JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 105, NO. D15, PAGES 19,771-19,778, AUGUST 16, 2000
Central role of carbonyl compounds in atmospheric
chemistry
D. J. Lary
Department of Geophysics and Planetary Sciences, Tel Aviv University, Tel Aviv, Israel
D. E. Shallcross 2
Centre for Atmospheric Science, Cambridge University, Cambridge, England, United Kingdom
Abstract. With the exception of acetone it is not generally recognized how
important atmospheric carbonyls and alkyl radicals are in the lower stratosphere
and upper troposphere. Carbonyl compounds are the crucial intermediate species
for the autocatalytic production of OH. For example, in the upper troposphere and
lower stratosphere it is calculated based on data assimilation analysis of Atmospheric
Trace Molecule Spectroscopy Experiment (ATMOS) data that CH3 production due
to the degradation of carbonyls contributes around 40% to the overall production
of CH3, a key initiation step for HOx production, with the contribution due to the
photolysis of CH3CHO being comparable to that of acetone. So correctly modeling
the alkyl radical concentrations is of central importance and has not be given the
attention it deserves to date. The reactions of carbonyls with Br and C1 are also
major sources of HBr and HC1. In short, carbonyl compounds play a central role
in atmospheric chemistry close to the tropopause, and this is directly relevant to
issues such as the assessment of the impact of air traffic, and ozone depletion.
1. Introduction
The recent World Meteorological Organization [1998]
(WMO) report report concludes that our understand-
ing of tropospheric OH is incomplete, especially with
regard to sources of upper tropospheric OH and the
chemistry of polluted conditions. This is particularly
significant when we recall that HOx cycles are the dom-
inant catalytic ozone loss cycles in midlatitudes below
20 km [Lary, 1997; WMO, 1998].
Aldehydes and ketones are key intermediates in at-
mospheric chemistry. In the atmosphere they are pro-
duced by the oxidation of hydrocarbons, and some are
also emitted directly (e.g., acetone). Because they un-
dergo a wide variety of reactions, both chemical and
photolytic, they play a major role in many atmospheric
processes. Notably, the >C=O group is a chromophore
allowing absorption to occur at long wavelengths. The
reactivity of these compounds arises largely through two
features of their structures: the polarity of the carbonyl
1Also at the Centre for Atmospheric Science, Cambridge Uni-
versity, Cambridge, England, United Kingdom.
2Now at the School of Chemistry, University of Bristol, Bristol,
England, United Kingdom.
Copyright 2000 by the American Geophysical Union.
Paper number 1999JD901184.
0148-0227 / 00 / 1999 JD901184509.00
group and the acidity of any alpha hydrogens that are
present.
The sections which follow review the major roles of
atmospheric carbonyls. However, before examining the
atmospheric chemistry of carbonyls in more detail, let
us examine the credentials of the photochemical model.
2. Model Calculation
The numerical model used in this study is the exten-
sively validated AutoChem model [Lary et al., 1995b;
Lary, 1996; Fisher and Lary, 1995; Wang et al., 2000].
The model is explicit and uses the adaptive-timestep,
error monitoring, [Stoer and Bulirsch, 1980] time inte-
gration scheme designed by Press et al. [1992] for stiff
systems of equations.
P hotolysis rates are calculated using full spherical ge-
ometry and multiple scattering as described by Lary and
Pyle [1991a, b] after Meier et al. [1982]; Nicolet et al.
[1982] with the treatment of spherical geometry after
Anderson [1983]. The photolysis rate used for each time
step is obtained by 10 point Gaussian-Legendre integra-
tion after Press et al. [1992]. The time step usually used
is 15 min.
AutoChem uses kinetic data largely based on DeMote
et al. [1997] and Atkinson et al. [1997], with the recent
NO2 and HNO3 kinetics of Donahue et al. [1997], Fulle
et al. [1998], Brown et al. [1999a], and Portmann et al.
[1999] evaluated by Brown et al. [1999b], and the recent
OH+C10 kinetics of Kegley Owen et al. [1999].
19,771

19,772 LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS
In addition, it was the first model to ever have the fa-
cility to perform 4D variational data assimilation (4D-
VAR) [Fish½l' a71d La'y, 1995]. This is a \"state of the
art\" technique used extensively in meteorology. The
use of data assimilation in atmospheric chemistry has
been reviewed by Lary [1999]. It allows us to bring
together within a mathematical framework our obser-
vational and theoretical knowledge together with their
associated uncertainties. In a least squares sense it pro-
vides the best fit simulation of the model to any avail-
able observations within the assimilation period. In this
case it was a 1-day period of data from the space shuttle
borne Atmospheric Trace Molecule Spectroscopy Ex-
periment (ATMOS) instrument [Abrams et al., 1996;
Newchurch et al., 1996]. Observations of 16 species mea-
sured by ATMOS and the Upper Atmosphere Research
Satellite (UARS) were simultaneously used, namely:
03, NO, NO2, N205, HNO3, HObNOb, HCN, C10,
C1ONO2, N20, CO, CO2, CH4, C2H6, and H20, to-
gether with Stratospheric Aerosol and Gas Experiment
(SAGE) aerosol observations. The model is therefore
highly constrained, and so any deficiency in our ki-
netic description is readily highlighted. As this \"state
of the art\" technique is not yet widely used in atmo-
spheric chemistry it is valuable to mention the works
by Menke [1984], Courtier and Talagrand [1987], Cohn
[1997], Courtier et al. [19931, Fisher and Lary [1995],
Khattatov et al. [1999], and Lary [1999].
In this study no description of atmospheric trans-
port or convection has been used. Instead the 4D-VAR
data assimilation was performed using a set of stacked,
independent, boxes. The boxes were stacked in the
equivalent-PV latitude theta flow tracking coordinate
system [Lary et al., 1995a] at an equivalent PV latitude
of 40øS. 40øS was chosen as we used data from the STS-
45/ATLAS 1 mission which was launched on March 24,
1993 from the Kennedy Space Center. During its 8
days of operation, the ATMOS instrument made obser-
vations spanning a substantial portion of the globe. The
53 measurements taken at orbital sunrise covered the
midlatitude and equatorial regions of the Earth from
30øS to 30øN. The 41 sunset observations were made at
25øS to 55øS. For the duration of ATLAS-1 the equiva-
lent PV latitude for which the vertical profiles covered
the largest range of altitudes, and for which the largest
number of species were observed, was centered on about
40øS. The assimilation window used was 1-day so that
we can have one complete diurnal cycle.
2.1 Validation
Since the concentrations of the hydrocarbons used is
such a crucial factor in the reliability of the analysis
presented here, Figure 1 shows a comparison of the
analyses produced with data assimilation for C2H6 and
CH4 and the ATMOS observations. We can see that
in both cases the model falls well within the observed
error bars. Space does not here permit to give further
10
l 40-
::3 50-
(/) 60-
t._ 70-
13_ 80-
100
1E-012 1E-011 1E-010 1E-009 1E-008 1E-007 1E-006 1E-005
,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,,I , ,, .... ,
Data Assmlabon C2H6 ß
, ATMOS 0bservabons C2H6
.......... Data Assimilation CH4 
O ATMOS observations CH4
20-  -20
30- -eq -30
ee
.......  ....... '' '?'\"\"I .......  .......  ' \"\"\" '41\"' ....200
1E-012 1E-011 1E-010 1E-009 1E-008 1E-007 1E-006 1E-005
Concentration (v.m.r.)
200
lO
.-40
-50
-60
-70
-80
'2ø 1oo
Figure 1. A comparison of the observed C2H and
CH4 concentrations obtained using data assimilation
with the AutoChem model and the observations of AT-
MOS for 40øS at 17.45 local solar time on March 29,
1993. Observations of 15 species measured by ATMOS
and UARS were simultaneously used, namely: 03, NO,
NO2, N205, HNO3, HO2NO2, HCN, C10, C1ONO2,
N20, CO, C02, CH4, C2H6, and H20.
validation of the model. However, Lary and Shallcross
[in press, 2000] showed comparison with the observed
C1ONO2/HC1 and NOx/HN03 ratios and HBr obtained
using data assimilation with the AutoChem model and
the observations of ATMOS and other instruments and
showed that in both cases the model falls well within
the observed error bars. Likewise, later on in this paper
a detailed comparison is given with insitu aircraft data
from the Stratospheric Photochemistry Aerosol and Dy-
namics Experiment (SPADE).
2.2 Choice of Observations
Care needs to be taken when performing model data
comparisons that like is compared with like. So in this
study very careful data selection was performed to en-
sure that it really is suitable to use the observations we
have used in our analysis. For example, atmospheric
propane and acetone have been measured and would
have been useful, but as we use very strict selection cri-
teria in a flow tracking coordinate system (equivalent
PV latitude and theta) and local solar time, unfortu-
nately none were available for our analysis period. The
two data sets we have used (ATMOS and SPADE) in-
clude simultaneous observations of many species. To
the best of our knowledge, there are no simultaneous
measurements of NMHC or carbonyls; that is the only
reason that they were not used. We would, of course,
rather that such observations were available to us, as
then we could better constrain the calculations. We
have used such rigorous selection criteria as it is vital

LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS 19,773
for us to compare like with like, otherwise the compar-
ison is rather worthless.
2.3 A Unique Feature of Data Assimilation
It is worth noting something about the technique of
data assimilation. It gives us indirect information on
all the species within the chemical scheme via our theo-
retical model. This is a unique and powerful feature of
this type of analysis. It is not just a simple model obser-
vation comparison, it is an optimized fit of the model to
all of the observations. This unobserved specie informa-
tion may of course may be incomplete if our model is in-
complete, but as we can see from tile SPADE study the
shapes of diurnal cycles and the partitioning of species
contains information on other, nonobserved, species.
3. Hydrocarbon Oxidation and the
Possibility of Autocatalysis
The Earth's atmosphere is strongly oxidizing, and so
hydrocarbons released into the atmosphere are oxidized.
This eventually leads to the production of CO,2. Fig-
ure 2 shows schematically the oxidation of methane,
ethane, and propane.
Tile oxidation is usually initiated by the abstraction
of a hydrogen atom from the hydrocarbon by reaction
with OH, O(1D), or C1. Although for much of the tro-
posphere the abstrm'tion by OH is the most important,
in the upper troposphere and lower stratosphere the ab-
straction by C1 becnnes significant, even dominant. In
regions where we have had chlorine activation on cold
str'faces, then abstraction by C1 also dominates.
The alkyl radicals (CH3, C2Hs, or C3H7 generically
refereed to as R) formed by hydrogen abstraction will
immediately add to 02 to form a peroxy radical (RO2).
Peroxy radicals formed in the lower stratosphere or
upper troposphere will react witt NO, C10, or BrO,
thereby coupling RO, with N() C10, and BrOw, to
yield an alkoxy radical RO. Including these RO2 re-
actions can significantly alter the OH/HO2, NO/NO2,
C1/C10, and Br/BrO ratios. Fox' tile alkoxy radicals
investigated here, reaction with O2 will dominate. For
CHaO, C2HsO and n-CaH70, HO2 and RCHO will be
fbrmed. Where RO is i-Call70 acetone and HO2 will
be formed. The aldehydes will be photolyzed. Primar-
ily producing HCO radicals, which on reaction with O2
generate CO and HO2, and another alkyl radical, which
can feed back into the cycle depicted in Figure 2, giving
more OH.
3.1 Importance of Alkyl, R, Radicals
It should be ernphasised that the alkyl radicals, R, are
the first step in hydrocarbon oxidation once a hydrogen
has been abstracted from the alkanes. So their forma-
t, ion represents one of the key initiation steps of H O
R P
CHa H
C2Hr, CH3
c3a? C2H5 NO HOBr
)2
Figure 2. A schematic representation of nethane,
ethane, and propane oxidation.
production. As the diagram in Figure 2 shows, they are
also the radicals forrned in the autocatalysis where the
carbonyls are fragrnented, another key issue for HO
production. This autocatalysis becomes a more signifi-
cant source of CHa as you go down the series of alkanes
from C2 to C3, etc. So correctly modeling the alkyl
radical concentrations is one of central importance, and
has not be given the attention it deserves to date.
3.2 Source of R Radicals and Hence OH
What is seldom highlighted is the fact that in the up-
per troposphere and lower stratosphere the source of

19,774 LARY AND SHALLCROSS- CENTRAL ROLE OF CARBONYL COMPOUNDS
0 20 40
i
60 100
, I  10
0 2O 4O 6O
% Contribution to CH 3 production
0 20 40 60 80 100
I  I , I , 10   ß  r I  I
I Lor_.,l Sol Noor / I CI+CH4
14oøS / ] / I  C,+HCHO
I /  / I J
I I / ! I I CI+C2H50H
20/ | ' I   cC: ++(cCHH33 )020 c '
0 20 40 60 80
% Contribution to HCI production
80
I ,
 CI+CH4
OH+CH4
(CH3)2CO+hv
CH3CHO+hv
CH3CO3+NO
CH3CO3+CH300 -20
CH3CO3+CH3CO3
-30
, 200
80 100
-40
,.50
-60
-70
-80
.9o 1 O0
10
-20
-30
.40
-50
-60
-70
2ø 1 O0
, 200
loo
(b)
0 20 40 60 80
10  ,,, ,  , !  !
40 ø OH+C2H6
0(1D)+C2H6
=,,. C2H5CHO+hv
20
3O
4O
o '%
,o o
o 1009
,/
oo I I i  i I ; ,
0 20 40 60 80
% Contribution to C2H5 production
100
(d)
-2O
-3O
lO
-40
-50
-60
-70
-80
2ø1 O0
200
lOO
o 2o 4o 6o 8o lOO
Loc;11 Sol Noor Br+H02
40o,  BrO+OH
 Br+CH300
 Br+HCHO
A Br+CH3C(O)CHO
3- .7 Br+CH3CHO
/ ':
1 O0 9 ,- ! iiiii ..... --
200 '\" l 200
0 20 40 60 80 100
% Contribution to HBr production
Figure 3. The calculated percentage contribution for local solar noon on March 29, 1993, at
40øS of the various reactions which produce (a) CH3 (the background shading shows the total
contribution to CH3 production from carbonyls), (b) C2H5, (c) HC1 (the background shading
shows the contribution to HC1 production from organic species not in most stratospheric models),
and (d) HBr (the background shading shows the total contribution to HBr production from
carbonyls). Obtained using data assimilation with the AutoChem model of observations of 15
species measured by ATMOS and UARS, namely: 03, NO, NO2, N205, HNO3, HONO,, HCN,
C10, C1ONO2, NO, CO, CO2, CH4, CH6, and HO.
10
lOO
our R radicals (i.e. CH3, C2H5 etc.) from the de-
composition of the carbonyls is often of comparable
magnitude, and sometimes greater than that due to
the initial hydrocarbon hydrogen abstraction by OH,
O(1D), or C1 (Figure 3). For example, we have at least
three sources of CH3 from carbonyl compounds, acetone
((CH3)CO) photolysis [Gierczak et al., 1998; McKeen
et al., 1997; Jaegl et al., 1998], ethanal (CH3CHO) pho-
tolysis [Friedl et al., 1997], and the reaction of NO with
CH3CO. Consequently, we have an effective autocat-
alytic production of OH with the carbonyl compounds
being the crucial intermediate species, and with auto-
catalysis being the catalysis of a reaction by one of its
products.
Figure 3 shows the calculated percentage contribu-
tion for local solar noon on March 29, 1992, at 40øS of
the reactions which produce CH (Figure 3a) and C2H5
(Figure 3b). Figure 3a shows that in the upper tropo-
sphere and lower stratosphere one of most important
sources of CH3 is the photolysis of CH3CHO which is
comparable to that due to the photolysis of acetone.
Acetone is a major source of HOx. For example,
Muller and Brasseur [1999] recently reported model cal-
culations that suggest hat acetone photooxidation rep-

LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS 19,775
resents a large, almost ubiquitous source of HOx in the
upper troposphere (around 20-40% of the total primary
source in the main aircraft corridors, poleward of 40øN),
hile the convective injections of peroxides and alde-
hydes are the dominant sources in the tropics, above
the oceans and the continents, respectively. Muller and
Brasscur [1999] calculated that the presence of acetone
might enhance by about 20% the sensitivity of upper
tropospheric ozone to the current aircraft emissions of
In the case of C2H5, Figure 3b shows that hydrogen
abstraction by C1 is an important source of C2H5. This
ga; .... h; +he; .... + .... of hytogn ¬r-
tion by halogens.
Figure 4 shows the effect of a comparison between
observations and calculations obtained when the car-
bonyl sources of HO are included using the technique
of 4D-Var data assimilation. Data assimilation allows
us to simultaneously use all the observations together
with our numerical model, to give us the best fit model
simulation to the observations in a least squares sense.
In some cases there is an improvement when the car-
bonyl sources of HO are included (namely, C10 and
HC1), and in others there is a slight worsening of the fit
(morning OH). The comparison is not conclusive.
The various curves correspond to progressively more
complete chemical schemes. The solid line is a full
methane oxidation scheme with 59 species and 366 re-
actions, the dashed curve is for a full methane-ethane
oxidation scheme with 74 species and 431 reactions, the
dot-dashed curve is for a full methane-ethane-propane
oxidation scheme with 89 species and 492 reactions,
the dot-dot-dashed curve is for a full methane-ethane-
propane-ethene oxidation scheme with 107 species and
566 reactions. The overlying diamonds with error bars
are the SPADE observations.
So two points are worth emphasizing. First, in the
upper troposphere and lower stratosphere the reaction
of C1 with C2H6 is much more important than the re-
action of OH with C2H6. Second, the reaction of C1
with C2H6 producing the reservoir H C1 offsets the re-
action of C1 with 03. So we have additional pathways
for reducing the effectiveness of catalytic ozone loss by
chlorine.
3.3 Interaction With Organic Aerosols
D. J. Lary et al. (The potential role of carbonyl pro-
duction on organic aerosols, submitted to Journal of
Geophysical Research, 1999) consider the possibility of
atmospheric organic aerosols interacting with O3 lead-
ing to the formation of aldehydes such as HCHO and
CHgCHO. They conclude that if organic aerosols do re-
lease carbonyls on reaction with 03, then one of the
mst significant effects is likely to be the formation of
hydrogen halides. It is likely that this will play a role if
the product of the reaction probability and the organic
surface area exceeds 2x10 -4 ]zm 2 cm -3 If the teac-
tion probabilities of [deGouw and Lovejoy, 1998] apply
to these processes, then this corresponds to an organic
aerosol surface area in the range 0.2 to 2/m 2 cm -a.
The measurements of Novakov and Penner [1993] im-
ply an organic surface area of the order of 3/m  cm-a;
it is therefore possible that there is a role for these pro-
cesses in the atmosphere. The next most significant ef-
fect is the autocatalytic production of OH which in turn
will reduce the HNOa/NOy ratio. It is likely that this
will play a role if the product of the reaction probability
and the organic surface area exceeds 9x10 -4/m 2 cm -a.
Fruekilde et al. [1998] have also observed the formation
of carbonyls from_ the ozono!ysis of a leaf surface.
Recently Murphy et al. [19981 have reported that a
very high fraction of tropospheric aerosols contain or-
ganic molecules; some presumably reside on the sur-
face of the aerosols. Murphy et al. [1998] made in situ
measurements of the chemical composition of individual
aerosol particles at altitudes between 5 and 19 kin. The
measurements reveal that upper tropospheric aerosols
often contained more organic material than sulfate. No-
vakov and Penner [1993], Sheridan et al. I1994], and
Novakov et al. [1997] collected atmospheric particles in
the mid-latitude upper troposphere and lower strato-
sphere, and they found non sulfate materials including
carbon rich substances. Other measurements of organic
aerosols have also been made, [e.g., Middlebrook et al.,
1998; Shevchenko et al., 1999; Kavouras et al., 1998;
Vasconcellos et al., 1998; Chen et al., 19971.
4. The Production of Hydrogen Halides
Several observational and modeling studies have also
highlighted the inability of the models to reproduce the
observed HBr profile. These include the work of Nolt
et al. [1997], Chipperfield et al. [19971, Johnson et al.
[1995], and Carlotti et al. [1995]. The recent WMO
[1998] report concludes that 30 km models usually over-
predict the C10/HC1 and HOC1/HC1 ratios.
Figure 3c shows that in the troposphere it is calcu-
lated that the main source of HC1 is hydrogen abstrac-
tion by C1 from C2H6, CH4, and C2H5OH. As we ap-
proach the tropopause and then enter the lower strato-
sphere, it is calculated that the reaction of C1 with
CFHsOH becomes the most important. The role of car-
bonyl compounds is also important in the case of HBr
(Figure 3d).
5. Summary
Carbonyl compounds are crucial intermediate species
for the autocatalytic production of OH. It is calculated
that at around 20 km the CH3 production due to the
degradation of carbonyls is almost twice as much as that
due to the total direct H abstraction by OH, O(D), and
C1. This is particularly significant when we recall that

19,776 LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS
(a)
0 6 12 18 24
1.6E-012 .... 'Mh'h ' ' ' I ..... I,,,,, I 1.6E-012
e ane c eme Methane-Ethane*Propane Scheme Methane*Ethane-Propa he-Ethane Scheme , SPADE Observations 1.2E-012 1.2E-012
.
8E-013 - - 8E-013
.
4E-013 - - 4E-013
.
0 6 12 18 24
Local Solar Time (hours)
(c)
0 6 12 18 24
2E-006 , , , , , I , , , , , I , , , , , I , ,  , , 2E-006
1.8E-006 - 1.8E-006
- 1o6E-006 - 1.6E-006
\" 1.4E-006 - 1,4E-006
o
.... E.n. iii..; Sh
' sM; ; E ,;l,sh; rv a ,iPo n  E.... 03
1E-006 ..... I ..... I ..... I ..... 1E-006
0 6 12 18 24
Local Solar Time (hours)
(e)
0 6 12 18 24
6E-011 ..... I..... I ..... I '('l, 6E-011 Methane Scheme Methane-Ethane Scheme
Methen e- Et bane- Propane Scheme
Methane-Et bane- Propee-Eth erie Scheme
 SPADE Observations
 4E-011 4E-011
2E-011
- 2E-011
6 12 18
Local Solar Time (hours)
(b)
0 6 12 18 24
Methane Scheme
8E-012 , , M..,-Et,-o-  2
 SPADE Observeboris
.
\"\" 6E-012
.
v
( 4E-012 -
 .
2E-012 --
0 0
0 24
(d)
1.2E-009
\"-7'. 8E-010 -
E
 -
O
Z 4E-010
0
0
8 E-012
r
- 6E-012
- 4E-012
- 2E-012
i ! ! ! ! ! ! ii ! ! ! ! ! ! ! ! !
(f)
o
2E-009
1.6E-009 -
 .2E-009-
v
f 8E-010 -
.
4E-010 -
!
6 12 18
Local Solar Time (hours)
Methane Scheme
Methane-Ethane Scheme
Methane-Ethane-Propane Scheme
Metha ne-Etha ne-Propa ne-Ethene Scheme
6 12 18
Local Solar Time (hours)
24
1.2E-009
- 8E-010
- 4E-010
2
o
24
o o o o
o 24 o 24
ElIHIll I III IIIL ,,, ............
Methane Scheme
..... Methane-Ethane Scheme
Methane-Ethane-Propane Scheme
, SPADE Observations
'''''1'''''1'''''1'''''
6 12 18
Local Solar Time (hours)
- 1.6E-009
- 1.2E-009
- 8E-010
- 4E-010
Figure 4. 4D-Var analysis of SPADE data taken from the ER-2. The various curves correspond
to progressively more complete chemical schemes. The solid line is a full methane oxidation
scheme with 59 species and 366 reactions, the dashed curve is for-a full methane-ethane oxidation
scheme with 74 species and 431 reactions, the dot-dashed curve is for a full methane-ethane-
propane oxidation scheme with 89 species and 492 reactions, and the dot-dot-dashed curve is for
a full rnethane-ethane-propane-ethene oxidation scheme with 107 species and 566 reactions. The
overlying diamonds with error bars are the SPADE observations. (a)'OH, (b) HO2, (c) 03, (d)
NO2, (e) ClO, and (f) HC1.

LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS 19,777
HOx cycles are the dominant catalytic ozone loss cy-
cles in midlatitudes below 20 km [Lary, 1997; WMO,
1998]. In addition, at around 10 km the photolysis of
C2HsCHO contributes approximately 35% to the pro-
duction of C2H5. Both CH3 and C2H5 are at the start
of the CH4 and C2H6 oxidation chains.
The reaction of Br with carbonyls is the dominant
source of HBr in the troposphere and lower strato-
sphere. It is therefore not surprising that many models
have not been able to reproduce the observed HBr pro-
file as these halogen carbonyl interactions are often not
included in the model calculations [Nolt et al., 1997;
,w,je et al, ' ro   ((' Car!oti
et al., 1995]. In short, carbonyl compounds play a cen-
tral role in atmospheric chemistry.
Acknowledgments. It is a pleasure to acknowledge
the following: The government of Israel for an Alon Fellow-
ship; the Royal Society for a Royal Society University Re-
search Fellowship; the NERC and EU for research support;
and anonymous reviewers for their constructive criticisms
and comments.
References
Abrams, M., M. Gunson, A. Chang, C. Rinsland, and
R. Zander, Remote sensing of the earth's atmosphere from
space with highresolution Fourier-transform spectroscopy:
Development and methodology of data processing for the
ATMOS experiment, Appl. Opt., 35, 2774-2786, 1996.
Anderson, D., The troposphere-stratosphere radiation-field
at twilight - A spherical model, Planet. Space Sci., 31,
1517-1523, 1983.
Atkinson, R., D. Baulch, R. Cox, R. Hampson, J. Kerr,
M. Rossi, and J. Troe, Evaluated kinetic and photochem-
ical data for atmospheric chemistry: Supplement VI., IU-
PAC subcommittee on gas kinetic data evaluation for
atmospheric chemistry, J. Phys. Chem. Ref. Data, 26,
1329-1499, 1997.
Brown, S., R. Talukdar, and A. Ravishankara, Rate con-
stants for the reaction OH+NO=--HNOa+M under at-
mospheric conditions, Chem. Phys. Left., 299, 277-284,
1999a.
Brown, S., R. Talukdar, and A. Ravishankara, Reconsider-
ation of the rate constant of the reaction of hydroxyl rad-
icals with nitric acid, J. Phys. Chem., 103, 3031-3037,
1999b.
Carlotti, M., et al., Measurement of stratospheric HBr using
high-resolution far-infrared spectroscopy, Geophys. Res.
Left., 22, 3207-3210, 1995.
Chen, S., S. Liao, W. Jian, and C. Lin, Particle size distri-
bution of aerosol carbons in ambient air, Environ. Int.,
23,475-488, 1997.
Chipperfield, M., D. Shallcross, and D. Lary, A model study
of the potential role of the reaction BrO+OH in the pro-
duction of stratospheric HBr, Geophys. Res. Left., 2,
3025-3028, 1997.
Cohn, S., An introduction to estimation theory, J. Met. Soc.
Japan, 75,257-288, 1997.
Courtier, P., and O. Talagrand, Variational assimilation
of meteorological observations with the adjoint vorticity
equation., 2. Numerical results, Q. J. R. Meteorol. Soc.,
113, 1329-1347, 1987.
Courtier, P., J. Derher, R. Errleo, J. Louis, and T. Vukice-
vic, Important literature on the use of adjoint, variational-
methods and the Kalman filter in meteorology, Tellus,
Set. A., 45A, 342-357, 1993.
deGouw, J., and E. Lovejoy, Reactive uptake of ozone by
liquid organic compounds, Geophys. Res. Left., 25, 931-
934, 1998.
DeMote, W. B., C. J. Howard, S. P. Sander, A. R. Ravis-
hankara, D. M. Golden, C. E. Kolb, M. J. Hampson, R.
F. Molina, and M. J. Kurylo, Chemical kinetics and pho-
tochemical data for use in stratospheric modeling, JPL
Publ. 97- 12, 1997.
Donahue, N., M. Dubey, R. Mohrschladt, K. Demerjian,
and J. Anderson, High-pressure flow study of the reac-
tions OH+NOx- HNO: Errors in the falloff region, J.
Geophys. Res., 102, 6159-6168, 1997.
Fisher, M., and D. Lary, Lagrangian 4-dimensional varia-
tional data assimilation of chemical-species, Q. j. R. Me-
teorol. Soc., 121, 1681-1704, 1995.
Friedl, A., et al., Photochemistry of formaldehyde during
the 1993 Tropospheric OH Potochemistry Experiment, J.
Geophys. Res., 102, 6283-6296, 1997.
Fruekilde, P., J. Hjorth, N. Jensen, D. Kotzias, and
B. Larsen, Ozonolysis at vegetation surfaces: A source
of acetone, 4-oxopentanal, 6-methyl-5-hepten-2-one, and
geranyl acetone in the troposphere, Atmos. Environ., 32,
1893-1902, 1998.
Fulle, D., H. Hamann, H. Hippler, and J. Troe, Temperature
and pressure dependence of the addition reactions of OH
to NO and to NO., IV. Saturated laser-induced fluores-
cence measurements up to 1400 bar, J. Phys. Chem. Ref.
Data, 108, 5391-5397, 1998.
Gierczak, T., J. Burkholder, S. Bauerle, and A. Ravis-
hankara, Photochemistry of acetone under tropospheric
conditions, Chem. Phys., 231, 229-244, 1998.
Jaeg16, L., D. Jacob, Y. Wang, A. Weinheimer, B. Ridley,
T. Campos, G. Sachse, and D. Hagen, Sources and chem-
istry of NO in the upper troposphere over the United
States, Ceophys. Res. Left., 25, 1705-1708, 1998.
Johnson, D., W. Traub, K. Chance, and K. Jucks, Detection
of HBr and upper limit for HOBr: Bromine partitioning
in the stratosphere, Geophys. Res. Left., 22, 1373-1376,
1995.
Kavouras, I., N. Mihalopoulos, and E. Stephanou, Forma-
tion of atmospheric particles from organic acids produced
by forests, Nature, 395, 683-686, 1998.
Kegley Owen, C., M. Gilles, J. Burkholder, and A. Rav-
ishankara, Rate coefficient measurements for the reac-
tion OH+C10products, J. Phys. Chem. A, 103, 5040-
5048, 1999.
Khattatov, B. V., J. C. Gille, L. Lyjak, G. P. Brasseur, V. L.
Dvortsov, A. E. Roche, and J. W. Waters, Assimilation
of photochemically active species and a case analysis of
UARS data, J. Geophys. Res., 10,1, 18,715-18,737, 1999.
Lary, D., Gas phase atmospheric bromine photochemistry,
J. Geophys. Res., 101, 1505-1516, 1996.
Lary, D., Catalytic destruction of stratospheric ozone, J.
Geophys. Res., 102, 21,515-21,526, 1997.
Law, D., Data assimilation: A powerful tool for atmospheric
chemistry, Philosophical Transactions of the Royal Soci-
ety, Series A, Mathematical, Physical and Engineering
Sciences, 1763, 1999.
Lary, D., and J. Pyle, Diffuse-radiation, twilight, and pho-
tochemistry: 1., J. Atmos. Chem., 13,373-392, 1991a.
Law, D., and J. Pyle, Diffuse-radiation, twilight, and pho-
tochemistry: 2., J. Atmos. Chem., 13, 393-406, 1991b.
Lary, D., and D. Shallcross, The potential importance of the
reaction CO and HNO, J. Geophys. Res., in press, 2000.
Law, D., M. Chipperfield, J. Pyle, W. Norton, and L. Ri-
ishojgaard, 3-dimensional tracer initialization and gen-
eral diagnostics using equivalent pv latitude-potential-

19,778 LARY AND SHALLCROSS: CENTRAL ROLE OF CARBONYL COMPOUNDS
temperature coordinates, Q. J. R. Meteorok Soc., 121,
187-210, 1995a.
Lary, D., M. Chipperfield, and R. Toumi, The potential im-
pact of the reaction OH-C10 )HCI- O2 on polar ozone
photochemistry, J. Atmos. Chem., 21, 61-79, 1995b.
McKeen, S., et al., The photochemistry of acetone in the
upper troposphere: A source of odd-hydrogen radicals,
Geophys. Res. Left., 2,i, 3177-3180, 1997.
Meier, R., D. Anderson, and M. Nicolet, Radiation-field in
the troposphere and stratosphere from 240-1000 nm. 1.,
General-analysis, Planet. Space Sci., 30,923-933, 1982.
Menke, W., Geophysical Data Analysis: Discrete Inverse
Theory, Academic, San Diego, Calif., 1984.
Middlebrook, A., D. Murphy, and D. Thomson, Observa-
tions of organic material in individual marine particles at
Cape Grim during the First Aerosol Characterization Ex-
periment (ACE 1), J. Geophys. Res., 103, 16,475-16,483,
1998.
Muller, J., and G. Brasseur, Sources of upper tropospheric
HO: A three-dimensional study, J. Geophys. Res.,
1705-1715, 1999.
Murphy, D., D. Thomson, and T. Mahoney, In situ mea-
surements of organics, meteoritic material, mercury, and
other elements in aerosols at 5 to 19 kilometers, Science,
282, 1664-1669, 1998.
Newchurch, M., et al., Stratospheric NO and NO2 abun-
dances from ATMOS solar-occultation measurements,
Geophys. Res. Lett., 23, 2373-2376, 1996.
Nicolet, M., R. Meier, and D. Anderson, Radiation-field in
the troposphere and stratosphere .2. Numerical-analysis,
Planet. Space Sci., 30,935-983, 1982.
Nolt, I., et al., Stratospheric HBr concentration profile ob-
tained from far-infrared emission spectroscopy, Geophys.
Res. Left., 2,i, 281-284, 1997.
Novakov, T., and J. Penner, Large contribution of organic
aerosols to cloud-condensation-nuclei concentrations, Na-
ture, 365,823-826, 1993.
Novakov, T., D. Hegg, and P. Hobbs, Airborne measure-
ments of carbonaceous aerosols on the east coast of the
united states, J. Geophys. Res., 102, 30,023-30,030, 1997.
Portmann, R., S. Brown, T. Gierczak, , R. Talukdar,
J. Burkholder, and A. Ravishankara, Role of nitrogen ox-
ides in the lower stratosphere: A reevaluation based on
laboratory studies, Geophys. Res. Lett., 26, 2387-2390,
1999.
Press, W., S. Teukolsky, W. Vetterling, and B. Flannery,
Numerical Recipes in Fortran - The Art of Scientific Com-
puting, 2nd ed., Cambridge Univ. Press, New York, 1992.
Sheridan, P., C. Brock, and J. Wilson, Aerosol-particles in
the upper troposphere and lower stratosphere: Elemen-
tal composition and morphology of individual particles
in northern midlatitudes, Geophys. Res. Left., 21, 2587-
2590, 1994.
Shevchenko, V., A. Lisitzin, V. Kuptzov, H. VanMalderen,
J. Martin, R. VanGrieken, and W. Huang, Composition of
aerosols in the marine boundary layer over the seas of the
western Russian arctic, Okeanologiya, 39, 142-151, 1999.
Stoer, J., and R. Bulirsch, Introduction to Numerical Anal-
ysis, Springer-Verlag, New York, 1980.
Vasconcellos, P., P. Artaxo, P. Ciccioli, A. Cecinato,
E. Brancaleoni, and M. Frattoni, Chemical composition
of aerosol collected in the Amazon forest, Quim. Nova,
21,385-393, 1998.
Wang, K., D. Lary, and S. Hall, Improvement of a 3D CTM
and a 4D-VAR assimilation on a vector machine Cray J90
through a multitasking strategy, Comp. Phys. Comm.,
125, 142-153, 2000.
WMO, Scientific assessment of ozone depletion: 1998, Tech.
Rep. JJ, WMO Global Ozone Res. and Monitor. Proj.,
Geneva, 1998.
D. J. Lary, Centre For Atmospheric Science, De-
partment of Chemistry, Cambridge University, Lens-
field Road, Cambridge, CB2 1EW, England. (e-mail:
davi d.lary@atm. ch. cam .ac. uk)
D. E. Shallcross, School of Chemistry, University of Bris-
tol, Bristol, BS8 1TS, England, United Kingdom.
(Received June 22, 1999; revised December 6, 1999;
accepted December 10, 1999.)