analyses of the atmospheric precipitation collected during two field seasons show seasonal variations: high tritium concentration in snow (to a maximum of 207 TU) and low values of tritium concentration (<20 TU) in rain. High tritium concentrations are also observed in the surface water, in suprapermafrost ground waters, and in the upper part of permafrost. High tritium concentrations range between 30-40 TU and 750 TU in the studied modern ice wedges (active ice wedges), which let us believe that they are of modern growth. Such high tritium concentrations in ice wedges can not be associated with old thermonuclear tritium because of the radioactive decay. High tritium concentrations found in the snow cover in 1998/99, in the active layer and in the upper part of permafrost give evidence of modern (probably the last decade) technogenic tritium arrival from the atmosphere on to the Earth surface in the region. The comparison of the isotopic composition (d18O, dD and d-excess) of active ice wedges and modern winter precipitation in both sites shows: 1) the isotopic composition of snow correlates linearly with a slope close to 8.0 and parallel to the GMWL at both sites; 2) the mean isotopic composition of active ice wedges on Bykovsky Peninsula is in good agreement with the mean isotopic composition of modern snow; 3) the isotopic composition of active ice wedges and snow on Bol'shoy Lyakhovsky Island are considerably different. There are low values of d-excess in all studied active ice wedges (mean value is about 4.8 per mil), while in snow, the mean value of d-excess is about 9.5 per mil. Possible reasons for this gap are the following: 1) the modification of the isotopic composition in snow during the spring period; 2) changes in the isotopic composition of ice wedges due to the process of ice sublimation in open frost cracks during the cold period; 3) mixing of snowmelt water with different types of surface water during the spring period; 4) different moisture source regions.

Supplementary material: XRF core scanning data, CaCO3 values, foraminiferal carbon and oxygen isotope data, Mg/Ca data, clay mineralogy and helium isotope results

The Paleocene-Eocene Thermal Maximum (PETM; ~56 Ma) is associated with abrupt climate change, carbon cycle perturbation, ocean acidification, as well as biogeographic shifts in marine and terrestrial biota that were largely reversed as the climatic transient waned. We report a clear exception to the behavior of the PETM as a reversing climatic transient in the eastern North Atlantic (Deep-Sea Drilling Project Site 401, Bay of Biscay) where the PETM initiates a greatly prolonged environmental change compared to other places on Earth where records exist. The observed environmental perturbation extended well past the d13C recovery phase and up to 650 kyr after the PETM onset according to our extraterrestrial 3He-based age-model. We observe a strong decoupling of planktic foraminiferal d18O and Mg/Ca values during the PETM d13C recovery phase, which in combination with results from helium isotopes and clay mineralogy, suggests that the PETM triggered a hydrologic change in western Europe that increased freshwater flux and the delivery of weathering products to the eastern North Atlantic. This state change persisted long after the carbon-cycle perturbation had stopped. We hypothesize that either long-lived continental drainage patterns were altered by enhanced hydrological cycling induced by the PETM, or alternatively that the climate system in the hinterland area of Site 401 was forced into a new climate state that was not easily reversed in the aftermath of the PETM.

Depth tie points and ages of IODP Site 306-U1314

Integrated Ocean Drilling Program (IODP) Site U1314 of the North Atlantic is a critical sedimentary archive record of subpolar deep water from the southern Gardar Drift for which we derived an age model of orbital resolution for the last 1.8 Ma. This chronology combined with high-resolution (cm scale) X-ray fluorescence core scanning measurements of major elements allows tracking changes in terrigenous provenance during the last 1.1 Ma. Low Potassium to Titanium (K/Ti) ratios reflect enhanced transport of basalt-derived titanomagnetites during warm climate intervals, while high K/Ti ratios indicate a dominance of acidic sediment sources typical for glacial and stadial events. Changes in K/Ti and magnetic concentration at Site 1314 are coeval with fluctuations in smectite content and grain size data from nearby piston cores, suggesting that the provenance changes are mainly controlled by variable flow of the Iceland-Scotland Overflow Water, an important branch of North Atlantic Deep Water. Furthermore, K/Ti variations on orbital time scales show a striking similarity to the deep sea d13C record from ODP Site 607. Pervasive features of the K/Ti time series during and after the Mid-Pleistocene Transition are suborbital changes similar to Dansgaard/Oeschger and Bond oscillations that appear to be strongly amplified during ice growth phases when global benthic d18O was within the range of ~4.1-4.6 per mil. The strong increase in variability of sediment provenance and subsequently deep hydrography at benthic d18O values below ~4.1 suggests that the extent of glaciations and, therefore, sea level corresponding to this value constitutes an important physical threshold that was persistent at least for the last 1.1 Ma.

Carbon cycle change in the Eocene

The transition from the extreme global warmth of the early Eocene 'greenhouse' climate ~55 million years ago to the present glaciated state is one of the most prominent changes in Earth's climatic evolution. It is widely accepted that large ice sheets first appeared on Antarctica ~34 million years ago, coincident with decreasing atmospheric carbon dioxide concentrations and a deepening of the calcite compensation depth in the world's oceans, and that glaciation in the Northern Hemisphere began much later, between 10 and 6 million years ago. Here we present records of sediment and foraminiferal geochemistry covering the greenhouse-icehouse climate transition. We report evidence for synchronous deepening and subsequent oscillations in the calcite compensation depth in the tropical Pacific and South Atlantic oceans from ~42 million years ago, with a permanent deepening 34 million years ago. The most prominent variations in the calcite compensation depth coincide with changes in seawater oxygen isotope ratios of up to 1.5 per mil, suggesting a lowering of global sea level through significant storage of ice in both hemispheres by at least 100 to 125 metres. Variations in benthic carbon isotope ratios of up to ~1.4 per mil occurred at the same time, indicating large changes in carbon cycling. We suggest that the greenhouse-icehouse transition was closely coupled to the evolution of atmospheric carbon dioxide, and that negative carbon cycle feedbacks may have prevented the permanent establishment of large ice sheets earlier than 34 million years ago.

North Atlantic late Miocene stable-isotope stratigraphy, biostratigraphy, and magnetostratigraphy

Upper Miocene foraminiferal nannofossil ooze and chalk from DSDP Hole 552A in the northeast Atlantic Ocean have been closely sampled for biostratigraphic, paleomagnetic, and stable-isotopic studies. Sampling at 10-cm intervals resulted in an uppermost Miocene isotope stratigraphy with a 1000- to 3000-yr. resolution. Covariance in benthic (Planulina wuellerstorfi) and planktonic (Globigerina bulloides) foraminiferal d18O records is taken as evidence for variability in continental ice volume. Our best estimate is that glacial maxima occurred at -5.0 and ~ 5.5 Ma and lasted no more than 20,000 yrs. These events probably lowered sea level by 60 m below the latest Miocene average. There is little oxygen-isotope evidence, however, for a prolonged glaciation during the last 2 m.y. of the late Miocene. High- and low-frequency variability in the d13C record of foraminifers is useful for correlation among North Atlantic DSDP Sites 408, 410, 522, 610, and 611, and for correlation with sites in other oceans. Similar d13C changes are seen in P. wuellerstorfi and G. bulloides, but the amplitude of the signal is always greater in G. bulloides. Variability in d13C common to both species probably reflects variability in the d13C of total CO2 in seawater. Major long-term features in the d13C record include a latest Miocene maximum (P. wuellerstorfi = 1.5 per mil ) in paleomagnetic Chron 7, an abrupt decrease in d13C at -6.2 Ma, and a slight increase at -5.5 Ma. The decrease in d13C at -6.2 Ma, which has been paleomagnetically dated only twice before, occurs in the upper reversed part of Chronozone 6 at Holes 552A and 611C, in excellent agreement with earlier studies. Cycles in d13C with a period of ~ 10 4 yrs. are interpreted as changes in seawater chemistry, which may have resulted from orbitally induced variability in continental biomass. Samples of P. wuellerstorfi younger than 6 Ma from throughout the North Atlantic have d13C near lo, on average ~ l per mil greater than samples of the same age in the Pacific Ocean. Thus, there is no evidence for cessation of North Atlantic Deep Water production resulting from the Messinian "salinity crisis." Biostratigraphic results indicate continuous sedimentation during the late Miocene after about -6.5 Ma at Hole 552A. Nannofossil biostratigraphy is complicated by the scarcity of low-latitude marker species, but middle and late Miocene Zones NN7 through NN11 are recognized. A hiatus is present at -6.5 Ma, on the basis of simultaneous first occurrences of Amaurolithusprimus, Amaurolithus delicatus, Amaurolithus amplificus, and Scyphosphaera globulata. The frequency and duration of older hiatuses increase downsection in Hole 552A, as suggested by calcareous nannofossil biostratigraphy and magnetostratigraphy. Paleomagnetic results at Hole 552A indicate a systematic pattern of inclination changes. Chronozone 6 was readily identified because of its characteristic nannoflora (sequential occurrences of species assigned to the genus Amaurolithus) and the d13C decrease in foraminifers, but its lower reversed interval is condensed. Only the lower normal interval of Chronozone 5 was recognized at Hole 552A; the upper normal interval and the lowest Gilbert sediment are not recognized, owing to low intensity of magnetization and to coring disturbance. Interpreting magnetic reversals below Chronozone 6 was difficult because of hiatuses, but a lower normally magnetized interval is probably Chronozone 7. Correlation between DSDP Hole 552A and other North Atlantic sites is demonstrated using coiling direction changes in the planktonic foraminifer Neogloboquadrina. At most sites this genus changed its coiling preference from dominantly right to dominantly left during the late Miocene. At Hole 552A this event probably occurred about 7 m.y. ago. At the same time, P. wuellerstorfi had maximum d13C values. A similar d13C maximum and coiling change occurred together in Chron 7 at Hole 611C, and at Hole 610E. In sediment younger than -5.5 Ma, the coiling of small Neogloboquadrina species is random, but the larger species N. atlantica retains preferential left coiling.

Borehole temperature, submarine permafrost methane concentrations and pore water chemistry measured on borehole BK2, central Laptev Sea shelf

Submarine permafrost degradation has been invoked as a cause for recent observations of methane emissions from the seabed to the water column and atmosphere of the East Siberian shelf. Sediment drilled 52 m down from the sea ice in Buor Khaya Bay, central Laptev Sea revealed unfrozen sediment overlying ice-bonded permafrost. Methane concentrations in the overlying unfrozen sediment were low (mean 20 µM) but higher in the underlying ice-bonded submarine permafrost (mean 380 µM). In contrast, sulfate concentrations were substantially higher in the unfrozen sediment (mean 2.5 mM) than in the underlying submarine permafrost (mean 0.1 mM). Using deduced permafrost degradation rates, we calculate potential mean methane efflux from degrading permafrost of 120 mg/m**2 per year at this site. However, a drop of methane concentrations from 190 µM to 19 µM and a concomitant increase of methane d13C from -63 per mil to -35 per mil directly above the ice-bonded permafrost suggest that methane is effectively oxidized within the overlying unfrozen sediment before it reaches the water column. High rates of methane ebullition into the water column observed elsewhere are thus unlikely to have ice-bonded permafrost as their source.

(Table 1) Sulfur isotopes and sulfur content at DSDP Leg 64 Holes

Hydrothermal pyrite samples from Holes 477, 477A, and 478 have sulfur isotope values ranging from about 10 to 11 per mil. Samples with negative sulfur isotope values generally have low sulfide sulfur contents and reflect mixed bacterial and hydrothermal sulfur sources. At higher sulfur contents, the hydrothermal component predominates, producing positive isotope values. Hydrothermal sulfide derives from reduction of seawater sulfate and may contain a significant basaltic component. Hydrothermal anhydrite is restricted to a narrow zone beneath a dolerite sill at Site 477 and, because of partial sulfate reduction in the circulating waters, has isotopic values (23.5-25 per mil), heavier than seawater.

Compositional and isotopic data on core MD84641

The organic carbon isotopic record of the sapropels(S1 and S3-S10) and intercalated marl oozes has been determined in a 12-m piston core from the eastern Mediterranean. The d13C_organic values are systematically lighter (mean=-21.0±0.82 per mil) in all sapropels and heavier (mean=-18.8±1.07 per mil) in the marl oozes. These differences are not due to variable marine and terrestrial organic matter mixtures because all values are heavier than modern plankton in the Mediterranean, there is no relationship between the C_organic/N ratios and the isotopic values, and published information on the abundance and distribution of organic biomarkers shows that terrestrial material constitutes a minor fraction of the total organic matter. Temperature effects on isotope fractionation are also discounted because the change in d13C_organic values between glacial and interglacial horizons is in the opposite sense. Diagenesis, which can produce relatively small changes in the carbon isotopic composition of sedimentary organic matter under certain circumstances, is unlikely to have caused the observed differences because this mechanism would cause an enrichmet in 12C, implying that all values were even heavier originally, and there is no secular trend in the d13C_organic record. The observed differences in d13C_organic between the two lithologies are probably produced by changes in the isotopic composition and the concentration of dissolved CO2. First, freshwater flooding during the formation of the sapropels caused the isotopic composition of the dissolved inorganic carbon in the surface waters of the Mediterranean to become lighter because of the 13C deficiency in fresh waters. Hence photosynthesis would have produced isotopically lighter organic material. Second, changes in atmospheric pCO2 between glacial and interglacial periods, as shown by the Vostok ice core, caused marked changes in the concentration of free dissolved CO2 in the mixed layer; lower values during glacial maxima caused a smaller fractionation of the carbon isotopes by phytoplankton, whereas levels were less limiting during the interglacials. Concentrations of dissolved CO2 could also have been much higher during the deposition of the sapropels because of the supply of regenerated CO2 to the mixed layer by upwelling, and this could have further lightened the d13C_organic values in the sapropels themselves. Carbon isotope records may provide an alternative method for estimating atmospheric pCO2 levels over longer time periods than can be obtained from ice cores.

Organic geochemistry records from two Palaeocene-Eocene Thermal Maximum sections in Denmark

Analysis of sediments deposited at different latitudes around the world during the Palaeocene-Eocene Thermal Maximum (PETM; ~56 Ma) have revealed a globally profound warming phase, regionally varying from 5-8 °C. Such records from Europe have not yet been obtained. We studied the variations in sea surface and continental mean annual air temperatures (SST and MAT, respectively) and the distribution patterns and stable carbon isotopes of higher plant derived n-alkanes in two proximal PETM sections (Fur and Store Bælt, Denmark) from the epicontinental North Sea Basin. A negative carbon isotope excursion (CIE) of 4-7 per mil was recorded in land plant derived n-alkanes, similar to what has been observed for other PETM sections. However, differences observed between the two proximal sites suggest that local factors, such as regional vegetation and precipitation patterns, also influenced the CIE. The presence of S-bound isorenieratene derivatives at the onset of the PETM and increased organic carbon contents points to a rapid shift in depositional environment; from well oxygenated to anoxic and sulfidic. These euxinic conditions are comparable with those during the PETM in the Arctic Ocean. SSTs inferred from TEX86 show relatively low temperatures followed by an increase of ~7 °C across the PETM. At the Fur section, a remarkably similar temperature record was obtained for MAT using the MBT'/CBT proxy. However, the MAT record of the Store Bælt section did not reveal this warming.

Middle Eocene bulk sediment, benthic and planktic foraminiferal stable isotope values, and relative weight percent CaCO3 content from ODP Site 207-1260

Major ice sheets were permanently established on Antarctica approximately 34 million years ago, close to the Eocene/ Oligocene boundary, at the same time as a permanent deepening of the calcite compensation depth in the world's oceans. Until recently, it was thought that Northern Hemisphere glaciation began much later, between 11 and 5million years ago. This view has been challenged, however, by records of ice rafting at high northern latitudes during the Eocene epoch and by estimates of global ice volume that exceed the storage capacity of Antarctica at the same time as a temporary deepening of the calcite compensation depth 41.6 million years ago. Here we test the hypothesis that large ice sheets were present in both hemispheres 41.6 million years ago using marine sediment records of oxygen and carbon isotope values and of calcium carbonate content from the equatorial Atlantic Ocean. These records allow, at most, an ice budget that can easily be accommodated on Antarctica, indicating that large ice sheets were not present in the Northern Hemisphere. The records also reveal a brief interval shortly before the temporary deepening of the calcite compensation depth during which the calcite compensation depth shoaled, ocean temperatures increased and carbon isotope values decreased in the equatorial Atlantic. The nature of these changes around 41.6 million years ago implies common links, in terms of carbon cycling, with events at the Eocene/Oligocene boundary and with the 'hyperthermals' of the Early Eocene climate optimum. Our findings help to resolve the apparent discrepancy between the geological records of Northern Hemisphere glaciation and model results that indicate that the threshold for continental glaciation was crossed earlier in the Southern Hemisphere than in the Northern Hemisphere.

Macroturbidites of ODP Hole 103-641C

Barremian through uppermost Aptian strata from ODP Hole 641C, located upslope of a tilted fault block on the Galicia margin (northwest Spain), are syn-rift sediments deposited in the bathyal realm and are characterized by rapid sedimentation from turbidity currents and debris flows. Calcarenite and calcirudite turbidites contain shallow-water carbonate, terrigenous, and pelagic debris, in complete or partial Bouma sequences. These deposits contain abraded micritized bioclasts of reefal debris, including rudist fragments. The youngest turbidite containing shallow-water carbonate debris at Site 641 defines the boundary between syn-rift and post-rift sediments; this is also the boundary between Aptian and Albian sediments. Some Aptian turbidites are partially silicified, with pore-filling chalcedony and megaquartz. Adjacent layers of length-fast and -slow chalcedony are succeeded by megaquartz as the final pore-filling stage within carbonate reef debris. Temperatures of formation, calculated from the oxygen isotopic composition of the authigenic quartz, are relatively low for formation of quartz but are relatively warm for shallow burial depths. This quartz cement may be interpreted as a rift-associated precipitate from seawater-derived epithermal fluids that migrated along a fault associated with the tilted block and were injected into the porous turbidite beds. These warm fluids may have cooled rapidly and precipitated silica at the boundaries of the turbidite beds as a result of contact with cooler pore waters. The color pattern in the quartz cement, observed by cathodoluminescence and fluorescence techniques, and changes in the trace lement geochemistry mimic the textural change of the different quartz layers and indicates growth synchronism of the different quartz phases. Fluorescence petrography of neomorphosed low-Mg-calcite bioclasts in the silicified turbidites shows extensive zonation and details of replacive crystal growth in the bioclasts that are not observed by cathodoluminescence. Fluorescence microscopy also reveals a competitive growth history during neomorphism of the adjacent crystals in an altered carbonate bioclast. Barremian-Aptian background pelagic sediments from Hole 641C have characteristics similar to pelagic sediments from the Blake-Bahama Formation described by Jansa et al. (1979) from the western North Atlantic. Sediments at this site differ from the Blake-Bahama Formation type locality in that the Barremian-Aptian pelagic sediments have a higher percentage of dark calcareous claystone and some turbidites are silicified at Site 641. The stable isotopic composition of the pelagic marlstones from Site 641 is similar to those of other Berriasian-Aptian pelagic sediments from the Atlantic.

Pore water isotopic ratios of CO2 and SO4, and sediment CH4 content of ODP Leg 164 sites

A unique set of geochemical pore-water data, characterizing the sulfate reduction and uppermost methanogenic zones, has been collected at the Blake Ridge (offshore southeastern North America) from Ocean Drilling Program (ODP) Leg 164 cores and piston cores. The d13C values of dissolved CO2 (sum CO2) are as 13C-depleted as -37.7 per mil PDB (Site 995) at the sulfate-methane interface, reflecting a substantial contribution of isotopically light carbon from methane. Although the geochemical system is complex and difficult to fully quantify, we use two methods to constrain and illustrate the intensity of anaerobic methane oxidation in Blake Ridge sediments. An estimate using a two-component mixing model suggests that ~24% of the carbon residing in the sum CO2 pool is derived from biogenic methane. Independent diagenetic modeling of a methane concentration profile (Site 995) indicates that peak methane oxidation rates approach 0.005 µmol/cm**3/yr, and that anaerobic methane oxidation is responsible for consuming ~35% of the total sulfate flux into the sediments. Thus, anaerobic methane oxidation is a significant biogeochemical sink for sulfate, and must affect interstitial sulfate concentrations and sulfate gradients. Such high proportions of sulfate depletion because of anaerobic methane oxidation are largely undocumented in continental rise sediments with overlying oxic bottom waters. We infer that the additional amount of sulfate depleted through anaerobic methane oxidation, fueled by methane flux from below, causes steeper sulfate gradients above methane-rich sediments. Similar pore water chemistries should occur at other methane-rich, continental-rise settings associated with gas hydrates.

Table 1. Pressure, temperature and isotope ratios between H2O and CO2 for all experiments

Traditionally, the application of stable isotopes in Carbon Capture and Storage (CCS) projects has focused on d13C values of CO2 to trace the migration of injected CO2 in the subsurface. More recently the use of d18O values of both CO2 and reservoir fluids has been proposed as a method for quantifying in situ CO2 reservoir saturations due to O isotope exchange between CO2 and H2O and subsequent changes in d18OH2O values in the presence of high concentrations of CO2. To verify that O isotope exchange between CO2 and H2O reaches equilibrium within days, and that d18OH2O values indeed change predictably due to the presence of CO2, a laboratory study was conducted during which the isotope composition of H2O, CO2, and dissolved inorganic C (DIC) was determined at representative reservoir conditions (50°C and up to 19 MPa) and varying CO2 pressures. Conditions typical for the Pembina Cardium CO2 Monitoring Pilot in Alberta (Canada) were chosen for the experiments. Results obtained showed that d18O values of CO2 were on average 36.4±2.2 per mil (1 sigma, n=15) higher than those of water at all pressures up to and including reservoir pressure (19 MPa), in excellent agreement with the theoretically predicted isotope enrichment factor of 35.5 per mil for the experimental temperatures of 50°C. By using 18O enriched water for the experiments it was demonstrated that changes in the d18O values of water were predictably related to the fraction of O in the system sourced from CO2 in excellent agreement with theoretical predictions. Since the fraction of O sourced from CO2 is related to the total volumetric saturation of CO2 and water as a fraction of the total volume of the system, it is concluded that changes in d18O values of reservoir fluids can be used to calculate reservoir saturations of CO2 in CCS settings given that the d18O values of CO2 and water are sufficiently distinct.

Mineralogy, geochemistry and benthic foraminifera of the Gorrondatxe beach section

A distinctive low-carbonate interval interrupts the continuous limestone-marl alternation of the deep-marine Gorrondatxe section at the early Lutetian (middle Eocene) C21r/C21n Chron transition. The interval is characterized by increased abundance of turbidites and kaolinite, a 3 per mil decline in the bulk d13C record, a >1 per mil decline in benthic foraminiferal d13C followed by a gradual recovery, a distinct deterioration in foraminiferal preservation, high proportions of warm-water planktic foraminifera and opportunistic benthic foraminifera, and reduced trace fossil and benthic foraminiferal diversity, thus recording a significant environmental perturbation. The onset of the perturbation correlates with the C21r-H6 event recently defined in the Atlantic and Pacific oceans, which caused a 2°C warming of the seafloor and increased carbonate dissolution. The perturbation was likely caused by the input of 13C-depleted carbon into the ocean-atmosphere system, thus presenting many of the hallmarks of Paleogene hyperthermal deposits. However, from the available data it is not possible to conclusively state that the event was associated with extreme global warming. Based on our analysis, the perturbation lasted 226 kyr, from 47.44 to 47.214 Ma, and although this duration suggests that the triggering mechanism may have been similar to that of the Paleocene-Eocene Thermal Maximum (PETM), the magnitude of the carbon input and the subsequent environmental perturbation during the early Lutetian event were not as severe as in the PETM.

Hydrocarbon gar in sediments of the Costa Rica ubduction zone

At the active continental margin off Costa Rica substantial amounts of hydrocarbon gases are encountered in sediments. The molecular composition (C1-C3) of free hydrocarbon gas as well as the isotopic composition (d13C of methane and ethane and D of methane) was analysed on core samples (ranging between 50 and 380 m depth) collected at sites 1040-1043 which was drilled during ODP Leg 170. In addition, the molecular composition of the C1-C3 hydrocarbons and the d13C composition of C1 and C2 hydrocarbons was determined on adsorbed gas from selected depth intervals at Site 1041 (50-380 mbsf). The molecular composition, and stable carbon and hydrogen isotope signature of low molecular weight hydrocarbons from core sediments and gas pockets indicate that most of the gas was generated by microbial CO2-reduction. Beside d13C values of about -80 per mil for methane (which is typical for microbially- generated methane) extremely light d13C values of -55 per mil were measured for ethane. The carbon isotope composition of methane and ethane, as well as the C1/(C2+C3) ratio display distinct trends with increasing depth. Gas mixing calculations indicate that the percentage of thermally-generated ethane increases from 10% at about 75 mbsf to almost 80% at 380 mbsf. The fraction of thermogenic methane in this depth interval is calculated to range from 0.03 to 1.8% of the total methane. The small contribution of thermogenic methane would increase the d13C value by <1 per mil. Therefore, the increase of d13C of methane (by about 12 per mil) with depth cannot be explained by gas mixing alone. Instead, the observed d13C trend is caused by successive isotope depletion of the methane precursor within the sedimentary organic matter due to progressing microbial gas generation.