Atmospheric Environment 33 (1999) 1305—1321
Atmospheric methyl chloride
M.A.K. Khalil *, R.A. Rasmussen
Department of Physics, Portland State University, P.O. Box 751, Portland, OR 97207, USA
Department of Environmental Science and Engineering, Oregon Graduate Institute, P.O. Box 91000, Portland, OR 97291, USA
Received 8 September 1997; accepted 10 June 1998
Abstract
There are about 5 Tg of methyl chloride in the Earth’s atmosphere making it one of the largest reservoirs of gas-phase
chlorine. We discuss the time series of global measurements taken over the last 16 yr at seven locations distributed among
the polar, middle, and tropical latitudes of both hemispheres (1981—1997). Measurements were also taken at 20 more sites
between 1987 and 1989. The vertical distribution was measured during campaign experiments in the Arctic, Western
Atlantic, and over Brazil. Small, mostly decreasing trends are observed, showing that on average, there was 4% less
methyl chloride during the last three years (1994—1996) than there was in the first three years (1985—1987) of the
experiment. The latitudinal variation is marked by highest concentrations in the tropics and lowest in the polar regions.
Sites representing inland locations show higher concentrations, suggesting continental sources, mostly confined to the
tropics. There are seasonal variations at various latitudes that can be explained mostly by the cycles of OH radicals,
which are the dominant removal process for methyl chloride in the atmosphere. Based on these data, the expected
emissions can be calculated at the polar, middle, and tropical latitudes represented by the six long-term primary sites.
Using a photochemical model of OH, we estimate that a global source of about 3.7 Tg yr\ of methyl chloride is needed
to explain the observed concentrations. Other removal processes have been identified that may add to this estimate of the
global annual emissions. The results further establish that some 85% of the emissions must come from the half of the
earth’s surface between 30°S and 30°N, representing tropical and sub-tropical latitudes. Small emissions are estimated for
the middle latitudes, and no emissions are expected from the polar regions. 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: Methyl chloride; Chloromethane; Reactive chlorine; Global emissions; Troposphere
1. Introduction
Even though it is present at only about 600 pptv,
methyl chloride is one of the largest reservoirs of gaseous
chlorine in the earth’s atmosphere (5 Tg). It is thought to
be mostly of natural origins, and as such, it is probably
the largest natural source of chlorine to the stratosphere.
* Corresponding author.
Chlorine-containing gases catalytically destroy ozone in
the stratosphere, which amplifies their effect on the environment even if they exist at exceedingly rare concentrations. Thus, man-made chlorofluorocarbons and similar
compounds are causing ozone depletion and the Antarctic ozone hole by adding substantially to the natural
chlorine in the stratosphere (WMO, 1989, 1994). While
there has been a great deal of research on the chlorofluorocarbons, relatively little is known about methyl
chloride. In recent years there has been renewed interest
in the global budget of methyl chloride, chloroform, and
1352-2310/99/$ — see front matter 1999 Elsevier Science Ltd. All rights reserved.
PII: S 1 3 5 2 - 2 3 1 0 ( 9 8 ) 0 0 2 3 4 - 9
1306
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
other gases as sources of natural chlorine in the atmosphere (see Graedel and Keene, 1995, 1996).
The main source of information on the global balance
of methyl chloride (CH Cl) is the atmospheric record,
which in our experiments spans sixteen years. These data
establish the climatology of methyl chloride in the atmosphere. The purpose of this paper is to analyze the data
on the atmospheric distribution and trends of methyl
chloride and to deduce the expected global emissions,
which can then be attributed to various anthropogenic
and natural sources. Similar results on chloroform will be
discussed in a companion paper (Khalil and Rasmussen,
1998b).
Until now it was thought that most of the global
emissions of methyl chloride came from the oceans based
on early measurements reported by Singh et al. (1979,
1983). While this work established the idea that the
oceans were a major source of methyl chloride, new
intensive experiments on oceanic emissions suggest that
this source is at most 0.5 Tg yr\, which is much smaller
than previously thought (Moore et al., 1996; Tait et al.,
1994; Khalil and Rasmussen, 1998a). Biomass burning
contributes about 0.5 Tg yr\ (Palmer, 1976; Andreae,
1991; Reinhardt and Ward, 1995; Rudolph et al., 1995).
Some of the remaining emissions, up to 0.6 Tg yr\ at
most, may come from terrestrial biological processes in
the soils and from fungi responsible for cycling carbon in
forests (Harper, 1985; White, 1982; Harper, personal
communication). The identified sources are less than half
the amount needed to explain the methyl chloride observed in the atmosphere, which requires a source of
3.7 Tg yr\. These findings are a major shift in our understanding of the global budget of methyl chloride and
show that more work is needed to produce a closed
budget of methyl chloride which may include sources not
known at present.
2. The experiments
2.1. Sites and sampling frequency
Starting around 1980, we established a network of sites
for systematically collecting air samples in stainless steel
flasks (0.8 l) especially designed for sampling and
preserving clean air for the measurement of trace gases at
the parts per trillion to parts per billion levels. The
samples are sent back to our laboratory for analysis.
Many trace gases were measured in these flasks, depending on the changing instrumental capabilities of the laboratory over the years. Methyl chloride is one of the
gases that has been measured from the early 1980s to
now.
The samples are collected in triplicate each week. This
ensures a representative measurement and allows for an
evaluation of successful sampling. Most of the time
(&99%) all three samples agreed well and were averaged.
On occasion, the concentrations in the three samples
disagreed significantly from one another, and such data
were discarded. From the weekly data, monthly averages
are calculated, which are used for further analysis. We
consider the monthly average concentrations to be statistically independent of each other as they represent
different air masses.
We obtained samples from seven sites for much of the
last sixteen years (1981—1997): Pt. Barrow, Alaska
(71.16N, 156.5W); Cape Meares, Oregon (45.5N, 124W);
Cape Kumukahi and Mauna Loa, Hawaii (19.3N,
154.5W); Cape Matatula, Samoa (14.1S, 170.6W); Cape
Grim, Tasmania (42S, 145E), and Antarctica (South Pole
at 90S; and Palmer Station, 65.46S, 64.05W). These sites
were selected to capture the effects of the major global air
circulation patterns on the latitudinal dispersal of trace
gases. We have a site in the polar, middle and tropical
latitudes of each hemisphere. Most sites are near sea level
and represent the marine boundary layer.
For a short time, between 1987 and 1989, we also
obtained data from many other sites distributed globally.
Methyl chloride was measured at 20 of these sites, which
were: Alert (82.30N, 62.00W), NW Territories, Canada;
Ragged Point, Barbados (13.10N, 59.26W); Bermuda
(32.20N, 64.50W); Boola Boola (38.11S, 146.18E), Victoria, Australia; Galway (53.16N, 9.03W), Ireland; Jabiru
(5.35S, 64.05E), Amazonas, Brazil; Minqin (38.41N,
103.11E), Gansu, China; Panama (9.00N, 80.00W); Poker
Flats (69.00N, 154.00W), Alaska; Sable Island (43.57N,
60.00W), Nova Scotia, Canada; St. Croix (17.35N,
63.40W), Virgin Islands; Utlangan (56.02N, 15.50E),
Sweden; three sites in Brazil during the ABLE experiments: Belem (1.27S, 48.29W), Barcarena (1.3S, 48.38W),
Cashoeira Paulista (22S, 45.01W); two in Norway: Ship
M (57N, 5W) and Ny-Alsund (62N, 3W); two sites in
Florida: Key Biscayne (24N, 81W) and the Everglades
(26N, 81W); and Guam (13.5N, 144E). The duration of
the measurements at these sites is documented in the
archived data base.
In addition to these long-term sites and short-term
background sites, we have obtained data from campaign
experiments including vertical distributions at different
latitudes. The main data are from Arctic Gas and Aerosol
Sampling Program I, II, and III (AGASP) conducted
during April—May 1983, 1986, and 1989 and spanning
latitudes 60N—90N; Global Tropospheric Experiment
(GTE), based in Barbados, during 17—24 June 1984,
spanning latitudes 2.4N—18N, longitudes 52.2W—60.6W,
and altitudes of 0.15 to 5.3 km; GTE Pacific flights during 4—26 November 1983, spanning 30N—41N latitudes,
119W—147W longitudes and 0.25—10 km altitudes; the
same set of GTE flights also covered 8N—30N latitudes,
127W—189W longitudes, and 0.15—9.7 km altitudes;
Western Atlantic Oxidant Study (WATOX), flights
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
during 4—10 January 1986, spanning latitudes
31.7N—43.7N, longitudes 59.6W—75.9W, and altitudes of
about 0.15—6.1 km; and Amazon Boundary Layer Experiment 2A and 2B. ABLE 2A was during July—Aug,
1985, over 0—16N latitudes, 44—66W longitudes, and
0.15—5 km altitudes; and ABLE 2B was during
April—May, 1987, spanning latitudes 3S—30N, longitudes
49—74, and altitudes 0.1—4.6 km. These experiments provide data on the vertical distribution of methyl chloride
at various latitudes (see Harriss et al., 1988, 1990; Beck et
al., 1987; Albritton, 1989; Davidson and Schnell, 1993;
Khalil and Rasmussen, 1984a, 1988).
2.2. Laboratory measurements
The samples are analyzed within 2—3 weeks of collection, except from locations such as the South Pole or
Palmer Station. Samples from these sites are not available for analysis for several months after being collected.
Methyl chloride concentrations have been shown to remain constant in these containers for much longer storage times (Edgerton, 1985). The concentrations of methyl
chloride are measured using an Electron Capture Gas
Chromatograph by methods described in earlier publications (Rasmussen and Khalil, 1980; Rasmussen et al.,
1980; Edgerton, 1985). Concentrations are reported as
mixing ratios in dry air. The concentrations are determined by using a set of calibration standards that are
referenced against a primary standard which is also used
to establish the absolute concentration. Periodic
measurements of the concentrations in the secondary
standards are used to determine the stability of the calibration or the limits of the possible drifts in time. For
the two main tanks used in these experiments the drifts
were (!0.4$0.6) pptv yr\ (1990—1998) and !0.8$2
pptv yr\ (1983—1998) where the$values reflect the
standard error of the trend. Neither is large or statistically significant.
The primary calibration standards for methyl chloride
are prepared by individual investigators in the absence of
an available standard from a centralized location, or
available standards are diluted to make them usable at
the low concentrations measured in the non-urban
troposphere. The creation of independent standards
often leads to differences of absolute concentrations
between different investigators. These differences can be
corrected by inter-laboratory calibrations, which have
not been done for methyl chloride. Data on atmospheric
concentrations have been published by other investigators for similar locations. These can be used for an
inter-comparison to establish differences of absolute calibration. Although ideally, these comparisons should be
done at the same locations and for the same times, such
data are not available for methyl chloride. We analyzed
data from six independent experiments, two from our
own research and four others (Koppman et al., 1993;
1307
Singh et al., 1983; Atlas et al., 1993; Moore et al., 1996;
Tait et al., 1994). We found that there are relatively small
differences among investigators as shown in Fig. 1. In
this figure we have taken the average data in the latitudinal bands represented by our long-term measurements.
We then estimated the percent deviations of the observations of methyl chloride in these studies from the
long-term averages. Since the other experiments do not
represent long-term measurements, they are subject to
the effects of natural variability of concentrations in the
atmosphere. We estimated this variability as the relative
standard deviation of the monthly concentrations observed at our long-term sites (%Var"Standard Deviation/Mean Concentrations;100%). The estimates of
variability are based on de-seasonalized data (to be discussed later). The limits of natural variability are of
the order of 2% expressed as one standard deviation and
are shown on the figure. We see that our measurements
in different campaigns are more or less within these limits
of natural variability, as are the early measurements
of Singh et al., (1983) and the more recent measurements
of Atlas et al., (1993). Two experiments reported in the
literature are substantially below the limits of natural
variability and the absolute concentrations inherent in
our long-term studies (Koppman et al., 1993; Moore
et al., 1996 and Tait et al., 1994). These are both about
10% lower than our absolute calibration. Since these are
newer calibrations, they may be more accurate. We see
therefore, that the results of the present paper may be
subject to systematic errors of absolute calibration of the
order of 10%. The implication of this result is that our
data may be high by 10%. If so, we would overestimate
the global burden and the total global emissions, when
deduced from these data, by the same amount (10%). It is
noteworthy that while different experimenters may disagree on the absolute calibration they should agree on
the patterns of latitudinal distribution and the percent
rate of change. For the data we have on hand, the
different experiments do agree on the latitudinal pattern
(see Khalil et al., 1998), but there are no comparable data
to evaluate the agreement on trends.
3. Global distributions and trends
3.1. Seasonal and temporal trends
3.1.1. Time series
The monthly average concentrations of methyl chloride at the long-term sites are shown in Fig. 2. These time
series contain two main components: the seasonal cycles
and possible trends over the period of the study.
3.1.2. Seasonal cycles
The seasonal patterns for the long-term sites were
determined by using a moving average filter to subtract
1308
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
Fig. 1. A comparison of absolute concentrations of methyl chloride reported by independent laboratories and our experiments over
the oceans and at marine sites. The data from these studies (C) are averaged in bands representative of our long-term sampling sites. The
percent deviation is calculated relative to the average concentrations (C ) observed at our long-term sites (C/C !1);100%. The
*2
*2
dark lines represent the natural variability of methyl chloride at the locations sampled and is represented as the standard deviations of
the monthly concentrations observed at the long-term sites. SAGA-1 are data from a cruise of the Academik Korolev in 1983; K and
R short refers to data from our short-term sites, and the remaining experiments in the legend are referenced in the text. For locations
N and S represent Northern and Southern hemispheres; P, M and T represent Polar, Middle and Tropical latitudes.
cycles of periods longer than 12 months, which includes
long-term trends. The resulting time series were averaged
for each of the 12 months over the years when measurements were taken. The result is 12 indices which represent
our estimate of seasonal cycles. These seasonal indices
are then subtracted from the original monthly data to
obtain a de-seasonalized time series. The de-seasonalized
time series are used for determining trends and residual
variability. Methods for decomposing time series are
described in more detail by Khalil and Rasmussen (1990).
In Fig. 3a we show the cycles at the long-term sites.
The results for the northern hemisphere are further supported by data from the short-term sites in Fig. 3b.
Similar data are not available for the southern hemisphere. We calculated the cycle amplitude as the difference between the average concentrations during
February—April and August—October each year. The
average cycle amplitudes are shown in Fig. 4. These
choices of months provides the most robust estimate of
the cycle amplitude, taking into account the differences
among the sites. In the southern hemisphere however,
a different choice of 3-month periods can increase the
cycle amplitude by up to 5 pptv.
The two Hawaiian sites are close to each other, except
that Mauna Loa is at 3.4 km above sea level and Cape
Kumukahi is on the beach. These two sites provide data
on the long-term differences of concentrations of trace
gases with altitude. For methyl chloride, the most useful
data are from 1992 to the present, when reliable samples
were obtained from both sites. The concentrations for
this period are shown in Fig. 5. The seasonal cycles are
similar at both sites except that the minimum is deeper at
Cape Kumukahi compared to Mauna Loa.
It is apparent that the cycle amplitude decreases
with latitude, which is expected from the seasonal
pattern of changes in atmospheric hydroxyl radicals
that remove methyl chloride from the atmosphere. Moreover, the cycle amplitude is nearly half as large in the
southern hemisphere middle and higher latitudes as it is
at similar latitudes in the northern hemisphere. These
patterns are consistent with the results of our photochemical model and of similar models. The calculated
seasonal cycle of OH is smaller in the southern hemisphere by about the same amount as the methyl chloride
cycle. The smaller seasonality of southern hemisphere
OH is due in part to the much lower CO concentration in
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Fig. 2. The atmospheric concentrations of methyl chloride between 1981 and 1997 at Polar, Middle, and Tropical latitudes of both
hemispheres. Data are from (a) Barrow, Alaska; (b) Cape Meares, Oregon; (c) Cape Kumukahi, Hawaii; (d) Samoa; (e) Cape Grim,
Tasmania; (f) Palmer Station, Antarctica; and the South Pole. (g) The latitudinally weighted globally and monthly averaged
concentrations of methyl chloride.
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
1309
1310
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
Fig. 3. (a) The seasonal cycles of methyl chloride at the long-term sites (see Fig. 1). (b) The seasonal cycle at short-term sites in the
northern hemisphere (Barrow and Cape Meares composite included for comparison).
the southern hemisphere and hence a smaller CO seasonal cycle. Since CO is a major sink of OH, smaller
concentrations in the southern hemisphere lead to
a smaller seasonality of OH.
3.1.3. Trends
The long-term trends appear to be small and unsteady.
We used two methods to estimate the trends — a linear
regression model, and the average difference between the
first and last three years of measurements. For the period
between 1/1985 and 12/1996, we have data from all the
long-term sites. During this time there has been a decrease in the concentrations at Cape Meares, Barrow and
Mauna Loa and an increase at Tasmania with no trends
elsewhere (Fig. 6). For the more recent period the difference of concentrations between the two three year
periods (1/1991—12/1993) and (1/1994—12/1996), the same
patterns persist, except that negative trends are also seen
at Samoa.
These trends are unlikely to be caused by drifts in the
calibration standards, as no long-terms drifts were observed. The concentrations of methyl chloride would
have to be increasing in the standard containers to account for a decrease in the atmospheric concentrations,
but there is no evidence for this. Moreover, since the
same standards are used for all the sites, the trends, if due
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
1311
Fig. 4. The latitudinal distribution of the seasonal cycle amplitude. The cycle amplitude is defined as the difference between average
concentrations during February—April and August—October. The short-term sites (without error bars) from north to south latitudes are:
Alert, Poker, Flats, Galway, Suble Is., Bermuda, St. Croix and Barbados (see text for detailed information).
Fig. 5. The seasonal cycle at the two sites in Hawaii. Cape Kumukahi is at sea level and Mauna Loa is at 3.4 km.
to drifts in calibration standards, should all be in the
same direction and of similar magnitude, which is also
not the case for the observed concentrations. If these
negative trends are representative of the atmosphere, two
root causes are a decrease in emissions or an increase in
the destruction rate, which could occur if OH concentrations were increasing, for instance. At present, it is difficult to associate a specific cause to explain these observations, especially since the global budget is not balanced
by the emissions from known sources. The situation
favors the explanation due to changes in OH concentrations since the ‘‘missing’’ sources are believed to be natural as we will discuss later, and long-term trends in
natural emissions are less likely to occur. There has been
speculation on the increasing trend of OH for some time,
but definitive estimates do not exist. The most recent
assessment by Krol et al. (1998), based on the budget of
methyl chloroform, suggests a trend of OH at about
1312
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
Fig. 6. Trends of methyl chloride at various latitudes and the 90% confidence limits of the trend.
0.5% yr\ for the time period over which our measurements were taken. This would be roughly consistent with
the trends reported here. A simple one box model calculation shows that by keeping emissions constant for the
period of our observations at 3.7 Tg yr\, the decreasing
global trend of methyl chloride can be explained if OH
increased by 0.3%$0.1% yr\.
3.2. Spatial distribution
3.2.1 Latitudinal distribution
Based on the data from the long-term sites, a climatology of the latitudinal distribution can be established. The
patterns are quite remarkable and unlike the latitudinal
distributions of most other long-lived gases. The average
concentrations between 1985 and 1996 are shown in
Fig. 7a. The concentration of methyl chloride is lower at
higher latitudes and highest in the tropics. These data are
representative of the marine boundary layer, and, for
consistency, we have used only the data from Cape
Kumukahi in Hawaii to represent the northern tropics.
Averages over shorter periods are also shown in Fig. 7a.
Despite some variability, the general latitudinal pattern
described above is always the same. These results suggest
that the tropics are a sizable source of methyl chloride
and that there is likely to be a net transport to higher
latitudes of both hemispheres.
For the short-term sites, the latitudinal distribution is
shown in Fig. 7b. The seasonal cycles were subtracted
before plotting the data. This is to ensure that there is no
bias in the average because the data represent only part
of the year, and because there are only data for a few
years. Many of these sites do not represent the marine
boundary layer as the long-term sites do. A qualitative
classification is included in the figure showing that concentrations of methyl chloride are higher at sites that are
clearly continental and mixed at other sites that represent
marine air. At such sites, continental pollution can affect
the concentrations, as evidenced by high levels of manmade chlorofluorocarbons and other gases. Sites where
there is evidence of such anthropogenic influences are
classified as ‘‘Polluted Marine,’’ and the others are classified as ‘‘Clean Marine.’’ As these data are only from
1987—1989, the averages for the long-term sites are also
&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
Fig. 7. (a) The latitudinal distribution of methyl chloride at the long-term sites (1/1981—1/1997). The vertical bars are the standard
deviations of monthly concentrations at each latitude. The standard errors of the mean are too small to appear on the figure
(2.5—3.5 pptv). (b) The average concentrations of methyl chloride at the short-term sites where measurements were taken between 1987
and 1989. Data from the long-term sites (Fig. 1) are shown for the same period for comparison. The short-term sites from north to south
latitudes are: Alert, Poker Flats, Ny-Alsund, Ship M (Norway), Utlangen, Galway, Jungfraujoch, Sable Is, Minqin, Bermuda, Florida
Everglades, Key Biscayne, St. Croix, Guam, Barbados, Panama, Belem, Barcerana, Jabirou, Cachoeira Paulista, Boola Boola (see text
for detailed information). (c) The latitudinal distribution of methyl chloride during different seasons.
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
Fig. 7.
1313
1314
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
taken for the same period. The figure shows that there are
generally higher concentrations of methyl chloride when
measurements are taken at continental sites or where
there is evidence that the air masses are of continental
origin. These results suggest that there are sizable landbased sources. This is encouraging, since the current data
on oceanic emissions of methyl chloride do not account
for most of the methyl chloride in the atmosphere.
Since the ocean emissions are very diffuse compared to
land-based sources such as biomass burning, the contributions of the latter would be more readily observable in
the atmosphere than emissions from the oceans.
Often data for atmospheric methyl chloride are reported from ship-board measurements. In such cases, the
latitudinal variability is masked both by the variability of
the short-term measurements and also by the seasonal
changes in the latitudinal gradients. The latitudinal
gradients at any time during the year are a combination
of the seasonal variations at each site (Fig. 3) and the
prevailing mean latitudinal distribution (Fig. 7). This
combination causes very little latitudinal change over
most of the tropical and middle latitudes during fall and
smaller than usual gradients during the summer. The
seasonal latitudinal distributions are shown in Fig. 7c.
Earlier measurements from ship cruises may have been
affected by the seasonality of the latitudinal distribution
(Lovelock 1975; Singh et al., 1983; Moore et al., 1996;
Tait et al., 1994).
3.2.2. Altitudinal profiles
There are no long-term data on the vertical distribution of methyl chloride and for most other trace gases, as
frequent, systematic measurements are taken mainly at
ground level. Measurements taken at locations such as
Cape Kumukahi and Mauna Loa in Hawaii, are unique
in establishing an approximate long-term climatology of
the vertical distribution of trace gases. Most of our information on the change of concentrations with altitude
is based on samples collected on aircraft flights in various
parts of the world. Our data are from 8 such experiments
conducted between 1983 and 1989: AGASPs 1, 2, and 3;
GTE flights (Pacific and Atlantic); ABLE 2A and 2B; and
WATOX as described earlier in the paper. In these experiments samples were collected in the same manner as
at the long-term sites and analyzed using the same instruments and standards as for the site data. We can construct vertical concentration profiles for methyl chloride
from these data.
We classified the data from these experiments by latitude as ‘‘Arctic,’’ ‘‘Middle Latitudes,’’ and ‘‘Tropical Latitudes.’’ There are no data on vertical concentrations at
middle and higher southern latitudes. The patterns in the
data are best seen by taking averages of many measurements within specified heights, which we choose to be
0.5 km deep between 0 and 2 km, roughly representing
the boundary layer, and 1 km deep above 2 km to the
maximum height of the flights, which in some cases was
as low as 4 km and in others as high as 10 km. It should
be noted that the aircraft flights span relatively short
times of a few weeks at most. As such, the concentrations
observed can deviate significantly from the long-term
average concentrations shown earlier. For the sake of
comparison, however, we have included the average concentrations of methyl chloride at the comparable longterm sites for the same times. These results are shown in
Fig. 8. Earlier data from similar missions have been reported, extending well into the stratosphere (Penkett
et al., 1980; Fabian et al., 1996; Pierotti et al., 1980;
Schauffler et al., 1993).
The vertical distributions suggest two conclusions
about the budget of methyl chloride. First, that the higher tropical concentrations manifested in the long-term
time series are representative of the troposphere and not
just the surface concentrations. In tropical and other
lower latitudes the vertical concentrations of methyl
chloride are higher than the concentrations measured at
similar heights at middle or higher latitudes. Second, that
there are major sources of methyl chloride in the tropics.
The graphs show that the aircraft data are well within the
observations at the long-term sites for the arctic and
middle northern latitudes, but in the tropics the aircraft
data are generally higher than at the marine sites, and
these concentrations decrease with altitude, which is
a signature of ground-based sources. This is especially
notable in the 1985 ABLE experiments, which could
represent the effect of tropical biomass burning.
It is also noteworthy that although Mauna Loa is
3.4 km higher than Cape Kumukahi, the average concentrations at the two sites are almost the same (Mauna Loa
is 3$3 pptv higher than Cape Kumukahi, where the
$are the 90% confidence limits of the difference). Little
or no average differences are seen between Palmer station and the South Pole in Antarctica even though the
pole is at 2.8 km and Palmer is at sea level. The lack of
a difference may be due to the few data available from the
antarctic. These middle and high latitude cases are quite
unlike the inland tropics where concentrations are high
at ground level and decrease with altitude, eventually
reaching values close to those observed in the marine
boundary layer.
4. Estimates of emissions
One of the uses of these data is to estimate the total
emissions of methyl chloride at various latitudes. We
have used a low-resolution two dimensional box model
of the atmosphere and a separate detailed photochemical
model for OH to estimate the emissions based on the
data from the long-term sites (Lu and Khalil 1991;
Moraes, 1995). The mass balance model consists of
6 tropospheric boxes spanning 60—90° to represent the
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
1315
polar regions and the data from Barrow and Antarctica;
30—60° latitude representing the middle latitudes and the
data from Cape Meares and Tasmania; and 0—30° representing the tropical and sub-tropical areas and the sites at
Hawaii and Samoa. Each tropospheric box is associated
with a stratospheric box with which air masses are
exchanged to take into account the losses of methyl
chloride by reactions with stratospheric OH and photodissociation. The transport within the stratosphere is
neglected.
The source de-convolution model for tropospheric
sources (S ) is written as:
2
d
C"S"XC
(1)
dt
d C
S
X
X
C
21
2
2 " 2 ! 22
dt C
O
X
X
C
12
11
1
1
C2"[C , C , C , C , C , C ]
2
C2"[C , C , C , C , C , C ]
S2"[S , S , S , S , S , S ]
2
diag X
21
N
N
N
N
N
N
"! . gs , gs , g , g , gs , gs
N
N
N
N N
N
diag X "![g , g , g , g , g ]
1
1 1 1 1
12
diag X "![g , g , g , g , g , g ]
11
(2)
(3a)
(3b)
(3c)
(3d)
(3e)
(3f)
&&&&&&&&&&&&&&&&&&&&&&&&&&
Fig. 8. The vertical distribution of methyl chloride at various
latitudes. For the Arctic profiles we took the concentrations at
Barrow, Alaska, for comparison. There are three sets of points:
a single point representing the average concentrations during
1983—1986, which is the span of the Arctic experiments when we
measured methyl chloride; the span above this average is the
maximum and minimum monthly average concentration measured between 1983 and 1986, and the span above this is the
monthly average concentrations during Aprils of the years when
the two AGASP experiments were conducted. The lower concentration at the long-term sites correspond to the 1986 when
lower concentrations of methyl chloride were also observed
during the aircraft flights. For the middle latitudes, we used
measurements at Cape Meares for comparison. The sequence of
points plotted is the same as already described for the Arctic
profile. For the tropical profiles, we used Cape Kumukahi and
Samoa for comparison. In this case we have plotted the average
during 1983—1986 as a single point and the ranges as maximum
and minimum monthly concentrations observed during this
period. The lower set of points and range is for Cape Kumukahi
and the upper for Samoa. Here we have not plotted the average
concentrations during the months when the flights were conducted (as for the middle and Arctic latitudes). These results suggest
that the high tropical concentrations at the long-term sites are
probably representative of the troposphere as concentrations
aloft are higher in the tropics compared to middle and high
latitudes.
1316
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
diag(X )" (g #g #g ),
22
2
N
#g #g #g
g
,
2 N
2
1
N
#g #g #g ,
g
2 N
2
N
#g #g #g ,
g
2 N
2
N
#g #g #g ,
g
2 N
2
1
(g #g #g )
2
1
(3g)
Upper Diag.
N
N
, g
[X ]"! g , g , g , g
22
2 2 2 2 N 2 N
Lower Diag.
N
N
,g
N ,g ,g ,g
[X ]"! g
22
2 N 2 N 2 2 2
dC
"!X !X C
12
11 1
dt
C (t)"e!X11 d C (t!d)
1
1
#X\ [I!e!X11d ]X C
11
12 2
dC
S " 2#X C #X C
22 2
21 1
2
dt
(3h)
(3i)
(4a)
(4b)
(5)
C is for concentrations, here taken as measured mixing
ratios in pptv, and S is for the source (here pptv yr\). We
have split the system of equations into the tropospheric
components (T) and the stratospheric components (S) in
Eq. (2). The elements of the S and C are written in Eq. (3)
(as the transpose of the vectors to save space). The two
subscripts are for the troposphere with the first for the
northern (n) and the southern (s) hemispheres and the
second for the polar (p), middle (m), and tropical (t)
latitudes. For the stratospheric concentrations the additional subscript ‘‘h’’ is used to designate ‘‘high atmosphere’’. The X matrix is split into four matrices as in
Eq. (2). The parts X , X , and X have only diagonal
21 12
11
elements as shown in Eq. (3d)—(3f); all other elements are
zero. ) is more complicated and is expressed in three
22
Eqs. 3g—3i. These three equations for ) represent the
22
diagonal, upper-diagonal and lower-diagonal parts written in this form to save space. The upper-diagonal represents the elements that go directly above the diagonal and
the lower-diagonal terms that go do directly below the
diagonal. All other elements of ) are zero. In these
22
equations (3), the gs represent inverse transport times
and lifetimes. g , g , and g are the inverse strato 1
1
spheric—tropospheric exchange times for the polar,
middle, and tropical latitudes. g , g , and g are the
2 2
2
inverse exchange times between polar to middle, middle
to tropics and tropic to tropic latitudes. The remaining
g are the inverse lifetimes where i"n or s and j"p,m,
GH
or t as before. For the inverse stratospheric lifetimes the
gs have the additional subscript ‘‘h’’. All elements of
) are in units of 1 yr\. N where i"p, m, and t are the
G
number of molecules of air in the polar, middle, and
tropical tropospheres; N represent the analogous numFG
bers for the stratosphere (The numbers of molecules of air
are: troposphere total"8.5; stratosphere total"2.1;
troposphere: polar"0.45, middle"1.5, tropical"2.3;
stratosphere : polar"0.26,
middle"0.43,
tropical
"0.36 in units of 10 molecules). Since we assume that
there is no stratospheric source of methyl chloride, the
stratospheric concentrations are fully determined by
the tropospheric measurements and the chemical and
transport parameters of the model. Equations for the
stratospheric concentrations can be solved exactly, for
each month, using the tropospheric concentration data
as in Eqs. (4a)—(4b); thus specific measurements of
stratospheric concentrations are not required. d is the
time step for the solution, here taken to be one month.
The source de-convolution is written in Eq. (5). Here C,
dC/dt are obtained from the measurements discussed
earlier and the elements of X are obtained from the
photochemical model for the lifetimes and from meteorological data for the exchange rates between the regions.
The structure of these models and their solutions are
discussed in detail by Khalil and Rasmussen (1984b). The
general formulation allows for a variety of resolutions as
they suit a particular application — here constrained by
the six locations where measurements were taken.
The transport parameters are estimated from existing
meteorological data and are consistent with similar models published earlier (Newell et al., 1972). The transport
times between the tropospheric boxes from polar to
middle, middle to tropical, and across the tropical latitudes are about 0.3, 0.2, and 0.4 yr, respectively; the times
for the exchange of tropospheric air with the stratosphere
is 8 yr on average. The lifetimes of methyl chloride in
each of the 6 tropospheric boxes are calculated for each
month of the year using the photochemical model described by Lu and Khalil (1991) and Moraes (1995).
Annually averaged stratospheric lifetimes are used (0.7,
1.8, 3.7 yr in tropical, middle and polar regions). The
calculated average stratospheric concentrations are in
good agreement with data from Fabian et al. (1996). The
annual average tropospheric concentrations of OH in
this model are 0.2, 0.7, 1.7, 1.4, 0.5 and 0.2;10 molecule cm\ in the regions represented by np, n, nt, st, s and
sp, respectively. The average OH in the model is 10
molecule cm\ in the troposphere.
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
In each tropospheric box the calculation of the lifetime
takes into account the surface temperature, temperature
lapse rate, and changes of air density with altitude as in
Khalil and Rasmussen (1984c). The mean tropopause
height is also adjusted for each latitudinal band (8, 12,
and 16 km from polar to equatorial). The vertical mixing
ratio of methyl chloride is assumed to be constant in the
troposphere based on the observations discussed earlier.
The average global tropospheric lifetime of methyl chloride turns out to be 1.4 yr due to reaction with OH. The
total atmospheric lifetime may be shorter if there are
other sinks, such as losses in the oceans.
The calculations of the emissions at various latitudes
are shown in Fig. 9. According to these calculations, the
observed concentrations imply a source of about
3.7 Tg yr\. There are practically no emissions in the
polar regions and moderate emissions in the middle
latitudes (0.33 and 0.11 Tg yr\ north and south respectively). Some 85% of the emissions occur in the tropics
and sub-tropics, which represent half the earth’s surface
area in this model (1.62 and 1.54 Tg yr\ north and
south, respectively). This is expected because the concentrations of methyl chloride are higher in the tropics on
average and so is the concentration of OH. Both these
factors require substantially higher emissions in the
tropics compared to other locations. Crutzen and Gidel
(1983) drew qualitatively similar conclusions from their
analysis. It seems that all the known sources, namely
oceans, fungal degradation of wood and biomass burning
are concentrated in the tropics, consistent with the results
of the source de-convolution and the observations of the
latitudinal distribution of methyl chloride.
1317
The year-to-year variations of the estimated global
emissions are between 3.5 and 3.9 Tg yr\. There is some
seasonality in the calculated emissions, with higher than
normal northern tropical emissions between October
and March, and somewhat higher than average middle
northern hemisphere emissions from February to May.
A major portion of the observed seasonal variation of
methyl chloride concentration is therefore explained by
the seasonal cycle of OH at all latitudes. In the polar
regions, which constitute only about 13% of the surface,
the seasonality of emissions may also be driven by seasonal differences of transport, which we have not taken
into account.
There are a number of uncertainties inherent in these
calculations and in all similar calculations. Uncertainties
exist is all three components — measurements, chemistry,
and transport. For the measurements there are potential
errors in absolute calibration. Since the estimated emissions are directly proportional to measured concentrations, calibration errors lead to proportionately similar
errors in the estimated emissions. We believe that these
errors, in the accuracy of measurements, are unlikely
to exceed 10%. Errors in the chemistry can be more
complex as both the absolute oxidizing capacity of the
atmosphere (here OH concentrations) and its distribution in space and time are uncertain. The absolute error
in OH leads to a proportionate error in the total
estimated emissions. This we believe to be small, as
the current model is consistent with the average OH
determined from methyl chloroform measurements
(Spivakovsky et al., 1990; Prinn et al., 1992). There are
some differences in the estimated OH at various latitudes
Fig. 9. Calculated emissions of methyl chloride at various latitudinal bands represented by the long-term measurements shown in Fig. 2.
Slow T refers to slowing down the transport across latitudes to half the average values, and Fast T refers to speeding up the transport to
twice the average values adopted in this study.
1318
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
among the models available to us and discussed in the
literature (Spivakovsky et al., 1990; Lu and Khalil, 1991).
These uncertainties cannot be quantified at present because there are no observational data on the climatology
of OH at various latitudes. The rate constant for the
reaction between methyl chloride and OH is also subject
to errors. Recent data have led to a substantial revision
(DeMore et al., 1997). The remaining uncertainty in the
temperature dependence of the rate constant would lead
to a range of emissions between 2.8 and 4.6 Tg yr\.
Finally, there are uncertainties in the transport parameters. These are both in the mean transport efficiency
and in its seasonal changes. Here only the mean values
are used. As the transport is speeded up, the emissions
become more sharply confined near regions of higher
concentrations. The opposite happens when the transport is slowed down. In the present case, doubling or
halving the transport times leads to relatively small changes in the estimated emission rates at various latitudes,
and does not affect the total source required to balance
the losses. The results of such calculations are included in
Fig. 9. This insensitivity to transport is partly due to the
moderately short lifetime of methyl chloride and to the
smallness of the latitudinal gradients, which determine
how much is transported from one latitude to another.
Another set of uncertainties in estimating the global
emissions relates to additional sinks that are known to
exist or may exist. Our estimates suggest that these are
small compared to reaction with OH, but may be significant when taken together and would therefore increase
the disparity between the total emissions needed to explain the atmospheric concentrations and the known
emissions from specific sources. Recent experiments have
shown that at latitudes approximately above 50° in each
hemisphere, the oceans become a net sink of methyl
chloride (Moore et al., 1996; Tait et al., 1994; Khalil and
Rasmussen, 1998a; Lobert, personal communication).
The net oceanic sink amounts to about 33 and
50 Gg yr\ for the northern and southern polar regions
(60—90°). These have been added to results in Table 1 (but
not included in the figures). Two additional sinks are
possible but at present we do not have enough data to
add their potential contributions into the budget. First,
some soils are known to take up methyl chloride. This
could as much as 0.2 Tg yr\, but this estimate is highly
uncertain because the extent of the soils that take up
methyl chloride is unknown (Khalil and Rasmussen,
1998c). Second, reactions with marine boundary layer
chlorine radicals could constitute yet another sink in the
atmosphere, which too may be more effective in the
tropical latitudes. Based on rough calculations, this sink
could be up to 0.4 Tg yr\ using 5;10 molecules of Cl
in the marine boundary layer (1 km) and negligible elsewhere (Singh et al., 1996; Rudolph et al., 1996; Keene
et al., 1996; Aucott, 1997).
5. Discussion
In this paper we have discussed the long-term behavior
of methyl chloride in the earth’s atmosphere. The sixteen-year global data set is archived for readers to use in
their own research. The salient points of the data are
summarized in Table 1. The main features of the data are
Table 1
The observed and calculated balance of methylchloride in the atmosphere based on data from 1981—1996
Northern hemisphere
C (pptv)
C (Tg)
Seasonal cycle (%)
Variability (%)
Lifetime by OH (yr)
Emissions (Tg yr\)
Southern hemisphere
Polar
Middle
Tropical
579.7
0.35
15
2.8
11.7
0.02
596.9
1.05
10
1.9
2.2
0.33
614.7
1.33
10
2.2
0.8
1.62
Tropical
620.0
1.35
!3
1.9
0.9
1.54
Middle
597.7
0.85
!6
2.5
3.6
0.11
Polar
572.6
0.35
!5
2.6
12.3
0.04
Global
606
5.3
3
2.5
1.4
3.7
Notes: (1) The concentrations are expressed in parts per trillion. The global average is latitudinally weighted. The concentration is
converted to the number of Tg of methylchloride in each region: Polar (60—90°), Middle (30—60°) and Tropical (0—30°). The Seasonal
Cycle is the difference of the average concentrations during Feb—Oct and Aug—Oct. It represents the cycle amplitude which is expressed
as the percent of the mean mixing ratio at the same latitude. The cycle amplitude in pptv can be calculated by multiplying the numbers
by C (pptv)/100%. (3) Variability is the standard deviation of monthly average concentrations as a percent of the mean concentration
C(pptv). The standard deviations are derived after subtracting seasonal cycles and trends from the data. It represents the residual
variability of the data. (4) The lifetime is calculated as the effective reaction rate constant times the average OH concentration at each
latitude. The OH concentration is determined from an independent photochemical model. (5) Emissions are estimated using a transport-chemistry model.
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
that the average concentration is about 600 pptv, there
are small uncertain trends, and there is more methyl
chloride in the tropical latitudes than at higher latitudes.
This is an unusual latitudinal distribution compared
to other moderately lived or long-lived gases, and
it is especially different from the characteristic latitudinal
signature of gases contributed by industrial and other
anthropogenic processes, which tend to result in high
concentrations in the middle northern latitudes. Since
reactions with OH are believed to be the dominant pathway for the removal of methyl chloride from the atmosphere, it implies that tropical emissions must be substantially more than emissions at higher latitudes as OH
is more concentrated in the tropics also. We find that
some 3.2 Tg yr\ may be emitted from the tropics and
most of the rest from the middle latitudes to make up
3.7 Tg yr\ of emissions.
The source deconvolution calculations do not provide
any direct information as to what the sources might be,
much less how much could come from each source. Such
estimates are based on additional and specific experiments and global extrapolations. The current budget of
methyl chloride based on emissions from the oceans,
fungi, biomass burning, and other sources will be discussed elsewhere. It is noteworthy that the current best
estimates of the known sources and their emission rates
do not readily account for the 3.7 Tg yr\ or more required from the calculations in this paper. This leaves
open the possibility of undiscovered sources, which could
include atmospheric formation from the chloride in the
marine boundary layer. The current source model includes this possibility as it does not distinguish between
surface emissions and atmospheric formation.
The data we have discussed are primary observational
information on the global cycle of methyl chloride. We
have used them to estimate the emissions required, and
we have discussed some of the uncertainties. The task of
apportioning the estimated emissions and making them
compatible with the latitudinal observations of methyl
chloride is the next step towards understanding the role
of human activities on the cycle of methyl chloride in the
environment and its role in affecting the ozone layer and
the earth’s climate.
Acknowledgements
We thank Martha Shearer at Portland State University, Don Stearns and Bob Dalluge at the Oregon Graduate Institute, Bill Keene at the University of Virginina
and Bill Sturges at the University of East Angila, for their
contributions to this work. We thank the staff at the
Cape Grim Baseline Station in Tasmania and NOAA/
GMCC, now NOAA/CMDL, for collecting samples
at the long-term sites. This project was supported in part
by grants from NASA (CNAG-1-589), NSF (CDPP-
1319
8821320), and the Department of Energy (CDOE DEFG06-84ER60613). Financial support for the analysis of
the data was provided by the Chemical Manufacturers
Association (Chlorine Chemistry Council) and the European Chemical Industry Council (Euro Chlor) as part of
the GEIA/RCEI program. Additional support was provided by the resources of the Biospherics Research Corporation and the Andarz Co.
How to obtain the data and what is archived. To
obtain the data write to: Carbon Dioxide Information
Analysis Center, Oak Ridge National Laboratory, P.O.
Box 2008, Oak Ridge TN 37831-6335. The data set
consists of the monthly average concentrations from the
7 primary sites (Fig. 2), monthly data from the 20 shortterm sites (Fig. 7), and vertical concentrations as in
Fig. 8.
References
Albritton, D. (Ed.) 1989. Journal of Atmospheric Chemistry,
Special Issue: Arctic Haze, 9, 1—397.
Andreae, M.O., 1991. Biomass burning: its history, use and
distribution and its impact on environmental quality and
global climate. In: Levine, J.S. (Ed.), Global Biomass Burning. MIT Press, Cambridge, MA.
Atlas, E., Pollock, W., Greenberg, J., Heidt, L., Thompson, A.M.,
1993. Alkyl nitrates, nonmethane hydrocarbons and
halocarbon gases over the equatorial Pacific Ocean
during SAGA 3. Journal of Geophysics Research 98,
16 933—16 947.
Aucott, M.L., 1997. Chlorine atoms and the global biogeochemical chlorine cycle. (Ph.D. Dissertation, Rutgers University,
New Brunswick, NJ).
Beck, S.M., Bendura, R.J., McDougal, D.S., Hoell Jr. J.M.,
Gregory, G.L., Curfman, Jr. H.J., Davis, D.D., Bradshaw, J.,
Rodgers, M.O., Wang, C.C., Davis, L.I., Campbell, M.J.,
Torres, A.L., Carroll, M.A., Ridley, B.A., Sachse, G.W., Hill,
G.F., Condon, E.P., Rasmussen, R.A., 1987. Operational
overview of NASA GTE/CITE 1. Airborne instrument intercomparisons: carbon monoxide, nitric oxide, and hydroxyl
instrumentation. Journal of Geophysics Research 92 (D2),
1977—1985.
Crutzen, P.J., Gidel, L.T., 1983. A two dimensional photochemical model of the atmosphere. 2: the tropospheric budgets of
the anthropogenic chlorocarbons, CO, CH , CH Cl and the
effect of various N sources on tropospheric ozone. Journal
of Geophysics Research 88, 6641—6661.
Davidson, C.I., Schnell, R.C., 1993. Introduction Atmospheric
Environment 27A, 2695—2699.
DeMore, W.B. et al., 1997. Chemical kenetics and photochemical data for use in stratospheric sodeling, Evalation C12
NASA, JPL, Pasadena, CA.
Edgerton, S.A., 1985. Gaseous tracers in receptor modeling:
methyl chloride emissions from wood combustion. Ph.D.
dissertation, Oregon Graduate Center, Beaverton.
Fabian, P., Borchers, R., Leifer, R., Subbaraya, B.H., Lal, S.,
Roy, M., 1996. Global stratospheric distribution of halocarbons. Atmospheric Environment 30, 1787—1796.
1320
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
Graedel, T.E., Keene, W.C., 1995. Tropospheric budget of reactive chlorine. Global Biogeochemical Cycles 9, 47—77.
Graedel, T.E., Keene, W.C., 1996. The budget and cycle of
Earth’s natural chlorine. Pure and Applied Chemistry 68,
1689—1697.
Harper, D.B., 1985. Halomethane from halide ion — a highly
efficient fungal conversion of environmental significance.
Nature 315, 55—57.
Harriss, R.C., Wofsy, S.C., Garstang, M., Browell, E.V., Molion,
L.C.B., McNeal, R.J., Hoell, Jr., J.H., Bendura, R.I., Beck,
S.M., Navarro, R.L., Riley, J.T., Snell, R.L., (1988). The
Amazon boundary layer experiment (ABLE 2A): dry season
1985. Journal of Geophysics Research 93 (D2), 1351—1360.
Harriss, R.C. et al., 1990. The Amazon boundary layer experiment: wet season 1987. Journal of Geophysics Research 95,
16 721—16 736.
Keene, W.C., Jacob, D.J., Fan, S.-M., 1996. Reactive chlorine:
a potential sink for dimethylsulfide and hydrocarbons in the
marine boundary layer. Atmospheric Environment 30, i-iii.
Khalil, M.A.K., Rasmussen, R.A., 1984a. Statistical analysis of
trace gases in arctic haze. Geophysics Research Letters 11,
437—440.
Khalil, M.A.K., Rasmussen, R.A., 1984b. Modeling chemical
transport and mass balances in the atmosphere. Environmental Exposure From Chemicals, Vol. II, Chapter 2. CRC
Press, Boca Raton, Florida.
Khalil, M.A.K., Rasmussen, R.A., 1984c. The atmospheric lifetime of methylchloroform. Tellus, 36B, 317—332.
Khalil, M.A.K., Rasmussen, R.A., 1988. Trace gases over the
western Atlantic Ocean: fluxes from the eastern United
States and distributions in and above the planetary boundary layer. Global Biogeochemical Cycles 2(1), 63—71.
Khalil, M.A.K., Rasmussen, R.A., 1990. Atmospheric methane:
recent global trends. Environmental Science and Technology
24, 549—553.
Khalil, M.A.K., Rasmussen, R.A., 1998a. Ocean-Air exchange of
atmospheric trace gases. Report No. 01-1097, Department of
Physics, Portland State University, Portland, Gregon, submitted.
Khalil, M.A.K., Rasmussen, R.A., 1998b. Atmospheric chloroform, submitted.
Khalil, M.A.K., Rasmussen, R.A., 1998c. Soil—Atmosphere exchanges of radiatively and chemically active gases, submitted.
Koppmann, R., Johnen, F.J., Plass-Dulmer, C., Rudolph, J.,
1993. Distribution of methyl chloride, dichloromethane,
trichloroethene and tetrachlorethene over the north and
south Atlantic. Journal of Geophysics Research 98,
20 517—20 526.
Krol, M., van Leeuwen, P.J., Leilieveld, J., 1998. Global OH
trend inferred from methylchloroform measurements. Journal of Geophysics Research 103, 10 697—10 712.
Lovelock, J.E., 1975. Natural halocarbons in the air and the sea.
Nature 256, 193—194.
Lu, Y., Khalil, M.A.K., 1991. Tropospheric OH: model calculations of spatial, temporal and secular variations. Chemosphere 23, 397—444.
Moore, R.M., Groszko, W., Niven, S.J., 1996. Ocean-atmosphere exchange of methyl chloride: results from NW Atlantic and Pacific ocean studies. Journal of Geophysics Research 101, 28 529—28 538.
Moraes, F., 1995. Model calculations of atmospheric OH.
Ph.D. dissertation, Oregon Graduate Institute, Portland,
Oregon.
Newell, R.E., Kidson, J.W., Vincent, D.G., Boer, G.J., 1972.
The General Circulation of the Tropical Atmosphere
and Interactions with Extratropical Latitudes. MIT, Cambridge, Massachusetts.
Palmer, T.Y., 1976. Combustion sources of atmospheric chlorine. Nature 263, 44—45.
Penkett, S.A., Derwent, R.G., Fabian, P., Borchers, R., Schmidt,
U., 1980. Methyl chloride in the stratosphere. Nature 283,
58—60.
Pierotti, D., Rasmussen, R.A., Dalluge, R., 1980. Measurements of N O, CF Cl , CFCl , CH CCl , CCl , and
CH Cl in the troposphere and lower stratosphere over
North America. Journal of Geomagnetism Geoelectricty, 32,
181—205.
Prinn, R., Cunnols, D., Simmonds, P., Alyea, F., Boldi, R.,
Crawford, A., Fraser, P., Gutzler, D., Hartley, D., Rosen,
R., Rasmussen, R., 1992. Global average concentration
and trend for hydroxyl radicals deduced from ALE/GAGE
trichloroethane (methyl chloroform) data for 1978—1990.
Journal of Geophysical Research 97, 2445—2461.
Rasmussen, R.A., Khalil, M.A.K., 1980. Atmospheric halocarbons: measurements and analyses of selected trace gases. In:
Aikin, A.C. (Ed.), Proceedings, NATO Advanced Study Institute on Atmospheric Ozone. Department of Transportation, Washington D.C.
Rasmussen, R.A., Rasmussen, L.E., Khalil, M.A.K., Dalluge,
R.W., 1980. Concentration distribution of methyl chloride
in the atmosphere. Journal of Geophysics Research 85,
7350—7356.
Reinhardt, T.E., Ward, D.E., 1995. Factors affecting methyl
chloride emissions from forest combustion. Environmental
Science and Technology 29, 825—832.
Rudolph, J., Khedim, A., Koppmann, R., 1995. Field study of the
emissions of methyl chloride and other halocarbons from
biomass burning in Western Africa. Journal of Atmospheric
Chemistry 22, 67—80.
Rudolph, J., Koppermann, R., Plass-Dulmer, C., 1996. The
budgets of ethane and tetrachloroethene: is there evidence
for the impact of reactions with chlorine atoms in the troposphere. Atmospheric Environment 30, 1887—1894.
Schauffler, S.M., Heidt, S.M., Pollock, W.H., Gilpin, T.M., Vedder, J.F., Solomon, S., Lueb, R.A., Atlas, E.L., 1993. Measurements of halogenated organic compounds near the tropical
tropopause. Geophysics Research Letters 20, 2567—2570.
Singh, H.B., Salas, L., Shigeishi, H., Scribner, E., 1979. Atmospheric halocarbons, hydrocarbons and sulfur hexafluoride: global distributions, sources and sinks. Science 203,
899—903.
Singh, H.B., Salas, L., Stiles, R.E., 1983. Methylhalides in and
over the Eastern Pacific (40°N-32°S). Journal of Geophysics
Research 88, 3684—3690.
Singh, H.B., Thakur, A.N., Chen, Y.E., Kanakidou, M., 1996.
Tetrechloroethylene as an indicator of low Cl atom concentrations in the troposphere. Geophysics Research Letters 23,
1529—1532.
Spivakovsky, C.M., Yevich, R., Logan, J.A., Wofsy, S.C.,
McElroy, M.B., 1990. Tropospheric OH in a three dimensional chemical tracer model: an assessment based on
M.A.K. Khalil, R.A. Rasmussen / Atmospheric Environment 33 (1999) 1305—1321
observations of CH CCl . Journal of Geophysics Research
85, 18 441—18 471, 1990.
Tait, V.K., Moore, R.M., Tokarczyk, R., 1994. Measurements of
methyl chloride in the northwest Atlantic. Journal of Geophysics Research 99, 7821—7833.
1321
White, R.H., 1982. Biosynthesis of methyl chloride in the fungus
Phellinus ponaceus. Archives of Microbiology 132, 100—102.
WMO, 1989, 1994. Scientific Assessment of Stratospheric
Ozone: 1989 (two volumes), 1994. GORAMP report C20
and C37, WMO, Geneva.