Gordon Brailsford

Concentration and 13C records of atmospheric methane in New Zealand and Antarctica : evidence for changes in methane sources

Measurements of 13C in atmospheric methane made at Baring Head, New Zealand (41°S), over the 4-year period, 1989–1993, display a persistent but highly variable seasonal cycle. Values for δ13C peak in summer at about −46.9‰ and drop to around −47.5‰ in the late winter. Methane concentration shows a similar cycle, with winter peaks and summer minima. Similar features are observed at the New Zealand Antarctic station, Scott Base, at 78°S. While the phase of the δ13C cycle is consistent with a kinetic isotope effect that preferentially leaves methane enriched in 13C in the atmosphere after oxidation by OH, the amplitude of the cycle is much larger than expected from published laboratory measurements of the effect. We interpret the origin of this cycle and its inter-annual variability to be due to episodic southward transport of isotopically heavy methane from large-scale tropical biomass burning, possibly in conjunction with changes in the rate of interhemispheric transport in the troposphere. The Baring Head 13C data show no significant secular trend from 1989 to mid-1991, followed by a rapid trend toward methane less enriched in 13C. This indicates a major shift in the balance of the sources of atmospheric methane and precludes an increased sink strength. The trend in 13C since mid-1991 coincided with significant changes to the methane growth rate observed at Baring Head and at Scott Base: an elevated growth rate of about 15 parts per billion by volume (ppbv) during 1991 gave way to less than 3 ppbv yr−1 thereafter. A 2-box model of atmospheric methane (one box per hemispheric reservoir) suggests that (1) the recent decline in 13C in methane observed at Baring Head and Scott Base cannot have a solely northern hemispheric origin and (2) the most plausible origin is a recent reduction in methane released by biomass burning in the southern hemisphere, combined with a lower release rate of fossil methane in the northern hemisphere.

13C12C Fractionation of methane during oxidation in a temperate forested soil

We have made measurements of the 13C12C fractionation of methane (CH4) during microbial oxidation by an upland temperate soil from College Woods, New Hampshire, using both in situ and laboratory incubation measurements. Uptake rates of 1–4.8 mg CH4/m2/d were measured during the active season in New Hampshire while rates of uptake were 2.6–6.8 mg CH4/m2/d in jars used for incubation studies. The fractionation factor, calculated from field measurements, was α = 0.978 ± .004. This corresponds to a kinetic isotope effect (KIE) of ki2k13 = 1.022 ± .004. Only a small dependence on temperature was noted for air temperatures between 281 and 296 K. Our results indicate that the KIE of soil CH4 oxidation is controlled by physical parameters based on gaseous diffusion into the soil. The implications of these results are discussed with respect to the global CH4 budget and balancing CH4 sources and sinks through the use of δ13CH4 measurements.

The 1991–1992 atmospheric methane anomaly: Southern hemisphere 13C decrease and growth rate fluctuations

Measurements of atmospheric methane from 1989-1996 at Baring Head, New Zealand, and at Scott Base, Antarctica show a seasonal cycle in the mixing ratio with a peak to peak amplitude of 28 ppb. This is superposed on a trend varying between 16 ppb yr -1 and near zero. δ 13 C values also show a seasonal cycle, with an amplitude of 0.1-0.3‰, approximately 6 months out of phase with the mixing ratio cycle. A pronounced negative anomaly in δ 13 C occurred in 1992 with annual average values dropping from -47.08‰ to -47.28‰. From 1992 to 1996, average δ 13 C values recovered slowly at an average rate of about 0.04‰ yr -1 . The simultaneous changes in the mixing ratio growth rate and δ 13 C together with the rapid drop and slow recovery in the latter provide a stringent test of possible causes. Although a combination of causes cannot be ruled out, decreased emissions from an isotopically heavy source such as biomass burning best meet the constraints of the data.

14CH4 measurements in Greenland ice: investigating last glacial termination CH4 sources.

Radiocarbon measurements show that wetlands were responsible for the rapid increase of atmospheric methane concentration during the last deglaciation. Methane from Wetlands At the end of the cold climate interval called the Younger Dryas, approximately 11,600 years ago, global temperatures began their final ascent to the warmth of the Holocene, and the concentration of methane in the atmosphere increased rapidly and substantially. There has been much speculation about the cause of that increase, with most recent evidence pointing to wetlands as the source. The most direct proof of that explanation requires the measurement of the radiocarbon content of that methane. Petrenko et al. (p. 506; see the Perspective by Nisbet and Chappellaz) analyzed 1000 kilogramsized samples of Greenland ice, which have sufficient methane to allow measurement of its 14C content. They show that wetland sources indeed must have been responsible for the majority of the rise in atmospheric methane levels at the end of the Younger Dryas. The cause of a large increase of atmospheric methane concentration during the Younger Dryas–Preboreal abrupt climatic transition (~11,600 years ago) has been the subject of much debate. The carbon-14 (14C) content of methane (14CH4) should distinguish between wetland and clathrate contributions to this increase. We present measurements of 14CH4 in glacial ice, targeting this transition, performed by using ice samples obtained from an ablation site in west Greenland. Measured 14CH4 values were higher than predicted under any scenario. Sample 14CH4 appears to be elevated by direct cosmogenic 14C production in ice. 14C of CO was measured to better understand this process and correct the sample 14CH4. Corrected results suggest that wetland sources were likely responsible for the majority of the Younger Dryas–Preboreal CH4 rise.

Shipboard determinations of the distribution of 13C in atmospheric methane in the Pacific

Measurements of the mixing ratio and δ 13 C in methane (δ 13 CH 4 ) are reported from large, clean air samples collected every 2.5° to 5° of latitude on four voyages across the Pacific between New Zealand and the West Coast of the United States in 1996 and 1997. The data show that the interhemispheric gradient for δ 13 CH 4 was highly dependent on season and varied from 0.5‰ in November 1996 with an estimated annual mean of 0.2-0.3‰. The seasonal cycles in δ 13 CH 4 reveal three distinct latitude bands differentiated by phase. Maxima occur in January-February for the extratropical Southern Hemisphere, in September-October for the tropics, and in June-July for the extratropical Northern Hemisphere. The data are compared with results from a three-dimensional transport and atmospheric chemistry model that simulates the observed latitudinal structure of either δ 13 CH 4 or the methane mixing ratio well, but not both simultaneously. The requirement that a methane source-sink budget be consistent with both types of data clearly imposes stricter constraints than arise from either mixing ratio or isotopic data alone. The seasonal δ 13 CH 4 data in the extratropical Southern Hemisphere are used to estimate a value for the net fractionation in the CH 4 sink of 12-15‰, which is larger than can be explained by current laboratory measurements of a kinetic isotope effect for the OH + CH 4 reaction and soil sink processes. The hypothesis that the discrepancy is caused by competitive reaction of active chlorine with methane in the marine boundary layer is discussed.

Seasonal variations in methane flux andδl3CH4 values for rice paddies in Japan and their implications

We have made measurements of the methane (CH4) flux and δ13C value in CH4 from rice paddies in Ryugasaki, Japan. This study is the first we are aware of in which a significant change in the δ13C signature of emitted CH4 has been documented over the rice growing season. Nutrient treatments studied were of two kinds: compound mineral fertilizer either with or without rice straw from the previous growing season incorporated into the inorganic fertilizer. The calculated annual emission rates during the 1990 growing season were 43.1 g/m2 (straw) and 40.6 g/m2 (no straw) for the two treatments. In both treatments, CH4 started out relatively enriched in 13C, became lighter in 13C, and then became more enriched again during the latter part of rice growth. The 1991 growing season showed a lower integrated flux in both nutrient treatments than for 1990 but plots of the fluxes versus time had the same general shape as the flux curves in 1990 and a similar although less pronounced trend in δ13CH4 signal. Seasonal changes in δ13C are probably related to changes in CH4 production and oxidation and plant-mediated transport. The likelihood of each process occurring and its effect on δ13C values is discussed. The range of δ13CH4 values from seasonal effects was ∼12‰ in 1991 for both nutrient treatments. The δ13CH4 range for 1991 was ∼10‰ (straw) and ∼5‰ (no straw). Our data indicate that when using flux-weighted isotopic signatures to put constraints on the tropospheric CH4 budget, attention should be paid to seasonal changes in isotopic signatures from rice paddy CH4 in a manner similar to that previously suggested from measurements in natural wetlands.

A new gas chromatograph‐isotope ratio mass spectrometry technique for high‐precision, N2O‐free analysis of δ13C and δ18O in atmospheric CO2 from small air samples

A new gas Chromatograph-isotope ratio mass spectrometry (GC-IRMS) technique for the first N2O-free, high-precision (<0.05‰) isotopic analysis of δ13C and δ18O in atmospheric CO2 from small air samples has been developed. On-line GC separation of CO2 and N2O from a whole air sample is combined with IRMS under elevated ion source pressures. A specialized open split interface is an integral part of the inlet system and ensures a continuous flow of either sample gas or pure helium to the IRMS. The analysis, including all flushing, uses a total amount of 45 mL of an air sample collected at ambient pressure. Of this, three 0.5 mL aliquots are injected onto the GC column, each providing ∼0.8 nmol CO2 in the IRMS source. At this sample size, δ13C precision obtained is at the theoretical shot noise limit. For typical ambient air samples collected in the Southern Hemisphere, demonstrated precisions for δ13C, δ18O, and the CO2 mixing ratio (all measured simultaneously) are 0.02‰, 0.04‰, and 0.4 ppm, parts per million (ppm) respectively. Since these data are achieved from small air samples without contamination by atmospheric N2O or the use of cryogen, the technique will be a valuable tool in global carbon cycle research.

The Effect of N2O, Catalyst, and Means of Water Vapor Removal on the Graphitization of Small CO2 Samples

The effect of nitrous oxide (N2O) upon the graphitization of small (~40 g of carbon) CO2 samples at the ANSTO and University of California, Irvine, radiocarbon laboratories was investigated. Both laboratories produce graphite samples by reduction of CO2 over a heated iron catalyst in the presence of an excess of H2. Although there are significant differences between the methods employed at each laboratory, it was found that N2O has no effect upon the reaction at levels of up to 9.3% by volume of CO2. Further, it was systematically determined that more effective water vapor trapping resulted in faster reaction rates. Using larger amounts of the Fe catalyst generally resulted in higher yields or reaction rates (but not both). The effects of changing the type of Fe catalyst on the final yield and reaction rate were less clear.

A New Method for Analyzing 14C of Methane in Ancient Air Extracted from Glacial Ice

We present a new method developed for measuring radiocarbon of methane (14CH4) in ancient air samples extracted from glacial ice and dating 11,000-15,000 calendar years before present. The small size (~20 µg CH4 carbon), low CH4 concentrations ((CH4), 400-800 parts per billion (ppb)), high carbon monoxide concentrations ((CO)), and low 14C activity of the samples created unusually high risks of contamination by extraneous carbon. Up to 2500 ppb CO in the air sam- ples was quantitatively removed using the Sofnocat reagent. 14C procedural blanks were greatly reduced through the construc- tion of a new CH4 conversion line utilizing platinized quartz wool for CH4 combustion and the use of an ultra-high-purity iron catalyst for graphitization. The amount and 14C activity of extraneous carbon added in the new CH4 conversion line were determined to be 0.23 ± 0.16 µg and 23.57 ± 16.22 pMC, respectively. The amount of modern (100 pMC) carbon added during the graphitization step has been reduced to 0.03 µg. The overall procedural blank for all stages of sample handling was 0.75 ± 0.38 pMC for ~20-µg, 14C-free air samples with (CH4) of 500 ppb. Duration of the graphitization reactions for small (<25 µg C) samples was greatly reduced and reaction yields improved through more efficient water vapor trapping and the use of a new iron catalyst with higher surface area. 14C corrections for each step of sample handling have been determined. The resulting overall 14CH4 uncertainties for the ancient air samples are ~1.0 pMC.

Verifying agricultural emissions of methane

The most direct way to establish the level of surface emissions of greenhouse gases is to measure and interpret concentration gradients in the atmosphere. We have tested the efficacy of this approach for inferring average methane fluxes from regions of pastoral agriculture a few tens of km in extent In its simplest form, vertical concentration profiles are measured upwind and downwind of the target region, based on air samples collected from light aircraft. Using simple mass balance models, the profile contrasts can be related to the mean surface flux over the intervening region. The inferred flux can then be compared with ‘bottom-up’ estimates based on livestock density and per-animal emissions. However, such simple models may poorly simulate air flows over the New Zealand terrain, and as an alternative, we deploy a state-of-the-art mesoscale meteorological model, RAMS, coupled to an atmospheric dispersion model. RAMS is used prognostically to guide the timing and siting of measurement campaigns, and diagnostically to simulate regional wind fields which are validated against local meteorological data. Source-oriented and receptor-oriented dispersion modelling techniques, in combination with aircraft-based sampling and laboratory gas analysis, provide ‘top-down’ methane flux estimates that compare favourably with ‘bottom-up’ estimates. These techniques thus enhance confidence in national emission inventories based on bottom-up estimation. However, the challenge for similar verification of nitrous oxide emission is more imposing.