13C/12C Ratio in Methane From the Flooded Amazon Forest

Analyses for C t-C, (cid:127) hydrocarbon concentrations and the 13C/12C ratio in CH,(cid:127) were performed on two air samples collected in the Amazon jungle (3.5øS, 59øW) after the nearby release of biogenic gas bubbles. The CH,(cid:127) concentrations of each sample were greatly enhanced (4100 and 310 ppmv) over the back- ground concentration (1.6 ppmv) for remote locations at that latitude and time. The (cid:127)3,/(cid:127)2C ratio in this biogenic methane is depleted in 13C (-64(cid:127),) relative to atmospheric CH,(cid:127) (-47(cid:127),), as is CH,(cid:127) from almost all other biogenic sources. Because laboratory measurements to date indicate only a very small 13C/12C isotope effect in the reaction of CH,(cid:127) with HO, an apparent discrepancy remains between the 13C/12C ratios of the known CH,(cid:127) sources and that of atmospheric CH,(cid:127). Five other hydrocarbons (C2H6, C2H,p C3H 8, i-C,(cid:127)Hlo, n-C,(cid:127)H(cid:127)o ) were also found at the 1 to 35 ppbv level in the air sample with 4100 ppmv CH,(cid:127). These concentrations are not large enough to indicate any major importance for this source in C2-C, (cid:127) hydrocarbon budgets on either a global or regional basis.


INTRODUCTION
The interactions both of methane and of nonmethane hydrocarbons (NMHC) have been of increasing scientific interest as their importance has been recognized relative to such diverse situations as the oxidizing capability of the atmosphere, the greenhouse effect, the reactions of atomic chlorine in the stratosphere, urban smog and other atmospheric chemical problems [NAS, 1984]. The most abundant hydrocarbon in the earth's atmosphere is methane with a worldwide average tropospheric concentration in mid-1985 of !.65 parts per million by volume (ppmv) Rowland, 1985, 1986a]. A very large number of other hydrocarbons ranging from C 2 compounds to the terpenes have also been iden.tified in the atmosphere in regions somewhat distant from the probable sources . The observed concentrations of these other compounds have generally been much lower, often in the 0.0001 to 0.01 ppmv range, and are much more variable than that of CH 4. Both the lower concentrations and the variability are directly related to their atmospheric lifetimes which are very much shorter than the 10 years estimated for CH 4 , and the cumulative carbon flux through the atmosphere in these chemical forms may be comparable to, or larger than, the 340 megatons of carbon per year estimated for methane alone.
The concentration of methane in the troposphere has been increasing at a rate of about 0.017 ppmv year over at least the past eight years to its present value [Rasmussen andKhalil, 1981, 1984;Blake et al., 1982;Khalil and Rasmussen, 1983;Rowland, 1985, 1986a]. Retrospective examination of atmospheric infrared spectra suggest that an increase of about 1'7o per year in CH 4 concentration has been occurring at least since 1951 [Rindand et al., 1985], while measurements of the composition of air bubbles trapped in ice cores indicate that the concentration of CH 4 in the atmosphere may have been only about 0.7 ppmv as recently as two or three hundred Copyright 1987 by the American Geophysical Union.
Paper number 6D0517. 0148-0227/87/006D-0517502.00 years ago [Craig and Chou, 1982' Rasmussen and Khalil, 1984' Stat•#kr et al., 1985. This increase in tropospheric methane concentration has raised important questions about the location and strength of the sources of methane and other hydrocarbons being emitted to the atmosphere. The major sources of atmospheric CH 4 involve anaerobic biology , including swamps, rice paddies, the rumen of cattle, etc.

Cross comparison of the concentrations of CH 4 and
CH3CCl 3 in air samples collected in or near the Amazon region have shown enhanced CH4, leading to a semiquantitative estimate that as much as 10% of the world's CH 4 is emitted in Amazonia . Much more detailed experimental measurements of the magnitude and extent of these source-enhanced concentrations of tropospheric CH 4 have been carried out in a NASA-sponsored program during 1985. Measurements have also been made which indicate that NMHC compounds such as ethane have higher concentrations in some tropic regions than are found at similar latitudes elsewhere Greenberg et al., , 1985. One known source for C2H 6 in Amazonia is biomass burning [Crutzen et al., 1979[Crutzen et al., , 1985 Rice and Claypool, 1981;Stevens and Rust, 1982].
Biomass burning, largely in tropical areas, is an important source of light hydrocarbons, and may be the dominant tropical source for C2-C 4 hydrocarbons . The •3C/•2C ratio in the material cornbusted during most biomass burning contains about -25 to -30%,, [Craig, 1953;Bender, 1968Bender, , 1971Troughton et al., 1974] and probably does not undergo substantial isotopic fractionation during combustion. The methane from biomass burning is therefore probably In the open areas, the surface of the water held scattered masses of floating or partially sunken decaying organic matter, from which bubbles appeared intermittently. These masses could be perturbed by forcing them below the water level, and then allowing them to be buoyed back up naturally with the emission of larger and more frequent bubbles. The two perturbed samples were collected from a canoe floating over ten feet of water by poking the buoyant organic mass with the oars, and then opening the evacuated air sample canisters while held only 2 or 3 inches above the debris. The main source of the gases drawn into the canisters was simply the ambient atmosphere, but an appreciable admixture of the gas bubbles emitted from the debris was included as well. The bubbles had almost reached the actual stage of emission to the atmosphere before being perturbed, and it is likely that any bacterial modification of the 13C/12C isotopic ratio during contact with the water had essentially been completed by then. The disturbance was limited to the floating debris with no perturbation of sediments.

Stable Carbon Isotope Ratios
The stable carbon isotope ratios were measured with a Nuclide 6-60 RMS isotope ratio mass spectrometer at the National Center for Atmospheric Research in Boulder. Details of the experimental procedure for preparation of samples for measurement, and calibration for similar samples will be presented elsewhere . The minimum sample size for these measurements is about 5 micromoles of CH,•, or 75 liters of air with CH,• at its normal background level. The 2-liter sample canisters furnished sufficient CH,• for the standard measurement because the CH,• concentrations were greatly enhanced over the 1.6-1.7 ppmv characteristic of the background. The •3C/•2C ratios for the two samples are reported in per mil variation relative to the conventional PDB carbonate standard. The working standard in this apparatus was cross-calibrated with that used by C. M. Stevens at Argonne National Laboratory. The value for this standard as measured at NCAR was --26.7%(, and at Argonne -26.8%0, both relative to PDB carbonate.

Hydrocarbon Analysis
The analyses for CH4 were performed on aliquots of the gaseous samples by standard gas chromatographic methods using flame ionization detection [Blake and Rowland, 1986a]. Aliquots of samples A-1 and A-2 were diluted on the vacuum line with zero air to bring the measured concentrations into the calibrated range for our instrument. The precision and accuracy are reduced to perhaps 2-3ø/; for these samples, rather than the usual +0.4ø/,,, but none of the conclusions are dependent upon high accuracy in the data. The trace components volatile from a -20cC bath were cryogenically trapped from as much as one-liter STP of air, and then analyzed for C2-C• compounds on a 3-foot Spherocarb column programmed from -10•C to 350øC. This procedure provides a sensitivity generally in the 0.1 ppbv range for one-liter STP air samples [Blake and Rowland, 1986b].

C 2-C 4 Hydrocarbons
The measured concentrations of CH4 and five other hydrocarbons are given in Table 1

for samples A-1 and A-2. The canister (A-l)with
the highest concentration of CH4 contained an excess of all five of the C2-C 4 hydrocarbons measured in these experiments. The accuracy of the C2-C 4 analyses on A-1 and A-2 is judged to be _+5ø/,,, but the representative nature of these samples is unknown because only two are available. The likely sources for all are the same bubbles which produced the excess methane, and the observations are indicative of the probable formation of such compounds in minor yield by anaerobic biological processes. A third air sample collected in the same area with only minor disturbance of the vegetation contained just 4 ppmv CH•, and a fourth with no disturbance contained 1.7 ppmv CH4. Neither of these latter two samples was retained for later analysis for NMHC compounds.
The C•-C4 analyses for air from three additional samples from the southern tropics are also given in Table 1 for comparison. Two of these air samples were collected on Pacific Islands and one was taken on the Brazilian coast ten days prior to the collection of the jungle samples. The air samples from Nauru and Bora Bora are generally typical of background oceanic air from the southern hemisphere with seasonally-dependent concentrations of C2H 6 and C3H 8 and usually <0.1 ppbv of the C4 alkanes [Blake and Rowland, 1986b].
Brazilian "background" air samples, such as B in Table 1, have higher concentrations of C2H4, C2H 6 and C3H 8 than found in samples collected in similar seasonal periods in Pacific Island locations, as in samples C and D, and are presumed to contain additional hydrocarbons from the well-known regional emissions . The atmospheric lifetime of C3H 8 in tropical latitudes has been calculated to be no more than two or Parts per billion by volume, 10-9 three weeks from a comparison of the diminishing concentrations of C3H 8 and 222Rn with increasing distance from continental locations [Bonsang et al., 1985]. The lifetime of

C2H 6 in the tropics is probably about two months [Blake and
Rowland , 1986b].
The concentration of the C2-C 4 compounds found in sample A-2 are within the normal range found for Amazonian continental samples such as B, and show no evidence for appreciably enhanced concentrations from the specific local environment with its perturbed bubble emission. The scatter in such measurements is large enough, however, that contributions from an immediate local source could be present in the tenths of ppbv range. Sample A-l, with 13 times greater CH 4 enhancement than A-2, definitely shows local enhancement of the yields of all five hydrocarbons in Table 1. These data demonstrate that other hydrocarbons in addition to CH 4 are emitted in parallel to the well-known methane emission, and we presume that these C2- x 103 without producing any measurable increase in the concentrations of these three compounds. No quantitative evaluation is possible from our data for the two C4 compounds, but the observed yields in A-1 are small enough relative to that of CH4 to make appreciable enhancement of these unlikely as well. We conclude from these data on the C2-C 4 hydrocarbons that minor yields of such compounds are released coincident with CH4, but that these minor yields are not an important contribution to the regionally enhanced concentrations of these hydrocarbons found in Amazonia, and are even less important to the global atmospheric release of C2-C 4 compounds. These observations are not inconsistent with biomass burning as the major source for C2-C, • compounds in Amazonia ].

•3C/•2C Isotope Ratio in Methane
The CH,• from sample A-1 had an isotope ratio of -64.5 + 0.3%0, while that from sample A-2 had a ratio of -63.3 + 0.3%0. Correction of the isotope ratio measured for sample A-2 for the presence of 1.7 ppmv of atmospheric CH,• with about --47%0 would change the measured ratio for the 308 ppmv of CH4 directly emitted from the Amazon wetlands by only 0.1%o to -63.4 +_ 0.3%0. The correction for sample A-1 is even less significant because of the much higher CH4 concentration found in that canister. The •3C/•2C ratios in both of these samples indicate substantial depletion in •3C relative to the atmosphere, i.e., about -64%0, and therefore do not represent the sought-for missing source of •3C-enriched methane. It is obvious that many more studies are needed before reasonable extrapolations can be made to the entire Amazon basin. Nevertheless, these first two samples suggest that the resolution of the inconsistency between the 13C/12C ratio in methane sources and in atmospheric methane may well not lie in the tropical emission of •3C-enriched CH4.
With essentially all important biological sources of CH4 depleted in •3C relative to atmospheric CH4, the physicochemical processes needed to rationalize these respective isotopic ratios remain to be identified.