In search of an ice core signal to differentiate between source-driven and sink-driven changes in atmospheric methane
2011; American Geophysical Union; Volume: 116; Issue: D5 Linguagem: Inglês
10.1029/2010jd014878
ISSN2156-2202
AutoresJ. G. Levine, Eric Wolff, A. E. Jones, M. A. Hutterli, Oliver Wild, G. D. Carver, J. A. Pyle,
Tópico(s)Cryospheric studies and observations
ResumoJournal of Geophysical Research: AtmospheresVolume 116, Issue D5 Composition and ChemistryFree Access In search of an ice core signal to differentiate between source-driven and sink-driven changes in atmospheric methane J. G. Levine, J. G. Levine javi@bas.ac.uk British Antarctic Survey, Cambridge, UKSearch for more papers by this authorE. W. Wolff, E. W. Wolff British Antarctic Survey, Cambridge, UKSearch for more papers by this authorA. E. Jones, A. E. Jones British Antarctic Survey, Cambridge, UKSearch for more papers by this authorM. A. Hutterli, M. A. Hutterli British Antarctic Survey, Cambridge, UK Now at TOFWERK AG, Thun, Switzerland.Search for more papers by this authorO. Wild, O. Wild Lancaster Environment Centre, University of Lancaster, Lancaster, UKSearch for more papers by this authorG. D. Carver, G. D. Carver Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, Cambridge, UK National Centre for Atmospheric Science, University of Cambridge, Cambridge, UKSearch for more papers by this authorJ. A. Pyle, J. A. Pyle Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, Cambridge, UK National Centre for Atmospheric Science, University of Cambridge, Cambridge, UKSearch for more papers by this author J. G. Levine, J. G. Levine javi@bas.ac.uk British Antarctic Survey, Cambridge, UKSearch for more papers by this authorE. W. Wolff, E. W. Wolff British Antarctic Survey, Cambridge, UKSearch for more papers by this authorA. E. Jones, A. E. Jones British Antarctic Survey, Cambridge, UKSearch for more papers by this authorM. A. Hutterli, M. A. Hutterli British Antarctic Survey, Cambridge, UK Now at TOFWERK AG, Thun, Switzerland.Search for more papers by this authorO. Wild, O. Wild Lancaster Environment Centre, University of Lancaster, Lancaster, UKSearch for more papers by this authorG. D. Carver, G. D. Carver Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, Cambridge, UK National Centre for Atmospheric Science, University of Cambridge, Cambridge, UKSearch for more papers by this authorJ. A. Pyle, J. A. Pyle Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, Cambridge, UK National Centre for Atmospheric Science, University of Cambridge, Cambridge, UKSearch for more papers by this author First published: 11 March 2011 https://doi.org/10.1029/2010JD014878Citations: 14AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Abstract [1] The concentration of atmospheric methane increased from around 360 ppbv at the last glacial maximum (∼20 ka before present) to about 700 ppbv in the pre-industrial era (∼200 years before present). The sources and/or sinks of methane must therefore have changed during this period; however, the relative sizes of the source- and sink-driven changes in methane concentration remain uncertain. We take the first “bottom-up” approach to identifying any chemical signals preserved in the ice record that could help us to determine these. Using an atmospheric chemistry-transport model, we explore the effects of source- and sink-driven changes in methane on a wide range of chemical species in the Antarctic boundary layer. Though we identify several potentially useful atmospheric signals, a simple and robust constraint on the sizes of the source- and sink-driven changes cannot be readily identified, owing to their preservation in the ice, limitations to the information they hold, and/or ambiguity surrounding their interpretation. This includes the mass-independent fractionation of oxygen isotopes in sulfates, and the concentration of formaldehyde, in which there has been considerable interest. Our exploration is confined to a domain in which NOx emissions and climate remain constant. However, given the uncertainties associated with the changes in these factors, we would anticipate that their inclusion would make it harder still to identify a robust signal. Finally, though formaldehyde cannot provide this, we propose how it might be used to synchronize the gas- and aqueous-phase Antarctic ice records and thus determine the relative phasing of glacial-interglacial changes in Southern Hemisphere CO2 and temperature. 1. Introduction [2] The concentration of atmospheric methane trapped in Antarctic ice shows large variations over the last 800,000 years that appear to track changes in temperature on orbital, and shorter, time scales [Jouzel et al., 2007; Loulergue et al., 2008]; see, for example, Chappellaz et al. [1993a] for details of the phase relationship between changes in methane and climate in Greenland. This study focuses on the increase in methane from around 360 ppbv at the last glacial maximum (LGM; ∼20 ka before present) to approximately 700 ppbv in the pre-industrial era (PI; ∼200 years before present). Fundamentally, the increase must have resulted from a change in methane sources, such as emissions from wetlands–the largest natural source of methane–and/or a change in methane sinks, the dominant one being oxidation by the hydroxyl radical (OH), but the balance between these changes remains uncertain. Our aim is to identify any atmospheric chemical signals that could be used to differentiate between changes in methane emissions and changes in OH, and assess the likelihood they are preserved in Antarctic ice. [3] Estimates of the changes in methane emissions between the LGM and the PI vary, with much of the discussion focusing on the change in emissions from wetlands. Based on a reconstruction of vegetation, Chappellaz et al. [1993b] estimated that wetland emissions increased by 80% between the LGM and the PI, underpinning a 46% increase in total methane emissions that could explain around half the 94% increase in methane concentration (from 360 to 700 ppbv). Bottom-up studies employing dynamic global vegetation models, on the other hand, have calculated much smaller increases in wetland emissions, ranging from effectively no change [Kaplan et al., 2006] to an increase of 36% [Valdes et al., 2005], and consequently more modest increases in total methane emissions, of 19% and 31%, respectively. Recent calculations by Weber et al. [2010], based on climate simulations from the second phase of the Paleoclimate Modeling Intercomparison Project (PMIP2) [Braconnot et al., 2007], suggest wetland emissions increased by between 54 and 72%, and highlight the uncertainty that remains in the size of the source-driven component of the LGM-PI change in methane concentration. [4] The estimated changes in methane emissions cited above suggest changes in methane sources alone cannot fully account for the LGM-PI change in methane, and a significant fraction of this must have been sink-driven (i.e., the result of an increase in the atmospheric lifetime of methane). Valdes et al. [2005] and Kaplan et al. [2006] showed how the concentration of OH could have decreased, and hence the lifetime of methane increased, as a result of an increase in the amount of nonmethane volatile organic compounds (NMVOCs) emitted from vegetation. An increase in NMVOC emissions between the LGM and the PI is consistent with vegetation reconstructions [see, e.g., Adams et al., 2001] and model studies of the influence that rising temperatures have on vegetation [e.g., Lathière et al., 2005]. However, the size of this increase remains uncertain [see, e.g., Arneth et al., 2007], since some laboratory studies have identified a reduction in NMVOC emissions as the concentration of CO2 to which vegetation is exposed is increased [e.g., Wilkinson et al., 2009], and the concentration of CO2 rose between the LGM and the PI [Monnin et al., 2001]. The possible ‘recycling’ of OH consumed in isoprene oxidation [Lelieveld et al., 2008] adds to the uncertainty in the influence NMVOC emissions had on OH and, hence, the size of the sink-driven component of the change in methane concentration. [5] Changes in the amount of nitrogen oxides (NOx = NO + NO2) emitted from climate-sensitive sources, such as soils, lightning and biomass burning, could have also affected the lifetime of methane with respect to OH. All else being equal, we would expect an increase (decrease) in NOx emissions to have led to an increase (decrease) in the concentration of ozone (O3), and hence an increase (decrease) in the concentration of OH, since OH is formed from the reaction between excited O(1D) oxygen atoms (derived from O3) and water vapor. Valdes et al. [2005] estimated that NOx emissions from lightning increased from 3.4 to 4.2 Tg N yr−1, while those from soils decreased from 5.7 to 5.1 Tg N yr−1, leading to an overall 6% reduction in the lifetime of methane in the PI, relative to the LGM. However, their estimates of the lightning and soil source strengths are subject to large uncertainties; the emissions from lightning, to which the lifetime of methane is especially sensitive, are uncertain in the present, never mind the past [Wild, 2007]. It is also unclear how NOx emissions from biomass burning changed during this period. Calculations by Thonicke et al. [2005] using a dynamic global vegetation model with an embedded fire module suggested these could have decreased in the tropics, and thus contributed to an increase in the lifetime of methane. However, based on a suite of Monte Carlo calculations, Fischer et al. [2008] concluded that the amount of biomass burning was roughly constant between the LGM and the PI, and a global synthesis of charcoal records by Power et al. [2008] identified the last glacial period (16–21 ka before present) as the period of least biomass burning in the last 21 kyr, and hence an increase in biomass burning emissions from the LGM to the PI. [6] The change in climate itself would have also affected the lifetime of methane. Contrary to the increase that could help to explain the observed increase in methane concentration, the increase in temperatures, and hence humidities, would have tended to reduce the lifetime of methane by increasing the (temperature-dependent) rate of reaction between OH and methane, and increasing the concentration of OH. Martinerie et al. [1995] estimated that the increase in humidity would have led to a 6% reduction in the lifetime of methane, while Valdes et al. [2005] estimated that this, together with the increased rate of reaction between OH and methane, would have led to a reduction of 14%. These estimates, however, are subject to significant uncertainties in the changes in temperature between the LGM and the PI [e.g., Braconnot et al., 2007], particularly in the tropics where roughly three quarters of methane oxidation occurs [see, e.g., Lawrence et al., 2001; Labrador et al., 2004]. Proxy data collated and analyzed within the Multiproxy Approach for the Reconstruction of the Glacial Ocean Surface Project [MARGO Project Members, 2009] show an average sea surface–temperature warming between 15°N and 15°S of 1.7(±1.0)°C between the LGM and the PI, and the large, relative uncertainty in this figure (almost 60%) is echoed in the wide range of warmings exhibited by state-of-the-art climate models, from 1.0 to 2.4°C [Otto-Bliesner et al., 2009]. What is not in question, is that the warming climate would have tended to reduce the lifetime of methane. The most likely explanation identified in the literature for a net increase in methane lifetime from the LGM to the PI, which would help to explain the increase in methane concentration, depends upon an increase in NMVOC emissions (discussed above). [7] Previous observational studies suggesting the concentration of OH has changed in the past have largely taken a ‘top-down’ approach, attributing changes in the concentration, or isotopic composition, of atmospheric constituents trapped or dissolved in polar ice to changes in oxidizing capacity. A variety of chemical signals with which to identify changes in OH have thus been proposed, including: the concentration of formaldehyde [Staffelbach et al., 1991]; the concentrations of hydrogen peroxide and methyl-peroxide [Gillett et al., 2000]; and the mass-independent fractionation of oxygen isotopes in sulfates [Savarino et al., 2003]. Studies taking a top-down approach implicitly assume the concentration, or isotopic composition, of a constituent trapped or dissolved in the ice bears some relation to the concentration, or isotopic composition, of that constituent in the boundary layer during the period in which gases could freely exchange between the atmosphere and the then uncompacted snow, or ‘firn.’ In the case of methane, its concentration in air trapped in the ice is almost identical to its concentration in the atmosphere. However, the relationship can be considerably more complicated, as is the case for formaldehyde, the concentration of which in ice is modified during and postdeposition [Hutterli et al., 1999, 2002]. [8] In contrast to previous studies, we take a ‘bottom-up’ approach, using the Cambridge parallelised-Tropospheric Offline Model of Chemistry and Transport (p-TOMCAT), which is described in section 2.1, to explore changes in the chemical composition of the Antarctic boundary layer accompanying source- and sink-driven changes in methane. We carry out six experiments: a PI model run, which is described in section 2.2, and five sensitivity experiments derived from this, which are described in section 2.3. Note that we are not trying to simulate the LGM-PI change in methane. Instead, these are highly idealized experiments, exploring extreme scenarios, in an effort to identify any chemical signal (a chemical species, or combination of species, among those included in p-TOMCAT) that shows potential to constrain the cause(s) of this change. We examine the results of these experiments in sections 3.1 and 3.2. In sections 3.3 and 3.4, we briefly assess the sensitivity of our results to interannual variations in meteorology and the definition of the Antarctic boundary layer, respectively, before discussing in section 4 the preservation in Antarctic ice of those signals which show some potential, and issues surrounding their interpretation. 2. Method 2.1. The Cambridge p-TOMCAT Model [9] For the purposes of this study, we use the Cambridge p-TOMCAT model of tropospheric chemistry and transport, which has been used in a variety of recent studies [Köhler et al., 2008; Yang et al., 2008; Cook et al., 2007; and Levine et al., 2007]. This is a three-dimensional Eulerian model driven by wind, temperature and humidity fields taken from the operational analyses of the European Centre for Medium-range Weather Forecasts (ECMWF). Here, the model is run at a horizontal resolution of approximately 2.8° × 2.8° on 31 levels, which stretch from the surface to 10 hPa with a typical spacing of about 100 m in the boundary layer and 1–1.5 km in the vicinity of the tropopause. Tracer advection is calculated with the second-order moments advection scheme of Prather [1986], as implemented by Chipperfield [1996]. Transport in the horizontal is driven directly by the ECMWF winds while vertical transport is calculated based on the convergence/divergence of winds in the horizontal. Strong convergence also triggers convection, which is simulated using the mass flux parameterization of Tiedtke [1989]. The model also contains a nonlocal vertical diffusion scheme for the boundary layer based on the parameterization of Holtslag and Boville [1993]; see Wang et al. [1999] for details of its implementation. [10] The model chemistry includes 52 chemical species and 174 reactions, which describe the gas-phase HOx/NOx chemistries of methane, ethane, propane and isoprene, the latter according to the Mainz Isoprene Mechanism [Pöschl et al., 2000]; the model also includes a simple ethene-like tracer, which is simply emitted, and removed by OH. The chemistry scheme is of medium complexity, comparable with the schemes employed in other tropospheric chemistry-transport models, such as the troposphere-only version of the UK Chemistry Aerosol Community Model currently under development (see http://www.ukca.ac.uk), and is suitable for global integrations spanning years to decades. The bimolecular and termolecular reaction-rate coefficients (last updated in March 2005) are taken from the International Union of Pure and Applied Chemistry (R. Atkinson et al., Summary of Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry, 2005, http://www.iupac-kinetic.ch.cam.ac.uk) (hereinafter Atkinson et al., http://www.iupac-kinetic.ch.cam.ac.uk, 2005) and the Master Chemical Mechanism (see http://mcm.leeds.ac.uk/MCM/home.htt), while the photolysis rates are calculated offline using the Cambridge 2D model [Law and Pyle, 1993], which is also used to provide top-boundary conditions for ozone and NOy. The concentrations of species are integrated forward in time using the ASAD code of Carver et al. [1997] and the IMPACT time integrator of Carver and Stott [2000]. The wet deposition of soluble species is linked to the parameterization of convection (and large-scale rainfall) according to Walton et al. [1988]. The dry deposition of species at the surface is calculated using prescribed 1 m deposition velocities [Valentin, 1990], extrapolated to the center of each grid box in the lowest level of the model according to Berntsen and Isaksen [1997]. For more details on the implementation and validation of the wet and dry deposition schemes, see Giannakopoulos et al. [1999]. [11] The model does not include halogen chemistry (including the oxidation of methane initiated by atomic chlorine), snow photochemistry or the loss of methane to soils. Platt et al. [2004] estimate that atomic chlorine is responsible for about 3% of global methane loss. We would not expect the inclusion of an additional sink of this size to alter our findings significantly, particularly those of a qualitative nature. Snow photochemistry will affect OH in the boundary layer in snow covered regions [see, e.g., Grannas et al., 2007]. We would have to include this if our aim was to reproduce absolute OH concentrations measured in such regions, but it should not have much effect on the global oxidizing capacity. Last, the loss of methane to soils is omitted on the basis that it accounts for only 5–10% of global methane loss, which is small compared to the uncertainty in the PI emissions of methane implemented in the model [Valdes et al., 2005]; see section 2.2. 2.2. The PI Model Run [12] For the purposes of this experiment, we have implemented as far as practically possible the same PI emissions in the Cambridge p-TOMCAT model as Valdes et al. [2005] used in the STOCHEM atmospheric chemistry-transport model. The emissions, summarized in Table 1, include: seasonally varying emissions of nitrogen dioxide (NO2), methane (CH4), carbon monoxide (CO), ethane (C2H6), propane (C3H8), acetone (CH3COCH3), isoprene (C5H8) and ethene (C2H4), in addition to constant emissions of formaldehyde (HCHO) and acetaldehyde (CH3CHO). Note that, unlike Valdes et al. [2005], we do not include emissions of butane, propene, methanol or α-pinene, owing to the lack of these species in p-TOMCAT; hydrogen is not emitted but is included in the model as a constant field. Lightning emissions of NO2 in p-TOMCAT are coupled in time and space to the parameterization of convection [Stockwell et al., 1999], and are therefore unlikely to be distributed in the same way, either spatially or temporally, as the lightning emissions of NO in STOCHEM. We note, however, that the total amount of NOx emitted from lightning in p-TOMCAT (normalized to 4.8 Tg N/yr) is not dissimilar to that in STOCHEM (4.2 Tg N/yr) [Valdes et al., 2005]. Table 1. Trace Gas Emissions Used in the PI Model Run, Expressed in Terms of Molecular Mass (Tg) per Yearaa Except for the emissions of NO2 from lightning, these emissions are as close as practically possible to those used by Valdes et al. [2005] in their study of the LGM-PI change in methane, using the STOCHEM chemistry-transport model; see section 2.2 for more details. Trace Gas Biomass Burning Oceans Vegetation Soil Lightning Wetlands Termites Total NO2 4.7 - - 16.8 15.8 - - 37.3 CH4 11.0 13.0 - - - 147.9 27.0 198.9 CO 100.0 50.0 150.0 - - - - 300.0 C2H6 0.7 - 3.5 - - - - 4.2 C3H8 0.2 0.5 3.5 - - - - 4.2 CH3COCH3 0.1 - 20.0 - - - - 20.1 C5H8 - - 673.7 - - - - 673.7 C2H4 1.4 - 20.0 - - - - 21.4 HCHO 0.3 - - - - - - 0.3 CH3CHO 0.8 - - - - - - 0.8 a Except for the emissions of NO2 from lightning, these emissions are as close as practically possible to those used by Valdes et al. [2005] in their study of the LGM-PI change in methane, using the STOCHEM chemistry-transport model; see section 2.2 for more details. [13] Figure 1 illustrates the distributions of these emissions (excluding lightning emissions of NOx), integrated over a year. The emissions of methane are dominated by emissions from wetlands in equatorial South America and Indonesia, and to a lesser extent, Central Africa and South East Asia. The relatively small amounts of methane coming from the oceans, biomass burning and termites (see Table 1) account for its near-ubiquitous, low-level emissions elsewhere around the globe. The emissions of isoprene, which exert a strong influence on OH as a result of isoprene's short lifetime with respect to oxidation (Atkinson et al., http://www.iupac-kinetic.ch.cam.ac.uk, 2005), are concentrated in equatorial regions, but also appreciable in the subtropics of both hemispheres. The remaining emissions categorized under ‘vegetation’ in Table 1 are all distributed in the same way as each other, mostly in the tropics and subtropics but with a significant contribution from midlatitudes, particularly in the Northern Hemisphere. This distribution closely resembles the distribution of acetone emissions, the latter being overwhelmingly dominated by emissions from vegetation (see Table 1). Figure 1Open in figure viewerPowerPoint Distributions of the trace-gas emissions used in the PI model run, integrated over 1 year, expressed in terms of molecular mass (Tg) per grid box. They are as close as practically possible to those used by Valdes et al. [2005] in their study of the LGM-PI change in methane. NB The distribution of NO2 emissions excludes those from lightning; see section 2.2 for more details. [14] The PI model run (and each sensitivity experiment) follows a certain format. The model is first run to equilibrium using repeated meteorology from a single year (1997). At ‘equilibrium,’ neither the global burden of methane nor the annual mean concentration of methane in the Antarctic boundary layer (AntBL; defined to comprise all boxes in the lowest level of the model south of 70°S) changes by more than 0.02% per year. The model is then run with meteorology from three further years (1998–2000) to explore interannual variability. These three years were chosen on the basis that they include contrasting phases of the El Niño–Southern Oscillation (ENSO), a major contributor to interannual variations in the distributions of trace gases within the troposphere [e.g., Chandra et al., 1998; Ziemke and Chandra, 2003]; the Northern Hemisphere winters of 1997/1998 and 1998/1999 were El Niño and La Niña, respectively, while 2000 was less affected by the ENSO. The data gathered in these three years are then used to characterize the chemical composition of the AntBL. Note, we assume recent meteorological analyses adequately represent the meteorology of the PI. 2.3. The Sensitivity Experiments [15] We carry out five sensitivity experiments, each starting from the PI model setup; these are schematically illustrated in Figure 2. The first two experiments, Sink 1 and Source 1, are designed to identify any chemical signal that responds in a substantially different manner to purely source-driven and purely sink-driven changes in methane (of the same size), making no assumptions about the cause(s) of the latter. In Source 1, we reduce the concentration of methane in the AntBL to roughly that which was present at the LGM by scaling down the emissions of methane (uniformly) by 45%. In Sink 1, we increase the production of OH by a factor of 2.5 to effect the same change in methane concentration. We do so by increasing the number of OH radicals produced as each excited oxygen atom, O(1D), reacts with a water molecule. This is simply a way of increasing the concentration of OH without making any assumptions about the cause of this increase, which would have chemical consequences. For example, if we were to instead increase the rate of this reaction, we would not only increase the rate of OH formation, but also the rate of O(1D) removal. The third and fourth experiments, Sink 2 and Source 2, are designed to assess whether there exists a signal capable of differentiating between source- and sink-driven changes in methane, when the latter are the result of changes in the amount of NMVOCs emitted from vegetation. In Sink 2, all emissions from vegetation (i.e., those categorized under ‘vegetation’ in Table 1) are switched off, leading to a substantial reduction in methane, but a smaller one than in Sink 1 or Source 1. The same reduction in methane is achieved in Source 2, as in Sink 2, by scaling down the emissions of methane by 17%. In the fifth experiment, Sink+Source, we remove all emissions from vegetation (as in Sink 2) and scale down the emissions of methane by 31% to reduce the concentration of methane to that which was present at the LGM (as in Sink 1 and Source 1). By comparing the results to this experiment with those to Source 1, we can assess a signal's ability to differentiate between a purely source-driven change in methane, and one which is part source-driven, part sink-driven. Figure 2Open in figure viewerPowerPoint Schematic illustration of the five sensitivity experiments, described in section 2.3, each starting from the PI model setup, described in section 2.2. In Source 1, the emissions of methane are reduced (ΔECH4) so as to reduce the annual mean concentration of methane in the AntBL, [CH4], to roughly that which characterized the LGM, while the rate of OH production is increased (ΔPOH) in Sink 1 to effect the same change in [CH4]. In Sink 2, all NMVOC emissions from vegetation are switched off (ΔENMVOCs), while the emissions of methane are reduced in Source 2 (ΔE′CH4) to effect the same change in [CH4]. Finally, in Sink+Source, all NMVOC emissions from vegetation are switched off (ΔENMVOCs) and the emissions of methane are reduced (ΔE″CH4) so as to reduce [CH4] to roughly that which characterized the LGM, in line with Sink 1 and Source 1. [16] Though we explore in Sink 1 the consequences of a change in OH generated in such as way as to make no assumptions about the cause(s) of this change, we do not explore changes in OH driven explicitly by changes in NOx emissions or climate, which would have influences on the atmospheric composition besides their effects on OH, and presumably different ones to the changes in NMVOC emissions that we explore in Sink 2 and Sink+Source. As outlined in section 1, there are significant uncertainties associated with the changes in NOx emissions and climate between the LGM and the PI, and their influences on the oxidizing capacity were probably subsidiary to the influence of changes in NMVOC emissions. Our approach is therefore to first try to find a signal capable of differentiating between source- and sink-driven changes in methane, subject to constant NOx emissions and climate, and only if we find such a signal (and believe it will be preserved in the ice record on the necessary time scales), explore whether it would prove robust to such changes. 3. Results 3.1. Methane in the Antarctic (and Arctic) Boundary Layer [17] Figure 3 illustrates the concentrations of methane in the AntBL and the Arctic boundary layer (ArcBL; defined to comprise all boxes in the lowest level of the model north of 70°N) that we calculate in the PI model run (Figure 3, top) and the five sensitivity experiments: Sink 1, Source 1 and Sink+Source (Figure 3, bottom); Sink 2 and Source 2 (Figure 3, middle). In each case, the data correspond to the last three years of the model's run to equilibrium, using repeated 1997 meteorology, and the subsequent three years employing meteorology from 1998 to 2000. Figure 3Open in figure viewerPowerPoint The monthly mean concentration of methane modeled in the Arctic boundary layer (ArcBL) and the Antarctic boundary layer (AntBL), plotted as a function of the meteorology used to drive the model in (top) the PI run, and in the five sensitivity experiments: (bottom) Sink 1, Source 1, and Sink+Source and (middle) Sink 2 and Source 2. In each case, the data shown correspond to the last 3 years of the run to “chemical equilibrium” (3 × 1997) and the subsequent 3 year run to gather data (1998–2000). [18] The PI model run (Figure 3, top) yields methane concentrations of 705–720 ppbv in the AntBL and 740–755 ppbv in the ArcBL, amounting to annual mean concentrations, averaged over 1998–2000, of 709 ppbv and 744 ppbv, respectively. (From here on, all annual mean concentrations refer to averages over 1998–2000.) There is thus an average interhemispheric gradient of 35 ppbv, which arises
Referência(s)