Methane Emissions Deduced from a Two-Dimensional Atmospheric ...

0 downloads 0 Views 3MB Size Report
(Matthew and Fung, 1987; Aselman and Crutzen,. 1989). Therefore ..... and Engelkemeir, 1988; Quay et al., 1991) were ... the period 1987-1989 from Quay et al.
Journal

of the Meteorological Society of Japan, Vol. 76, No. 2, pp. 307-324, 1998

Methane

Emissions Transport

Deduced Model

from and

a Two-Dimensional Surface

By Tazu Saeki, Takakiyo

307

Nakazawa,

Atmospheric

Measurements

Masayuki

Tanaka

Center for Atmospheric and Oceanic Studies, Graduate School of Science, Tohoku University Sendai 980-8578, Japan and IKaz Higuchi Carbon Cycle Research Laboratory, ARAM, Air Quality Branch, Atmospheric Environment Service 405 Duferin Street, Downsview, Ontario M3H 5T4, Canada (Manuscript received27 October 1997, in revisedform 6 February 1998) Abstract Latitudinal and temporal distributions of CH4 emission were estimated by an iterative inverse method using a two-dimensional atmospheric transport model and the 1983-1994 CH4 concentration data from the National Oceanic and Atmospheric Administration/Climate Monitoring and Diagnostics Laboratory Global Sampling Network. The atmospheric transport of the model was validated by simulating the concentrations of 85Kr, CFC-11, CFC-12 and CO2 observed at various locations world wide. A zonally averaged OH field for the destruction of CH4 in the atmosphere, originally derived from a three-dimensional photochemical transport model, was adjusted to simulate the observed concentration of atmospheric CH3CC13. A.calculated average latitudinal distribution of CH4 emission showed a large north-south difference, with about 75% of the total global emission residing in the northern hemisphere. The CH4 emission varied seasonally in high latitudes of the northern hemisphere, with a maximum in the summer season, while no seasonality of the CH4 release was found in the southern hemisphere. Averaged global emission of the natural and anthropogenic CL for the period 1984-1994 was estimated to be 559+9Tg/yr, with chemical loss of 528+10Tg/yr and the atmospheric increase of 31Tg/yr. In sensitivity experiments of the model results, the global emission of CH4 was found to be sensitive to the OH concentration and the atmospheric temperature but less to the atmospheric transport coefficientsand the CH4 concentration data used. The latitudinal CH4 emission distribution was dependent largely on the specification of the horizontal transport coefficients. It was also found that the 613C value of a bacterial source associated with a large amount of CH4 emission, as well as the soil absorption process of CH4 with a large kinetic isotopic fractionation, significantly impacts the determination of S13Cin atmospheric CH4.

1. Introduction Methane (CH4)is one of the most important trace gases, not only for atmospheric radiation but also for atmospheric chemistry. Since CH4 has strong absorption bands in the infrared region, it plays an important role in the radiation budget of Earth's atmosphere. The atmospheric CH4 concentration Corresponding author: Tazu Saeki, Center for Atmospheric and Oceanic Studies, Graduate School of Science, Tohoku University, Sendai 980-8578. E-mail: saeki@ caos-a. geophys. tohoku. ac. jp (c)1998, Meteorological Society of Japan

has been increasing rapidly due to human activities since the pre-industrial/pre-agricultural era; analyses of the air bubbles in polar ice cores have shown that the CH4 concentration in the atmosphere was approximately 700 ppbv in the 17th-18th centuries, and has risen since then to about 1750 ppbv at present (Etheridge et al., 1988; Chappellaz et al., 1993; Nakazawa et al., 1993a). Systematic CH4 measurements in the atmosphere have also established that the CH4 concentration has been increasing recently at a rate of 0.8-1%/yr, with a clear seasonal cycle (Steele et al., 1987, 1992; Aoki et al.,

308

Journal

of the

Meteorological

1992; Dlugokencky et al., 1994a, 1995; Sugawara et al., 1994). Such an increase of CH4 would enhance the greenhouse effect and bring about additional warming of global climate. As part of the atmospheric chemistry, CH4 reacts with hydroxyl radicals (OH), CH4+OH-CH3+H2O,

(1)

which is a major sink of atmospheric CH4. Through a series of chemical reactions starting from reaction (1), CH4 influences changes in concentrations of other atmospheric species, such as CO and 03, and is finally oxidized to CO2. In the stratosphere, CH4produces H2O by reacting with OH and reduces chlorine (Cl) atoms, which destroy 03 significantly (Cicerone and Oremland, 1988). In view of the importance of CH4 in atmospheric chemistry and climate change, it is crucial that we increase our understanding of its global cycle. Major sources of CH4 are natural wetlands, rice cultivation, fossil fuel production, ruminant animals, biomass burning, landfills, and termites (Cicerone and Oremland, 1988; Crutzen, 1991; Fung et al., 1991). These sources are not only distributed heterogeneously in a very complicated fashion on Earth's surface, but also have different magnitudes of CH4 releases; in addition, CH4 releases from sources such as rice paddies and natural wetlands vary seasonally(Matthews and Fung, 1987;Aselman and Crutzen, 1989; Fung et al., 1991). Such a situation makes it difficult to quantitatively estimate CH4 emissions from various sources. Also, it is not easy to measure temporal and spatial variations of the atmospheric OH concentration directly and systematically due to its extremely short lifetime (Spivakovskyet al., 1990;Prinn et al., 1992). To investigate the global budget of CH4, it is useful to develop CH4 cycle models. Indeed, some model studies have been done for this purpose, using two- or three-dimensional atmospheric or chemical transport models (Fung et al., 1991; Hough, 1991; Taylor et al., 1991; Brown, 1993, 1995; Law and Pyle, 1993;Hemnet al., 1997). However,global CH4 emissions estimated by these studies have ranged widely between 400 and 600 Tg/yr. Spatial and temporal distributions of CH4 emission have also not been determined satisfactorily. Therefore, further studies are needed for a better understanding of the CH4 cycle on Earth's surface. In most previous model studies, CH4 source scenarios were explicitly given. However,in this study, we have deduced latitudinal and temporal distributions of CH4 emissionsfrom the surface, using a twodimensional atmospheric transport model and CH4 concentration data from recent systematic measurements at ground stations. The results obtained have been closelyexamined in terms of changes in the atmospherictransport, the OH concentration and the

Society

of Japan

Vol. 76, No.

2

atmospheric temperature of the model, as well as of longitudinal variation of the CH4 concentration. We have also calculated 613Cin atmospheric CH4 based on the model-derivedCH4emissionsand compared them with the observational results, to obtain estimates of relative contributions from various bacterial and non-bacterial CH4 sources with different 813C-in-CH4signatures, as well as information about CH4destructive processes in the atmosphere. 2. Two-dimensional model

atmospheric

transport

In this section, we provide a brief description of the two-dimensionalatmospheric transport model, as well as validation of the transport parameterization and the OH field. 2.1 Model description The model developed in this study was a tonally averaged two-dimensional atmospheric transport model based on Higuchi (1983). The horizontal and vertical axes of the model were expressed in sine of latitude and pressure, respectively. The model consisted of 20 latitude zones with equal area from 90S to 90N (divided at 90, 64.2, 53.1, 44.4, 36.9, 30.0, 23.6, 17.5, 11.5, 5.7 in both hemispheres and 0, and 9 vertical layers from 1000to 10 hPa (divided at 1000, 950, 750, 600, 450, 350, 250, 150, 50 and 10hPa). The concentrations of atmospheric CH4 were determined by atmospheric transport and activities of its sources and sinks. The equation governing the CH4 concentration variation can be written as 8c+di at v Vc=-divF+S

(2)

where c is the mixing ratio of CH4, p is the atmospheric density, V is the advective velocity, F is the eddy flux, and S is the source and sink strength of CH4. The advective velocity is represented as a stream function 8, V=-0'1' a

a z ay

3

where y and z are horizontal and vertical coordinates, respectively. The eddy flux is accomplished through the flux-gradient relationship, F=pK K=yy K

grad c, Kyz zy Kzz

(4) (5)

where K is the eddy diffusion tensor. Subscripts of K, y and z, denote the horizontal and vertical directions, respectively. In this study, we employed the zonally averaged stream function and diffusion coefficientsobtained from the work of Plumb and Mahlman (1987) in which a mean meridional circulation and a diffusiontensor describing the eddy

April

1998

T. Saeki,

T. Nakazawa,

M.

motions were derived from the results of imaginary tracer experiments with a general circulation model at the GeophysicalFluid Dynamics Laboratory (GFDL). The transport parameters were available for each month of the year. In order to estimate CH4 emissions using Eq. (2), it is necessary to take into account the destruction of CH4 in the atmosphere. Chemical reactions involving CH4 are very complex in the atmosphere. However, for the purpose of simplicity and ease of interpretation, we included only the oxidation process of CH4 with OH, since it is the major sink of atmospheric CH4. Since OH is quite shortlived in the atmosphere, it is difficult to measure its global concentration. Therefore, the OH concentration data calculated using photochemical models have been usually used in the past. Spivakovsky et al. (1990) calculated the atmospheric OH concentration as a function of many parameters, such as temperature, radiation and the concentrations of H2O, CO, O3, CH4 and NOt (=NO+NO2+N03+ 2N205+HNO2+HNO4), using a three-dimensional photochemical transport model. In this study, we used longitudinally-averagedvalues of their calculated monthly mean OH concentration (C.M. Spivakovsky,private communication). The temperature dependent rate coefficientk for reaction (1) was taken from Vaghjiani and Ravishankara (1991) to be 1.59x10-20xT2.84 exp(-978/T) cm3/mol/s where T is the atmospheric temperature in Kelvin. The coefficientk for each model box in each month was calculated using the tonally averaged monthly mean air temperatures which were obtained by averaging 14-year (1979-1992)data from the National Meteorological Center (NMC) (K. Yamazaki, private communication). In this study, CH4emissionsat the respective latitude zones were derived by using an iterative procedure, so that the CH4concentration calculated for the lowest layer of the model agrees with the values assigned by the surface observations. At the beginning of the procedure, the model was initialized by reproducing the observed latitudinal distribution of the CH4 concentration in April 1983, and then integrated until 1994. At every 10 days during the integration, the CH4 concentrations calculated for the lowest boxes of the model were compared with the observed values. If the differencesof more than 0.1 ppbv were found between both values, CH4 emissions were adjusted to reduce them. This adjustment was further repeated until the concentration differences between the two successive time steps fell into the values less than 0.1 ppbv for all model boxes, keeping the agreement of the calculated and observed CH4 concentrations in the lowest layer of the model. It is important to note that the interannual variations in the atmospheric transport and the OH field were not incorporated into the model

Tanaka

and

K. Higuchi

309

integration. However,we do discuss the sensitivity of the model results to perturbations in these parameters in Section 4.4. 2.2 Modelvalidation In order to examine the atmospheric transport incorporated into the model, we performed several validation experiments, using Krypton 85 (85Kr),CFC11, CFC-12, and CO2. 85Kr has been released into the atmosphere by reprocessing nuclear fuel since 1945, and its lifetime is 10.76 years (Jacob et al., 1987), which is significantly longer than atmospheric mixing. The 85Kr concentration was initially set to 0 pCi/m3 for all model boxes, corresponding to the 1945 condition; the model was then integrated until 1983 using the emission inventory of Zimmermann et al. (1989). The surface 85Kr concentrations, thus calculated, were compared to those observed in the Atlantic Ocean (Weiss et al., 1983; Jacob et al., 1987). For CFCs, their concentrations were calculated until 1991, after setting them to 0 pptv in the model atmosphere to match the 1940condition. The latitudinally-dependent emissioninventoriesfor CFCs were taken from Prather et al. (1987) and Khalil and Rasmussen (1993), who derived them by assuming proportionality to industrial activities. The observational data from the Atmospheric Lifetime Experiment (Cunnold et al., 1986) were used for the comparison with the calculated concentrations. The CO2 field was employed to examine the vertical transport in the model. The model was constrained so that the calculated and observed CO2 concentrations near the ground surface agreed with each other, and then, the calculated seasonal cycles of upper tropospheric CO2 were compared with the observational results over Japan (Tanaka et al., 1987; Nakazawa et al., 1993b). The concentrations calculated by the above procedures for the four species (85Kr, CFC-11, CFC-12, and CO2) were in relatively good agreement with the observation overall, but apparent discrepancies between the calculation and the observation were also found. To reduce such discrepancies, wetried to modifyparameters governingthe atmospheric transport. For example, the seasonal cycles calculated for upper tropospheric CO2 were more attenuated and delayed in the model, compared with the observed cycles over Japan. This implied that the vertical transport in the model was slower than the real atmosphere. Notwithstanding the inconsistency in comparing the zonally averaged model results with the observation over Japan, it was also reported in previous studies that the vertical diffusion coefficients used by Plumb and Mahlman (1987), and employed in this study, were too small (Plumb and Mahlman, 1987; Plumb and McConalogue, 1988; Tans et al., 1989; Enting and Mansbridge, 1991).

310

Journal

of the

Meteorological

Therefore, vertical diffusioncoefficientswith values lower than 8m2/s were set equal to 8m2/s (P.P. Tans, private communication). From the comparison of the observed values of 85Kr, CFCs and CO2 in the equatorial region with the calculated values, the horizontal transport in mid and high latitudes of the southern hemispherewas also found to be somewhat slower than in the actual atmosphere. The horizontal transport in theeesouthern hemisphere was, therefore, enhanced to some extent as a sine of latitude. Plumb and Mahlman (1987) and Plumb and McConalogue (1988) pointed out that the tropical tropopause calculated by their transport parameters was too high. In fact, such a situation was also found to be the case in our own numerical experiments with CO2; when compared with the annual mean CO2 concentration obtained in the upper troposphere (Nakazawa et al., 1991),the model results showed higher values. To reduce the vertical transport, we decreased the vertical diffusioncoefficients at 50 and 150hPa between 18N and 18S (P.P. Tans, private communication). We also validated the OH field by using methyl chloroform (CH3CCl3), whose chemical reaction with OH in the atmosphere constitutes its major sink. CH3CCl3 has recently been released into the atmosphere from its usage as an industrial solvent, and its origin is therefore fairly well known. To compare model performance with the results from direct measurements, the CH3CCl3concentration was calculated from 1951 to 1991 after setting it initially to 0 pptv in 1951. The integration was carried out using the global inventory (Prinn et al., 1995) and relative latitudinal distribution of its release (same as that of CFCs) obtained from Prather et al. (1987) and Khalil and Rasmussen (1995). The rate coefficient for the reaction of CH3CCl3with OH was assumed to be 1.8x10-12 exp(-1550/T) cm3/mol/s, which was recently derived by the Jet Propulsion Laboratory (1997). The CH3CCl3 concentrations, thus calculated, were systematicallyhigher by about 20 ppty than the observedvalues. Therefore, we increased the OH concentration by 1.25 to reduce such a difference. The resultant globally averaged OH concentration was 10.0x105 radicals/cm3, which is in agreement with (9.7+0. 6)x105 radicals/cm3 derived by Prinn et al. (1995). The CH3CCl3concentrations calculated for Adrigole/MaceHead, Ireland (52/53N, 1,0W), Cape Meares, Oregon (45N, 124W), Ragged Point, Barbados (13N, 59W), Point Matatula, American Samoa (14S, 170E), and Cape Grim, Tasmania (40S, 144E) are shown in Fig. 1, together with the measurement results (Prinn et al., 1987, 1992, 1995). As seen from this figure, the observedconcentration has increased steadily at all sites. In general, the variations of the CH3CCl3concentrations are well reproduced by the present model, especiallyin the southern hemi-

Society

of Japan

Vol. 76, No.

2

sphere. The calculated CH3CCl3concentrations are somewhat higher at Ireland and Oregon and lowerat Barbados, compared with the observed values. The lifetime of CH3CCl3in the lower atmospherewas estimated to be 4.8 years from its loss amount calculated by the model; the lifetime is in close agreement with an averaged value of 4.6+0. 3 years derived by Prinn et al. (1995). 2.3 Verticalprofile of the CH4 concentration As described above, we derived CH4emissions by constraining the model with the observed CH4 concentration data near the ground surface. In order to examine the present model in terms of upper atmospheric CH4, the calculated vertical profiles of the tropospheric CH4 concentration were compared with those observed over Japan and Australia. Average values of the tropospheric CH4 concentration over Japan during the period 1993-1994(our unpublished data) are plotted in Fig. 2, together with those calculated for latitudes 30-37N for the same period. To account for the differencesin concentrations of the CH4 standard gases between Tohoku University and the NOAA/CMDL (Nakazawa et al., 1993a), the observed CH4 concentration was shifted down by 22.6 ppbv. The observed CH4 concentration decreased gradually with height, with a low value at 2-4km, suggestingthat there are strong CH4 sources at the ground surface in this region. The calculated CH4 concentration near the surface is lower by about 30 ppbv than that observed at 0-2km. Even if the period to be considered was extended to 7 years from 1988 to 1994, this situation was not improved. This difference may have arisen from the fact that the CMDL sites are remote from strong CH4 sources, while the Tohoku data were taken in the proximity of Japan and the Asian continent. The vertical gradients of the calculated CH4 concentration are in close agreement well with that of the observed profile. Aircraft sampling has also been done for the tropospheric CH4 concentration by the Commonwealth Scientific and Industrial Research Organization (CSIRO) above Cape Grim since 1991 (Langenfelds et al., 1993; L.P. Steele, private communication). An average vertical CH4 profile for the period 1993-1994obtained from this program is also plotted in Fig. 2. This vertical profile shows that the CH4 concentration increases with increasing height, opposite to that over Japan. The concentration differencebetween the surface and 8 km is about 10 ppbv. The fact that the destruction of CH4 with OH is enhanced in the lower troposphere (Langenfeldset al., 1993) and that the northern hemisphericair with high CH4 concentrations is transported to the southern hemispherethrough the upper troposphere (Pearman and Beardsmore, 1984; Nakazawa et al., 1991) is responsible for such a ver-

April

1998

T. Saeki,

T. Nakazawa,

M. Tanaka

and

K. Higuchi

311

Fig. 1. Comparison of the calculated CH3CC13concentrations (solid lines) with the observed values (dots and crosses) at several selected locations.

Fig. 2. Average values of the tropospheric CH4 concentrations observed between 0 and 10km over Japan (solid line with open circles) and Australia (dotted line with closed circles) during the period 1993-1994,and those calculated for the corresponding latitude bands for the same period (dashed and dotted lines).

tical CH4 profile. On the other hand, the calculated values of the tropospheric CH4concentration for latitudes 37-44 S decrease slightly from the surface to the middle troposphere and then increase with height. The calculated CH4 concentration near the ground surface is almost identical to the observed values, but the difference between the calculated and observed CH4 concentrations in the upper troposphere amounts to about 10ppbv. The cause may be partly ascribed to a regional effect in which the CH4 concentrations in the upper and middle troposphere over Australia are strongly affected by the transport of the northern hemispheric air with high CH4 concentrations especially in the summer season, due to the monsoon circulation in Southeast Asia (Pearman and Beardsmore, 1984; Nakazawa et al., 1991). 3. Data fitting To derive CH4 emissions in the present study, we needed to prepare CH4concentration data with reg-

312

Journal

of the

Meteorological

Table 1. Summary of the NOAA/CMDL

Society

of Japan

Vol. 76, No. 2

sites from which the data were used in this study

ular intervals in time and in the north-south spatial direction. In this study, the concentration data were taken from 25 land site measurements and 1 shipboard measurement of the National Oceanic and Atmospheric Administration/Climate Monitoring and Diagnostics Laboratory (NOAA/CMDL) Global Sampling Network, from 1983to 1994,covering latitudes from 90S to 82N (Steele et al., 1987, 1992;Dlugokencky et al., 1994a, 1995). Information about the sites are summarized in Table 1. The data at high-elevation sites were not included, except for those from the South Pole at 2810m. The CH4concentration at the South Pole is known to be almost the same as those of other sites in mid and high latitudes of the southern hemisphere(Steele et al., 1987; Dlugokencky et al., 1994a). In order to interpolate each data set to an equal interval of time, as well as to separate signals from noises, a digital-filtering technique including Fourier harmonics, linear interpolation, Reinsch-type spline and Butterworth filter (Nakazawa et al., 1991, 1993b, 1997a, b) was applied. A brief outline of the technique is given below. We first estimated an approximate seasonal cycle and an approximate long-term trend by using a Fourier function and the Reinsch-type cubic spline with a cutoff period of 5 years. In this study, the number of Fourier harmonics was set equal to 3.

The fits with the Reinsch-type spline and the Fourier function were alternately repeated until the approximate long-term trend was unchanged. The observed data lying outside more than 3 standard errors of fit away from the fitted curve were regarded as outliers and rejected from further analysis. The approximate long-term trend and the approximate seasonal cycle were removed from the refined data set with no outliers. To derive interannual variations of the long-term trend, the daily concentration values, obtained by linearly interpolating the refined data set, were smoothed by the 26th-order Butterworth filter with a cutoff period of 21 months. Then, the average seasonal cycle was derived by fitting the Fourier function to the refined data after subtracting the long-term trend obtained above. These procedures with the Fourier function and the Butterworth filter were repeated until interannual variations of the long-term trend became almost unchangeable. Finally, intraannual variations of the seasonal cycle were determined by applying the 16th-order Butterworth filter with a cutoff period of 4 months to the daily concentration data after subtracting the average seasonal cycle and the long-term trend from the refined data. The best fit curve of the observed data was represented by a sum of the approximate longterm trend, its interannual variations, the average seasonal cycle and its intraannual variations.

April

1998

T. Saeki,

T. Nakazawa,

M. Tanaka

and

K. Higuchi

313

Fig. 3. Observed CH4 concentrations at five selected sites of the NOAA/CMDL air sampling network and their best fit curves (solid lines). Dots and crosses represent the baseline data and outliers, respectively. Dotted lines denotes the long-term CH4 trend.

As an example, at Point Barrow, Cape Grim, and 3, together with trends obtained From this figure, space variations

the CH4 concentrations measured Cape Kumukahi, Christmas Island, the South Pole are shown in Fig. their best fit curves and long-term by the technique described above. some general features of time and of the CH4 concentration on the

globe can be seen; the CH4 concentration increases secularly at all sites, accompanied by the seasonal cycle, and is higher in the northern hemisphere than in the southern hemisphere. For use in constraining the model, the CH4 concentration values were first taken every 10 days from the fitted curve to the data observed at each site for the period from 1983 to 1994. The values at all sites for each day were then fitted by using the Reinsch-type spline to obtain its latitudinal distribution. Then, the CH4 concentrations in latitudes corresponding to the surface boxes of the model were read from these fitted curves.

4. Results and discussion The results from the model are discussed in terms of the latitudinal distribution and seasonal variation of CH4 emission, and the secular change of global CH4 budget. We also discussthe results of sensitivity tests for our model. 4.1 Latitudinal distribution of CH4 emission Figure 4 shows the average latitudinal distribution of annual CH4 emission calculated for the period 1984 to 1994. The result for 1983 was not included in this average, because the CH4 concentration data were available from only few sites before March of this year. It is obvious from this figure that the CH4 emission was strong in mid and high latitudes of the northern hemisphere and weak in the southern hemisphere. In high latitudes of the southern hemisphere, CH4 emission of 6-7Tg/yr was calculated for each latitude zone. In Fig. 4, the total CH4 emission in the latitudes 30-90S amounted to about 7 % of

314

Journal

of the

Meteorological

Society

of Japan

Vol. 76, No. 2

Fig. 4. Latitudinal distribution of the average annual CH4 emission derived from the present model and the NOAA/CMDL surface observational data for the period 1984-1994.

the total global emission. The possible CH4 source in this region is thought to be the ocean, but the amount of CH4 released from the global ocean was estimated by Cicerone and Oremland (1988) to be about 10Tg/yr. Therefore, CH4 emission derived for high latitudes of the southern hemisphere appears to be too large, and the cause may be attributed to a larger-than-real CH4 chemical loss and/or smaller southward atmospheric transport in the model. As far as the atmospheric transport is concerned, it was suggested by Murayama et al. (1995) that the northern hemispheric air is transported to the Antarctic region through the upperr troposphere from the late boreal fall to winter, and the low latitude air of the southern hemisphere through the lower troposphere during other seasons. These atmospheric transport processes might not have been represented well in the model, in spite of some modifications of its transport parameters to match the observed distributions of the trace gases during the model validation exercise. In the equatorial region, CH4 emission is much more enhanced than in the southern hemisphere. The CH4 emission released from 18N to 30S is 176+9Tg/yr. Major CH4 sources in this region are thought to be wetlands (peat-poor swamps), biomass burning, and termites (Fung et al., 1991). According to Matthews and Fung (1987), possible annual CH4 emission from global wetlands is approximately 110Tg/yr, and about 25% of it comes from peat-poor swamps in latitudes 20N to 30S. A global CH4 emission from biomass burning was estimated to be 30Tg/yr by Hao and Ward (1993) and 11-53Tg/yr by Crutzen and Andreae (1990). About 80% of it is thought to be released in the

tropics (Hao and Ward, 1993). Emission of CH4 from termites is quite uncertain; the estimates given by Crutzen et al. (1986) and Fung et al. (1991) were 2-6 and 15-100Tg/yr, respectively. These sources contribute to the CH4 emissions derived by our model for the tropical region. To quantify individual contributions, isotopic composition of atmospheric methane will give us some information, as will be described later. The total calculated amount of CH4 released in the northern hemisphere was 420+7Tg/yr, about 75% of its global emission. Figure 4 shows that CH4 emission was especially enhanced in mid and high latitudes, with a maximum at 44-55N. Anthropogenic CH4 is thought to be released into the atmosphere in these latitudes, due to natural gas venting, pipeline leakage of natural gas, coal mining, landfills, rice cultivation, cattle breeding and biomass burning. The total emission from these anthropogenic sources was estimated to be 40-75% of the global emission(Khalil and Rasmussen, 1990; Fung et al., 1991). Rice paddies are one of the important anthropogenic CH4 sources due to its large amount of CH4 emission. Its estimates range from 60 to 170Tg/yr (Cicerone and Oremland, 1988; Aselman and Crutzen, 1989;Fung et al., 1991). Major rice-producingareas are located around 25N in such countries as India and China (Aselman and Crutzen, 1991). Therefore, CH4 from rice paddies could be responsible for high emissions in mid latitudes. In addition to these anthropogenic sources, natural wetlands are also one of the most important CH4 sources in this region. Estimated emission is around 100Tg/yr (Cicerone and Oremland, 1988; Aselmann and Crutzen, 1989; Fung et al., 1991).

April

1998

T. Saeki,

T. Nakazawa,

M. Tanaka

and

K. Higuchi

315

Fig. 5. Calculated seasonal variation of the latitudinal distribution of CH4 emissions (Tg/month/latitude band of the model).

It was also reported by Matthews and Fung (1987) that about 60% of the global emission is released from wetlands in latitudes 50-70N. CH4 originating from various anthropogenic activities, as well as from natural wetlands, plays an important role in the large emissions calculated for high latitudes of the northern hemisphere. 4.2 Seasonal variation of CH4 emission Figure 5 shows the calculated seasonal variation of CH4 emission averaged over the model simulated years from 1984to 1994. Seasonality in the sources and sinks of CH4, as well as in the atmospheric transport, is the major contributor to the latitudinal distribution of the seasonal cycle in Fig. 5. Major CH4 sourceswith seasonally-differentmagnitude are natural wetlands, rice cultivation, and biomass burning. The seasonal variation of CH4 emission is found to be very small in the southern hemisphere where CH4 sources are very weak. Thus, the seasonal cycle of the atmospheric CH4 concentration observed in the southern hemisphere is thought to be produced mainly by the seasonally varying atmospheric transport and chemical destruction of CH4. In low latitudes, the seasonal variation of CH4 emission reaches a maximum in December or January, with a value of 3-4Tg/month/latitude band of the model. Since the CH4 source with seasonally varying magnitude in this region is biomass burning, with a maximum emission in the dry season (Fung et al., 1991), the seasonal variation of CH4 emissiondeduced in the tropics would be affected by this source. Another possible cause could be atmospheric transport. Near the equator, the Intertropical ConvergenceZone (ITCZ) and the Southern Pa-

cific Convergence Zone (SPCZ) exist, and the interhemispheric air exchange is largely influenced by their locations and activities. During the austral summer, the ITCZ is located south of Christmas Island and Seychelles,allowing the advection of the northern hemispheric air with high CH4 concentrations to these sites. Depending on the locations of the ITCZ and the SPCZ, Samoa can also be affected significantly by the northern hemispheric air during the austral summer, and by the southern hemispheric air in other seasons. Since such a regional feature of the atmospheric transport cannot be adequately reproduced by the present zonally averaged global model, the rapid increase of the CH4 concentration due to the northern hemispheric air leads to a large amount of CH4 emission calculated in the model tropics. Rice is produced in latitudes 10-40N, and its cultivation is a seasonal CH4 source with a maximum emission in the summer season (Aselman and Crutzen, 1989). However,the model results do not showa clear evidencefor such a pronounced summer maximum in these latitudes. One possible reason is ascribed to the fact that the CMDL sites are remote from rice paddies. It was pointed out, for example, by Fung et al. (1991) in their three-dimensional model study that the seasonality of CH4 emission from rice cultivation was not captured faithfully by the CMDL sites in the subtropical region. Figure 5 also shows a very clear CH4 emission maximum in mid and high latitudes of the northern hemisphere during the late summer. One of the most important seasonal CH4 sources in this region is the natural wetlands. It is well known that much of the annual release of CH4 from the wetlands in

316

Journal

of the

Meteorological

Fig. 6. Secular trends of (a) global mean annual CH4 emission and (b) four semi-hemispheric annual CH4 emissions during the period 1984-1994. Dotted lines represent the linear fits to the data.

high latitudes occurs mainly in the summer season (Matthew and Fung, 1987; Aselman and Crutzen, 1989). Therefore, the model results shown in Fig. 5 are consistent with the seasonality of CH4 emission expected from the wetlands. Emissions of CH4 during the colder seasons of the year are relatively uniform and might originate from anthropogenic activities. 4.3 Global CH4 emission and its secular trend Average annual global emission and chemicalloss of CH4 over the period 1984-1994 were calculated to be 559+9 and 528+10Tg/yr, respectively,with a lifetime of 9.1 years (calculated from amount of chemicalloss in the model atmosphere). The difference of 31 Tg/yr between the source and the sink corresponds to the observed average atmospheric CH4 increase rate during this period. Recent annual global emissions of CH4 were also estimated, for example, to be 505Tg/yr by Crutzen (1991), 500Tg/yr by Fung et al. (1991), 640Tg/yr by Hough (1991), 575Tg/yr by Law and Pyle (1993) and 420-620Tg/yr by Khalil and Rasmussen (1990).

Society

of Japan

Vol. 76, No. 2

The emission rate of 559Tg/yr obtained in this study is within these estimates. However, since soil sink was not considered in our model, its contribution is reflected in the CH4 emission. Taking this into account, the CH4emission of the present study may be increased to some extent. In this regard, the estimates of the annual uptake by the soil range between 6 and 58Tg/yr (Aselmann and Crutzen, 1989; Born et al., 1990; Crutzen, 1991; Fung et al., 1991). Figure 6 shows the calculated global secular trend of the annual mean CH4 emission during the period from 1984 to 1994. Also shown in Fig. 6 are the corresponding secular trends in 4 different latitude zones, two in the northern hemisphere and two in the southern hemisphere, i,e., northern and southern lower (0-30) and higher (30-90) latitudes. As seen in this figure, the global CH4 emissions were almost constant over the period or decreased only slightly with time. The atmospheric CH4concentration data obtained at the CMDLsites showed a gradual slowingdown in its upward trend, especially after the mid-1980s (Steele et al., 1992, Dlugokenckyet al., 1994b). Since we assumed in our study that the seasonal cycles of the OH field, the atmospheric temperature field and the atmospheric transport field were always the same from year to year, the observed deceleration of the atmospheric CH4 increase resulted in slight decrease in the calculated CH4 emission. In this respect, the rapid nature of the deceleration of the CH4 increase in the atmosphere led Steele et al. (1992) and Dlugokencky et al. (1994b)to speculate that the deceleration might be caused by a reduction in the CH4 emissions from changeable anthropogenic sources such as fossil-fuel exploitation. In fact, our results showed that the decrease in CH4 emission is most pronounced in the region from 30 to 90N, which is highly populated and industrialized. By contrast, CH4 emissions from the low latitudes of the northern hemisphere increased gradually with time, due to such anthropogenic sources as rice cultivation. South of the equator, CH4 is released into the atmosphere at a relatively constant rate throughout the hemisphere. Figure 6 also shows that there was a relatively rapid decrease in the global CH4 emission, mainly from 1991 to 1992. Inspection of the results for 4 latitude zones indicates that this decrease was caused by a very pronounced decrease in the high latitude zone of the northern hemisphere. This deceleration in the CH4 increase suggests a decrease in the emission by about 50Tg at latitudes north of 30N. Taking account of the fact that Mt. Pinatubo, in the Philippines erupted in June 1991, the cause of the low CH4 growth rate in 1992 may be ascribed to changes in CH4 released from natural wetlands due to global climate change or changes in

April

1998

T. Saeki,

T. Nakazawa,

M. Tanaka

and

K. Higuchi

317

CH4 sink amount due to volcanic aerosols added to the stratosphere (Dlugokencky et al., 1994b; Tans et al., 1996). The reduction in CH4 emissions from biomass burning in the southern hemisphereand fossil fuel in the northern hemisphere, as well as the enhancement of the chemical CH4 destruction in association with ozone depletion in the stratosphere, may also be responsible for this phenomenon (Bekki et al., 1994; Dlugokencky et al., 1994b; Lowe et al., 1994). 4.4 Sensitivity tests of the model results We examined the sensitivity of the CH4 emission and its latitudinal distribution derived in Subsection 4.1 (used as the reference case) to changes in the atmospheric transport coefficients,the OH field, the atmospheric temperature field and the concentration data used. Figure 7a shows the latitudinal distributions of CH4 which were obtained by changing the diffusion coefficientsKyy, by +50%. Global emissions derived by increasing and decreasing Kyy were both 559 Tg/yr, which are the same as that for the reference case, but their latitudinal distributions were quite different. In the case of 50% increase in Kyy, CH4 emissionincreased and decreased in the northern and the southern hemispheres, respectively, increasing the differencein the two hemispheric CH4 emissionsby 10Tg/yr. This is due to the fact that the air exchange between the two hemispheres was enhanced by increasing Kyy. The negative emissions shownin Fig. 7a for the case of 50% increase in Kyy in mid latitudes of the southern hemispherecould be attributable therefore to the rapid southward transport. In contrast, the differencein the northern and southern hemispheric CH4 emissions was decreased by 9Tg/yr by decreasing Kyy overall by 50%, owing to slowerair exchangebetween the northern and the southern hemispheres. The CH4 emissions derived by changing Kzz by +50% are shown in Fig. 7b. Respective global CH4 emissionsobtained by increasing and decreasingKzz were 559 and 557Tg/yr. The latitudinal distribution of CH4 emission for the case of increased Kzz is almost the same as that in the referencecase, except for the mid and high latitudes of the northern hemisphere, where CH4 emissions became slightly smaller. In the case of decreased Kzz, CH4 emissions were larger in mid and high latitudes of both hemispheres but smaller in low latitudes, as compared to the reference case. However, the amount of change in CH4 emission distribution calculated for the both cases was not as large as was the case with changes in Kyy. We also found that changes in Kyz and Kzy have little effect on the results, implying that the transport by these components is extremely small. As mentioned before, it is difficult to measure the

Fig. 7. Latitudinal distributions of CH4 emissions obtained by changing the transport coefficientsof (a) Kyy and (b) Kzz by +50%, and (c) the OH concentration by +10%. Solid lines denotes the results for the reference case (cf. text), and dashed and dotted lines represent those calculated by increasing and decreasing the coefficientsor the OH concentration, respectively.

OH concentration directly and globally. The Jet Propulsion Laboratory (1990) also derived the rate coefficientfor the reaction of CH3CC13with OH to be 5.0x10-12 exp(-1800/T) cm3/mol/s, which is different from that employed in this study. This rate coefficientis larger than the coefficientused in this study by about 20% at 300K. To examine uncertainties of the model results arising from the OH concentration, we calculated the CH4 emissions by increasing and decreasing their concentration by 10%. The results obtained are given in Fig. 7c. The CH4 emissions derived for both cases showed latitudinal distributions which are similar to that for the referencecase. The annual global CH4 emissionwas found to be changed by about 50Tg/yr by the 10% change in the OH concentration; the emission was increased and decreased by increasing and decreasing OH, respectively,to compensate for the chemical loss of CH4 with OH. The change in CH4 emissions in low latitudes is somewhat larger than those in other regions, due to the fact that the atmosphere is rich in OH and air temperature is high. In this study, the latitudinal distribution of CH4 emission was calculated with the assumption of time

318

Journal

of the

Meteorological

Society

of Japan

Vol. 76, No. 2

Fig. 8. Latitudinal distributions of the average annual CH4 emission derived from five different data sets of the CH4 concentration constructed arbitrarily on the basis of systematic measurements at the NOAA/CMDL sites (cf. text). The distribution given in Fig. 4 is also shown by "reference case". The error bars represent the standard deviation at each latitude for the reference case.

invariant seasonal cycles of the. OH concentration field. It was pointed out by Prinn et al. (1995), on the basis of their CH3CC13measurements, that the atmospheric OH concentration was almost constant from 1978 to 1990, the rate of change in OH being 0.0+0.2%/yr. On the other hand, Madronich and Granier (1992) suggested that the recent slowing down of the CH4 increase is caused by an increase in the tropospheric OH due to the depletion of stratospheric O3. Therefore, in order to estimate the degree of impact that changing OH field could have on our calculation of the CH4 emission, we calculated the annual mean CH4 emissions by increasing and decreasing the atmospheric OH concentration at a rate of 1%/yr. The results showed that the global CH4 emission increased gradually at a rate of 4.9Tg/yr2 during the period covered by this study, with an average annual emission of 590+19Tg/yr, and decreased at -5.8Tg/yr2, with an average emission of 496+8Tg/yr, by increasing and decreasing the OH concentration, respectively. It was also found that the CH4 emissions for higher and lowerlatitude bands in the northern hemisphere were almost invariable with time by increasing and decreasing OH at a rate of 1%/yr, respectively. From the rate coefficientfor the reaction of CH4 with OH employedin this study, we found a change of 1% (about 2.6K) in the atmospheric temperature yields approximately 6.7% change in the reaction rate, corresponding to 6.7% change in OH. Therefore, the global CH4emissionis expected to be changed by about 30Tg/yr by the 1% temperature change, on the basis of the above results for changes

in OH. Indeed, the model results obtained by changing the temperature by +1% showedthat changesin magnitude of the annual global CH4 emission were about +27Tg/yr. The latitudinal distributions of CH4 emissionobtained for both cases are not so different from that for the referencecase. In order to examine the effect of the longitudinally-different CH4 concentrations on the modelderived CH4 emissions, we arbitrarily constructed five different concentration data sets (D1, D2, D3, D4 and D5) from the NOAA/CMDL data available for this study. D1 consists of the data from the ocean sites including Mould Bay, Ocean station "M", Cold Bay, Olympic Peninsula, Cape Meares, Terceira Island, Sand Island, Cape Kumukahi, Guam, Christmas Island, and shipboard, and D2 from the sites near or on the continents including Alert, St. Croix, Point Barrow, Cold Bay, Cape Meares, Ket Biscayne, Mould Bay, Ragged Point and Shemya Island, in addition to those from all sites in the southern hemisphere. For D3, the data from two continental sites, Tae-ahn Peninsula, Korea (36N, 126E) and Qinghai Province, China (36N, 100E), were added to the reference case which is comprised of those from the sites shown in Table 1. Annual mean CH4 concentrations at Qinghai Province are almost the same as those at maritime sites in similar latitudes, but those at Tae-ahn Peninsula is higher by about 70-80 ppbv. D4 and D5 were obtained artificially by adding and subtracting, respectively, one standard deviation of curve fit to and from the CH4 concentration data for each site. The CH4 emissions derived from these five data sets are plot-

April

1998

T. Saeki,

T. Nakazawa,

M. Tanaka

ted in Fig. 8 against latitude, together with those for the reference case. The latitudinal distributions of CH4 emission derived are consistent with each other, which are very similar to that for the reference case. The global CH4 emissions are 558, 558, 559, 561 and 556Tg/yr for D1, D2, D3, D4 and D5, respectively,the differencesamong the values being extremely small. These values are almost identical to the emission of 559Tg/yr for the reference case. However,more detailed inspection of the latitudinal distributions indicates that CH4 emissions are somewhat different, especially in northern middle latitudes from 30N to 50N where strong CH4 sources exist. Longitudinal variation of the CH4 concentration was also pointed out based on observations (e.g., Sugawara et al., 1996;Nakazawa et al., 1997c) and model studies (Fung et al., 1991; Hemn et al., 1997). For example, Sugawara et al. (1996) and Nakazawa et al. (1997c) found from their aircraft measurements over Russia that the lower tropospheric CH4 concentration over Siberia is considerably high even in the summer season, due to the release of a substantial amount of CH4 from natural wetlands and the leakage of natural gas. To reduce uncertainties of the global CH4 budget, it is necessary to systematically measure the CH4 concentration in the continents. 5. Latitudinal distribution spheric CH4

of o13C in atmo-

In order to examine the global budget of CH4 in terms of its carbon isotopic composition, we calculated the 813C value in atmospheric CH4 on the basis of the CH4 emissions derived above. The model was initialized by integrating it until a steady state condition in which -47.28 and -46.96%o for the atmospheric S13C in the northern and southern hemispheres,respectively (Stevens and Engelkemeir, 1988; Quay et al., 1991) were achieved. Then, the CH4 and 13CH4concentrations were calculated for the period 1983-1994, using the monthly mean CH4emissionsobtained in this study, and 813Cwas derivedfrom these two calculated concentrations, using the equation

and

K. Higuchi

319

al., (1991) on the basis of their observational data of S13Cand 14Ccontent in atmospheric CH4. The CH4 emission for each latitude band in each month was assignedto the respective sources, using the relative latitudinal distributions of CH4 sources estimated by Fung et al. (1991). The S13Cvalues of CH4 emitted from bacterial source, fossil fuel and biomass burning were assumed to be -60(+5), -41(+8) and -27 (+3)% respectively (Quay et al., 1991). The kinetic fractionation factor for the reaction of CH4 with OH was taken to be k13/k12=0.9946 (Cantrell et al., 1990). The latitudinal distributions of annual mean CH4 emissions from the respective sources, averaged over the period 19841994, are shown in Fig. 9. Also shown are the 613C values of CH4 emitted from the respective latitude bands. The globally averaged S13Cof CH4 sources is -53.2%o with relatively high values in the equatorial region due to emissions of isotopically heavy CH4 from biomass burning. The observational data of atmospheric 613C for the period 1987-1989 from Quay et al. (1991) and Lowe et al. (1991) are plotted in Fig. 10, to compare with the calculated values. The observed values were taken from systematic measurements at Cape Grim, Point Barrow, Mauna Loa (19N, 155W), Olympic Peninsula, Baring Head (41S, 175E). The observed values of S13Crange from -47.6%o in the north polar region to -47.0%o in the south polar region, the global mean being -47.3+0.2%o The 813Cvalue is almost constant in the southern hemisphere, and most of its north-south gradient occurs in the northern hemisphere. Measurements of the atmospheric S13C were also made over Siberia in the summer of 1993 and 1994, and the values of -47.9+0.3 and -47.8+0.2%o were found in the free atmosphere for the respective years (Sugawara et al., 1996). These values are slightly lower than those by Quay et al. (1991) and Lowe et al. (1991), due to isotopicallylight CH4 releasedfrom wetlands. The S13Cdata from Siberia in 1993-1994 are also plotted in Fig. 10 for reference. The latitudinal distribution of the calculated 613C for the lowest model level, averaged over 1987 to 1989, is shown by R1 in Fig. 10. The globally averaged atmospheric 613Cis -47.78%o and the difference between the northern and southern hemisphere is -0.40%o. 613Cis almost constant in the southern hemisphere and it decreases mostly in the northern hemisphere. This latitudinal distribution is very similar to the observed distribution. The interhemispheric differencein 8130 are due to the fact that a substantial amount of isotopicallylight CH4is released from bacterial sources into the atmosphere especiallyin northern middle and high latitudes and that the released CH4becomes enriched in 13Cduring the transport from the northern hemisphere to the southern hemisphere, due to the kinetic frac-

S13C-{[13cH4/(CFIq-13CFI4)]model (13c/12 -PDB

x1000(%0),

(6)

where (13C/12C)pDB is 0.011237 (Craig, 1957). Since the knowledge of the seasonal variations of CH4 emissions from individual sources and their 8130 valuesis currently insufficient,we roughly classified the global CH4 emission for each month into three sources composed of bacteria, fossil fuel and biomass burning, and we assumed their relative strengths to be 73, 16 (+12) and 11 (+4)%, respectively; these values were determined by Fung et

320

Journal

of the

Meteorological

Society

of Japan

Vol. 76, No. 2

Fig. 9. Latitudinal distributions of annual CH4 emissions for bacterial source, fossil fuel and biomass burning (histogram bars) and S13C calculated from these source scenario (solid line), averaged over the period 1984-1994.

Fig. 10. Calculated values of 813C in atmospheric CH4, averaged over the period 1987-1989 (cf. text), and the data observed by Quay et al. (1991) (open circles), Lowe et al. (1991) (closed triangles). The observed data over Siberia in 1993 and 1994 are also plotted for reference (closed squares).

tionation effect in the reaction with OH. A similar latitudinal distribution of S13C was also obtained by averaging its monthly values during the period 1984-1994, as well as by averaging its annual values derived by using annual mean CH4 emissions for the same period, instead of the monthly means. It is, however, obvious from Fig. 10 that the calculated 613C values are smaller by about 0.5%o than the observed values as a whole. To account for this discrepancy, it may be possible that the 813C values of CH4 sources assumed in this calculation are too low. Therefore, we calculated the atmospheric o13C again, by changing 513C of the respective CH4

sources within uncertainties estimated by Quay et al. (1991), i. e., +5%o for bacterial sources, +8%o for fossil fuel and +3%o for biomass burning. The results obtained by setting 613Cto -58%o for bacterial source or -33%o for fossil fuel, represented respectively by R2 and R3 in Fig. 10, fit well with the observation. The recent global 6130 mean value of about -47.3%o was also obtained by changing the relative magnitude of biomass burning and bacterial source to 16 and 68%, respectively, but the differencebetween the northern and southern hemispheric values was -0.50%o, which is larger than the observedvalue.

April

1998

T. Saeki,

T. Nakazawa,

The low calculated values of 8130 is also attributable to the CH4 loss by factors other than the reaction with OH. To confirm this effect, we recalculated the atmospheric 813C by including soil absorption of 10 and 30Tg/yr in the total CH4 emission. These absorption amounts are within previous estimates (Aselmann and Crutzen, 1989; Crutzen, 1991;Fung et al., 1991) and allocated to the respective latitude bands accorddingto Fung et al. (1991). The kinetic fractionation factor for soil absorption was assumed to be k13/k12=0.978 (Tyler et al., 1994). The respective results for soil absorption of 10 and 30Tg/yr are shown by S1 and 52 in Fig. 10. Compared with those for the source scenario without soil absorption (R1), the 8130 values are obviously increased by including soil absorption of 30Tg/yr (S2), in good agreement with the observed values. Since the global mean o13Cis -47.17%o the inclusion of soil absorption of 30Tg/yr yields the carbon isotopic enrichment of 0.57%o. The calculated o'3C values for the source scenario with soil absorption of 10Tg/yr (S1) are in between R1 and 52, suggesting that this amount of soil absorption is insufficientto obtain an agreement between the calculated and observed 613Cvalues only by this effect. It is obviousfrom this examination that the kinetic fractionation effect of soil absorption is effective for the atmospheric 613C,though the CH4 loss by itself is relatively small. This is also pointed out by Gupta et al. (1996). Recently, Gupta et al. (1996) suggested that the chemical reaction of CH4 with Cl radicals is especially important for the o13C value in the stratosphere and the marine boundary layer, because of its large kinetic fractionation. If this is the case, the atmospheric S13Cwould be increased by taking the reaction with Cl into consideration. However,spatial and temporal distributions of the concentration of Cl in the atmosphere are not well known. Also, the kinetic fractionation factor for the reaction of CH4 with Cl is quite uncertain; the respective values of 0.939 and 0.975 were derived by Saueressig et al. (1995) and Tanaka et al. (1996) for k13/k12 at 300K, the difference between both values being significantlylarge. To include this reaction effect in 613Cin atmospheric CH4, further studies are needed for knowledgeabout these facets. 6. Conclusions The global budget of CH4 was examined by using a two-dimensionalatmospheric transport model and the surface data from the NOAA/CMDL network for the period 1983-1994. The model transport provided by Plumb and Mahlman (1987) was validated by simulating 85Kr, CFC-11, CFC-12, and CO2. CH4 destruction by OH was introduced into the model in a simple way, using the longitudinallyaveraged OH concentrations which were originally

M. Tanaka

and

K.

Higuchi

321

derived by Spivakovskyet al. (1990) and adjusted in this study by simulating the atmospheric CH3CC13 concentrations. The CH4 concentrations collected at the NOAA/CMDL stations were analyzed using the digital filter technique and the data set was constructed to constrain the model's lower boundary. The comparison of the calculated vertical CH4 profiles with the observed values also indicated that it is important to model the atmospheric transport well, especially reproducing interhemispheric exchange across the equator and vertical mixing. About 75% of average global CH4 emission was found to be released from the northern hemisphere, especially in mid and high latitudes where highly populated and industrialized regions are located. It is thought that human activities and natural wetlands are responsible for such a large amount of CH4 release in these latitudes. The CH4 emission was found to be highly seasonal in high and mid latitudes of the northern hemisphere, with a maximum in the summer season, due mainly to CH4 emissions from natural wetlands. Such a seasonally-dependentCH4 emissionwas not found in the southern hemisphere. The average global annual CH4 emission for the period 1984-1994 was estimated to be 559±9 Tg/yr, with a chemical loss of 528+10Tg/yr. It was also found that the annual global CH4 emissions were almost constant during the period 1984-1994, and that the CH4 emissions decreased and increased secularly in northern higher and lowerlatitudes, respectively, while they were independent to time in the southern hemisphere. We also investigated the sensitivity of the model results to perturbations in the factors (such as the atmospheric diffusion coefficients,the OH concentration, the atmospheric temperature and the concentration data) which were fixed or prescribed in this study. It was found that the transport coefficients Kyy and the atmospheric temperatures are important for determining the latitudinal distribution of CH4 emission and the global CH4 emission itself, respectively. By changing the OH concentration, the calculated CH4 emission was significantly affected, but its latitudinal distribution was almost insensitive. The analyses of the CH4 data sets constructed arbitrarily on the basis of measurements at the NOAA/CMDL sites showed that the CH4 emissions and the latitudinal distributions derived were consistent with each other. However, it was suggested that the CH4 emissionis different longitudinally, especially in northern middle and high latitudes. Estimates of the latitudinal distribution of the relative contributions by different sources in each latitude zone to the zonally averagedCH4emissionwere obtained by calculating 813Cin atmospheric CH4 based on the global CH4 emissions derived in this study, as well as on the knowledgeof the latitudinal

322

Journal

of the

Meteorological

distribution and o13C values of CH4 sources from previous studies. It was found that the atmospheric o13C is significantly affected not only by the o13C value of bacterial source due to a substantial amount of CH4 released from it, but also by the CH4 loss bb soil absorption due to a large kinetic fractionation effect during its process. Acknowledgments We are grateful to Dr. P.P. Tans, NOAA/CMDL, Dr. C.M. Spivakovsky, Harvard University, Dr. K. Yamazaki, Hokkaido University and Dr. L.P. Steele, CSIRO for providing us with the respective data of the atmospheric transport parameter, the atmospheric OH concentration, the atmospheric temperature and CH4 concentration over Australia, and for their helpful comments. The CH4 concentration data used in this study were obtained from the NOAA/CMDL anonymous ftp site. References Aoki, S., T. Nakazawa, S. Murayama and S. Kawaguchi, 1992: Measurements of atmospheric methane at Japanese Antarctic Station, Syowa. Tellus, 44B, 273-281. Aselmann, I. and P. J. Crutzen, 1989: Global distribution of natural freshwater wetlands and rice paddies, their net primary productivity, seasonality and possible methane emissions. J. Atrnos. Chem., 8, 307358. Bekki, S., K. S. Law and J. A. Pyle, 1994: Effect of ozone depletion on atmospheric CH4 and CO concentrations. Nature, 371, 595-597. Born, M., H. Dorr and I. Levin, 1990: Methane consumption in aerated soils of the temperate zone. Tellus, 42B, 2-8. Brown, M., 1993: Deduction of emissions of source gases using an objective inversion algorithm and a chemical transport model. J. Geophys. Res., 98, 12639-12660. Brown, M., 1995: The singular value decomposition method applied to the deduction of the emissions and the isotopic composition of atmospheric methane. J. Geophys. Res., 100, 11425-11446. Cantrell, C. A., RE. Shetter, A. H. McDaniel, J. G. Calvert, J. A. Davidson, D. C. Lowe, S.C. Tyler, R. J. Cicerone and J. P. Greenberg, 1990: Carbon kinetic isotope effect in the oxidation of methane by the hydroxyl radical. J. Geophys. Res., 95, 22455-22462. Chappellaz, J., T. Blunier, D. Raynaud, J. M. Barnola, J. Schwander and B. Stauffer, 1993: Synchronous changes in atmospheric CH4 and Greenland climate between 40 and 8 kyr BP. Nature, 366, 443-445. Cicerone, R. J. and R. S. Oremland, 1988: Biogeochemical aspects of atmospheric methane. Global Biogeochem. Cycles, 2, 299-327. Craig, H., 1957: Isotope standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide. Geochim. Cosmochim. Acta, 12, 133-149.

Society

of Japan

Vol.

76, No. 2

Crutzen, P.J., I. Aselmann and W. Seiler, 1986: Methane production by domestic animals, wild ruminants, other herbivorous fauna, and humans. Tellus, 38B, 271-284. Crutzen, P.J. and M.O. Andreae, 1990: Biomass burning in the tropics: Impact on atmospheric chemistry and biogeochemicalcycles. Science, 250, 1669-1678. Crutzen, P.J., 1991: Methane's sinks and source. Nature, 350, 380-381. Cunnold, D.M., R.G. Prinn, R.A. Rasmussen, P.G. Simmonds, F.N. Alyea, C.A. Cardelino, A.J. Crawford, P.J. Fraser and R.D. Rosen, 1986: Atmospheric lifetime and annual release estimates for CFC13 and CF2C12from 5 years of ALE data. J. Geophys. Res., 91, 10797-10817. Dlugokencky, E. J., L. P. Steele, P.M. Lang and K.A. Masarie, 1994a: The growth rate and distribution of atmospheric methane. J. Geophys. Res., 99, 1702117043. Dlugokencky, E. J., K.A. Masarre, P.M. Lang, P.P. Tans, L. P. Steele and E. G. Nisbet, 1994b: A dramatic decrease in the growth rate of atmospheric methane in the northern hemisphere during 1992. Geophys.Res. Lett., 21, 45-48. Dlugokencky, E. J., L.P. Steele, P. M. Lang and K.A. Masarie, 1995: Atmospheric methane at Mauna Loa and Barrow observatories: Presentation and analysis of in situ measurements. J. Geophys. Res., 100, 23103-23113. Enting, I.G. and J. V. Mansbridge, 1991: Latitudinal distribution of sources and sinks of CO2: Results of an inversion study. Tellus, 43B, 156-170. Etheridge, D.M., G.I. Pearman and F. de Silva, 1988: Atmospheric trace-gas variations as revealed by air trapped in an ice core from Law Dome, Antarctica. Ann. Glaciol., 10, 28-33. Fung, I., J. John, J. Lerner, E. Matthews, M. Prather, L.P. Steele and P. J. Fraser, 1991: Three-dimensional model synthesis of the global methane cycle. J. Geophys. Res., 96, 13033-13065. Gupta, M., S. Tyler and R. Cicerone, 1996: Modeling atmospheric b13CH4and the causes of recent changes in atmospheric CH4amounts. J. Geophys.Res., 101, 22923-22932. Hao, W.M. and D.E. Ward, 1993: Methane production from global biomass burning. J. Geophys. Res., 98, 20657-20661. Hem, T., P. J. Crutzen and M. Heimann, 1997: An inverse modeling approach to investigate the global atmospheric methane cycle. Global Biogeochem. Cycles, 11, 43-76. Higuchi, K., 1983: Effect of two different atmospheric models on the absorptive rate of excess atmospheric carbon by the sea. Geophys.Res. Lett., 10, 869-872. Hough, A.M., 1991: Development of a two-dimensional global tropospheric model: Model chemistry. J. Geophys. Res., 96, 7325-7362. Jacob, D.J.; M.J. Prather, S.C. Wofsy and MB. McElroy, 1987: Atmospheric distribution of 85Kr simulated with a general circulation model. J. Geophys. Res., 92, 6, 614-6,626. Jet Propulsion Laboratory (JPL), 1990: Chemical kinetics and photochemical data for use in stratospheric

April

1998

T. Saeki,

T. Nakazawa,

modeling. JPL Publ. 90-1, 15-33. Jet Propulsion Laboratory (JPL), 1997: Chemical kinetics and photochemical data for use in stratospheric modeling. JPL Publ. 97-4, 14-125. Khalil, M.A.K. and R.A. Rasmussen, 1990: Constraints on the global sources of methane and an analysis of recent budgets. Tellus, 42B, 229-236. Khalil, M.A.K. and R. A. Rasmussen, 1993: The environmental history and probable future of fluorocarbon 11. J. Geophys. Res., 98, 23901-23106. Langenfelds, R. L., R. J. Francey, L.P. Steele, P. J. Fraser, S.A. Coram, MR. Hayes, D.J. Beardsmore, M.P. Lucarelli and F.R. deSilva, 1993: Improved vertical sampling of the trace gas composition of the troposphere above Cape Grim since 1991. Baseline '93, 46-57. Law, K.S. and J. A. Pyle, 1993: Modeling trace gas budgets in the troposphere 2. CH4 and CO. J. Geophys. Res., 98, 18401-18412. Lowe,D.C., C.A.M. Brenninkmeijer, S.C. Tyler and E. J. Dlugokencky, 1991: Determination of the isotopic composition of atmospheric methane and its application in the Antarctic. J. Geophys. Res., 96, 1545515467. Lowe, D.C., C.A.M. Brenninkmeijer, G.W. Brailsford, K.R. Lassey, A.J. Gomez and E. G. Nisbet, 1994: Concentration and 13C records of atmospheric methane in New Zealand and Antarctica: Evidence for changes in methane sources. J. Geophys. Res., 99, 16913-16925. Madronich, S. and C. Granier, 1992: Impact of recent total ozone changes on tropospheric ozone photodissociation, hydroxyl radicals, and methane trends. Geophys. Res. Lett., 19, 465-467. Matthews, E. and I. Fung, 1987: Methane emission from natural wetlands: Global distribution, area, and environmental characteristics of sources. Global Biogeochem. Cycles, 1, 61-86. Murayama, S., T. Nakazawa, K. Yamazaki, S. Aoki, Y. Makino, M. Shiobara, M. Fukabori, T. Yamanouchi, A. Shimizu, M. Hayashi, S. Kawaguchi and M. Tanaka, 1995: Concentration variations of atmospheric CO2 over Syowa Station. Antarctica and their interpretation, Tellus, 47B, 375-390. Nakazawa, T., K. Miyashita, S. Aoki and M. Tanaka, 1991: Temporal and spatial variations of upper tropospheric and lower stratospheric carbon dioxide. Tellus, 43B, 106-117. Nakazawa, T., T. Machida, M. Tanaka, Y. Fujii, S. Aoki and O. Watanabe, 1993a: Differences of the atmospheric CH4 concentration between the Arctic and Antarctic regions in pre-industrial/pre-agricultural era. Geophys. Res. Lett., 20, 943-946. Nakazawa, T., S. Morimoto, S. Aoki and M. Tanaka, 1993b: Time and space variations of the carbon isotopic ratio of tropospheric carbon dioxide over Japan. Tellus, 45B, 258-274. Nakazawa, T., S. Morimoto, S. Aoki and M. Tanaka, 1997a: Temporal and spatial variations of the carbon isotopic ratio of atmospheric carbon dioxide in the western Pacific region. J. Geophys.Res., 102, 12711285.

M. Tanaka

and

K. Higuchi

323

Nakazawa, T., M. Ishizawa, K. Higuchi and NBA. Trivett, 1997b: Two curve fitting methods applied to CO2 flask data. EnvironMetrics, 8, 197-218. Nakazawa, T., S. Sugawara, G. Inoue, T. Machida, S. Makshyutov and H. Mukai, 1997c: Aircraft measurements of the concentrations of C02, CH4, N20, and CO and the carbon and oxygen isotopic ratios of CO2 in the troposphere over Russia. J. Geophys. Res., 102, 3843-3859. Pearman, G.I. and D.J. Beardsmore, 1984: Atmospheric carbon dioxide measurements in the Australian region: Ten years of aircraft data. Tellus, 36B, 1-24. Plumb, R.A. and D.D. McConalogue, 1988: On the meridional structure of long-lived tropospheric constituents. J. Geophys.Res., 93, 15897-15913. Plumb, R. A. and J. D. Mahlman, 1987: The zonally averaged transport characteristics of the GFDL general circulation/transport model. J. Atmos. Sci., 44, 298-327. Prather, M., M. McElroy, S. Wofsy, G. Russell and D. Rind, 1987: Chemistry of the global troposphere: Fluorocarbons as tracers of air motion. J. Geophys. Res., 92, 6579-6613. Prinn, R., D. Cunnold, R. Rasmussen, P. Simmonds, F. Alyea, A. Crawford, P. Fraser and R. Rosen, 1987: Atmospheric trends in methylchloroform and the global average for the hydroxyl radical. Science, 238, 945-950. Prinn, R., D. Cunnold, P. Simmonds, F. Alyea, R. Boldi, A. Crawford, P. Fraser, D. Gutzler, D. Hartley, R. Rosen and R. Rasmussen, 1992: Global average concentration and trend for hydroxyl radicals deduced from ALE/GAGE trichloroethane (methyl chloroform) data for 1978-1990. J. Geophys. Res., 97, 2445-2461. Prinn, R.G., R.F. Weiss, BR. Miller, J. Huang, F. N. Alyea, D.M. Cunnold, P. J. Fraser, D.E. Hartley, P.G. Simmonds, 1995: Atmospheric trends and lifetime of CH3CC13and global OH concentrations. Science, 269, 187-192. Quay, PD., S.L. King, J. Stutsman, DO. Wilbur, L.P. Steele, I. Fung, R.H. Gammon, TA. Brown, G.W. Farwell, P.M. Grootes and F. H. Schmidt, 1991 Carbon isotopic composition of atmospheric CH4: Fossil and biomass burning source strengths. Global Biogeochem. Cycles, 5, 25-47. Saueressig, G., P. Bergamaschi, J. N. Crowley,H. Fischer and G.W. Harris, 1995: Carbon kinetic isotope effect in the reaction of CH4 with C1 atoms. Geophys.Res. Lett., 22, 1225-1228. Spivakovsky, CM., R. Yevich, J. A. Logan, S.C. Wofsy, M.B. McElroy and M.J. Prather, 1990: Tropospheric OH in a three-dimensional chemical tracer model: An assessment based on observations of CH3CC13. J. Geophys.Res., 95, 18441-18471. Steele, L.P., P.J. Fraser, R.A. Rasmussen, M.A.K. Khalil, T. J. Conway, A.J. Crawford, R.H. Gammon, K.A. Masarie and K.W. Thoning, 1987: The global distribution of methane in the troposphere. J. Atmos. Chem., 5, 125-171. Steele, L.P., E. J. Dlugokencky, P. M. Lang, P.P. Tans, R. C. Martin and K.A. Masarie, 1992: Slowing down of the

global

accumulation

of atmospheric

methane

324

Journal

of the

Meteorological

during the 1980s. Nature, 358, 313-316. Stevens, C.M. and A. Engelkemeir, 1988: Stable carbon isotopic composition of methane from some natural and anthropogenic sources. J. Geophys. Res., 93, 725-733. Sugawara, S., T. Nakazawa, G. Hashida, C. Sakai and M. Tanaka, 1994: Variations of atmospheric methane concentration over the Western Pacific Ocean. Proceedings of the International Symposium on Global Cycles of Atmospheric Greenhouse Gases, 121-128. Sugawara, S., T. Nakazawa, G. Inoue, T. Machida, H. Mukai, N.K. Vinnichenkoand V.U. Khattatov, 1996: Aircraft measurements of the stable carbon isotopic ratio of atmospheric methane over Siberia. Global Biogeochem. Cycles,10, 223-231. Tanaka, M., T. Nakazawa and S. Aoki, 1987: Time and space variations of tropospheric carbon dioxide over Japan. Tellus, 39B, 3-12. Tanaka, N., Y. Xiao and AC. Lasaga, 1996: Ab initio study on carbon kinetic isotope effect (KIE) in the reaction of CH4+Cl•. J. Atmos. Chem., 23, 37-49. Tans, P.P., T. J. Conway and T. Nakazawa, 1989: Latitudinal distribution of the sources and sinks of atmospheric carbon dioxide derived from surface observations and an atmospheric transport model. J. Geophys. Res., 94, 5151-5172. Tans, P.P., P.S. Bakwin, T. J. Conway, R.W. Dissly,

2次

Society

of Japan

Vol. 76, No. 2

E. J. Dlugokencky, L. S. Geller, D.W. Guenther, D.F. Hurst, D.R. Kitzis, P.M. Lang, MA. Masarie, J. B. Miller, P. C. Novelli, C. Prostko-Bell, M. Ramonet, K.W. Thoning, M. Trolier, L.S. Waterman, N. Zhang and C. Zhao, 1996: Carbon Cycle. Climate Monitoring and Deagnostics Laboratory, No. ~3, Summary Report 1994-1995, 29-49. Taylor, J. A., G.P. Brasseur, P.R. Zimmerman and R.J. Cicerone, 1991: A study of the sources and sinks of methane and methyl chloroformusing a global threedimensional Lagrangian tropospheric tracer transport model. J. Geophys. Res., 96, 3013-3044. Tyler, S.C., P. M. Crill and G.W. Brailsford, 1994: 13C/12C fractionation of methane during oxidation in a temperate forested soil. Geochim. Cosmochim. Acta, 58, 1625-1633. Vaghjiani, G.L. and AR. Ravishankara, 1991: New measurement of the rate coefficientfor the reaction of OH with methane. Nature, 350, 406-409. Weiss, W., A. Sittkus, H. Stockburger, H. Sartorius and K.O. Munich, 1983: Large-scale atmospheric mixing derived from meridional profiles of 85Kr. J. Geophys. Res., 88, 8574-8578. Zimmermann, P.H., J. Feichter, H.K. Rath, P.J. Crutzen and W. Weiss, 1989: A global three-dimensional source-receptor model investigation using 85Kr. Atmos. Environ., 23, 25-35.

元 大 気 輸 送 モ デ ル と 地 表 面 濃 度 の 観 測 値 か ら導 出 し た メ タ ン 放 出 量 佐 伯 田鶴

・中 澤 高 清 ・ 田 中 正 之

(東 北 大 学 大 学 院 理 学 研 究 科 大 気 海 洋 変 動 観 測 研 究 セ ン タ ー) Kaz

Higuchi

(カ ナ ダ 大 気 環 境 庁)

1983年

か ら1994年

にNOAA/CMDLネ

ル を用 い て解 析 し、CH4放 CFC-12、CO2濃

ッ トワ ー ク に お い て 観 測 さ れ たCH4濃

出 の 緯 度 分 布 お よ び時 間 変 動 を 推 定 した 。 モ デ ル の 大 気 輸 送 は85Kr、CFC-11、

度 を シ ミュ レー トす る こ と に よ り検 証 した 。CH4の

を 考 慮 し、3次 元 光 化 学 モ デ ル か ら導 か れ た 経 度 平 均 のOH場

消 滅 源 と し て は0Hラ

をCH3CCI3に

放 出 量 の 平 均 的 な 緯 度 分 布 は 大 き な 南 北 勾 配 を示 し、 全 体 の75%が られ た 。 北 半 球 高 緯 度 で のCH4放 て は 明 瞭 なCH4放 559±9Tg/yr、

出 の 季 節 変 化 は 見 られ な か っ た 。1984年

北 半 球 で 放 出 さ れ る と い う結 果 が 得

か ら1994年

大 気 増 加 は31Tg/yrで

全 球 放 出 量 はOH濃

の 平 均 的 な 全 球 のCH4放

出量 は

あ っ た 。 モ デ ル か ら得 られ た 結 果

度 と温 度 に敏 感 で あ る が 、使 用 す る 大 気 輸 送 係 数

度 の 観 測 値 の 違 い に よ る影 響 は 小 さい こ とが 判 明 した 。CH4放

送 係 数 に 大 き く依 存 す る こ と も分 か っ た 。 また 、大 気 中CH4の め る バ ク テ リ ア起 源 のCH4の

ジ カル との反応

よ っ て 検 証 し て用 い た 。FCH4

出 は 、 夏 に 最 大 値 に達 す る季 節 変 動 化 を 示 した 。 一 方 、 南 半 球 に お い

化 学 消 滅 量 は528±10Tg/Yr、

の 感 度 試 験 を行 っ た と こ ろ 、CH4の とCH4濃

度 を2次 元 大 気 輸 送 モ デ

出 め 緯 度 分 布 は水 平 方 向 の 大 気 輸

δ13Cの 分 布 に とっ て 、 放 出 量 の 大 部 分 を 占

δ13C値 と 同 位 体 分 別 効 果 の 大 きい 土 壌 吸 収 が 重 要 で あ る こ とが 示 唆 さ れ た 。