JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 101, NO. D?, PAGES 12,555-12,566, MAY 20, 1996
Interpretation of nitric oxide profile Observed in January 1992
over Kiruna
Y. Kondo, 1 S. R. Kawa, 2 D. Lary, 3 T. Sugita, 1 Anne R. Douglass, 2 E. Lutman, 3
M. Koike, 1 and T. Deshler 4
Abstract. NO mixing ratios measured from Kiruna (68øN, 20øE), Sweden, on January 22,
1992, revealed values much smaller than those observed at midlatitude near equinox and
had a sharper vertical gradient around 25 km. Location of the measurements was close to
the terminator and near the edge of the polar vortex, which is highly distorted from
concentric flow by strong planetary wave activities. These conditions necessitate accurate
calculation, properly taking into account the transport and photochemical processes, in
order to quantitatively explain the observed NO profile. A three-dimensional chemistry
and transport model (CTM) and a trajectory model (TM) were used to interpret the
profile observations within their larger spatial, temporal, and chemical context. The NOy
profile calculated by the CTM is in good agreement with that observed on January 31,
1992. In addition, model NOy profiles how small variabilities depending on latitudes, and
they change little between January 22 and 31. The TM uses the observed NOy values. The
NO values calculated by the CTM and TM agree with observations up to 27 km. Between
20 and 27 km the NO values calculated by the trajectory model including only gas phase
chemistry are much larger than those including heterogeneous chemistry, indicating that
NO mixing ratios were reduced significantly by heterogeneous chemistry on sulfuric acid
aerosols. Very little sunlight to generate NOx from HNO3 was available, also causing the
very low NO values. The good agreement between the observed and modeled NO profiles
indicates that models can reproduce the photochemical and transport processes in the
region where NO values have a sharp horizontal gradient. Moreover, CTM and TM model
results how that even when the NOy gradients are weak, the model NO depends upon
accurate calculation of the transport and insolation for several days.
1. Introduction
Reactive nitrogen plays important roles in the chemistry that
leads to ozone (03) losses in the winter Arctic stratosphere.
The sum of reactive nitrogen is defined as NOy.
NOy = NO + NO2 + NO3 + 2N205 + HNO3 + HO2NO2
+ C1ONO2 + BrONO2 + aerosol nitrate
Active nitrogen NOir (NO, NO2) sequesters active chlorine by
the reaction
(R1) C10 + NO2 + M  C1ONO2 + M
resulting in the effective reduction in the 0 3 loss rates by
catalytic cycles involving active chlorine. The main source of
NO 2 in the lower stratosphere in the Arctic winter and spring
is the photolysis of HNO3, which constitutes the major fraction
of NOy.
NOir is oxidized into HNO 3 indirectly via N20 s and C1ONO 2
lSolar-Terrestrial Environment Laboratory, Nagoya University,
Toyokawa, Japan.
2NASA Goddard Space Flight Center, Greenbelt, Maryland.
3Department of Chemistry, University of Cambridge, Cambridge,
England.
4Department of Atmospheric Science, University of Wyoming,
Laramie.
Copyright 1996 by the American Geophysical Union.
Paper number 96JD00578.
0148-0227/96/96JD-00578509.00
on the surface of sulfuric acid aerosols through the following
heterogeneous reactions:
(R2) N205 + H20 --> 2HNO3 3/2
(R3) C1ONO2 + H20---> HOC1 + HNO3 T3
where 3/2 and 3/3 are reaction probabilities for (R2) and (R3).
The importance of (R2) at high latitudes in winter was first
proposed by Evans et al. [1985], although the reliable reaction
rate constant was not available at that time. Laboratory exper-
iments showed that the reaction (R2) is insensitive to the
composition of aerosols and particle size [Hanson and Ravis-
hankara, 1991; Fried et al., 1994] and slightly dependent on
temperature [DeMore et al., 1994]. On the other hand, (R3)
becomes important only at low temperatures typical for polar
winter [Hanson et al., 1994, and references therein]. Attempts
were made to quantitatively test the role of (R2) by the in situ
measurements of reactive nitrogen on board the ER 2 aircraft
at midlatitudes [Fahey et al., 1993; Kawa et al., 1993]. They
showed the importance of (R2) in controlling the NOx/NOy
ratio in the lower stratosphere at midlatitude under back-
ground and volcanic aerosols conditions.
Compared with midlatitudes, high-latitude winter is charac-
terized by short daylight hours and lower temperatures. These
conditions can change the importance of heterogeneous effects
compared with midlatitudes. Low insolation especially reduces
aerosol surface area (SA) required to effectively lower the
NOx/NOy ratio as expected from the calculation by Mills e! al.
[1993]. NO mixing ratios were observed up to 30.5 km from
12,555

12,556 KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA
Table 1. Balloon Measurements From Kiruna on January
22 and 31, 1992
Date Species Method Investigator
January 22, 1992
Ascent and descent NO chemiluminescence Y. Kondo
Ascent 03 ECC Y. Kondo
Ascent aerosol particle counter Y. Kondo
Descent N20 grab sampler U. Schmidt
Ascent and descent actinic flux filter radiometer C. Schiller
Ascent
Ascent
Descent
January 31, 1992
NOy chemiluminescence Y. Kondo
03 ECC Y. Kondo
N20 grab sampler U. Schmidt
ECC, electrochemical cell.
Table 2. Solar Zenith Angle During Ascent of the Balloon
on January 22, 1992
Pressure, hPa Altitude, km SZA, deg PT, K
84.4 16.7 91.0 420
59.4 18.8 90.5 450
40.2 21.1 90.0 492
24.9 23.8 89.5 554
17.5 25.8 89.0 625
15.9 26.4 88.8 650
12.0 28.0 88.5 717
9.9 29.2 88.2 800
8.6 30.1 88.0 857
bbreviations are SZA, solar zenith angle, and PT, potential tem-
perature.
Kiruna (68øN, 20øE), Sweden, on January 22, 1992 [Aimedieu et
al., 1994]. The NO values up to 25 km were extremely low, as
compared with those observed at midlatitude in fall [e.g.,
Kondo et al., 1989; Y. Kondo et al., The effect of Pinatubo
aerosols on stratospheric NO, submitted to Journal of Geo-
physical Research, 1996 (hereinafter referred to as Y. Kondo et
al., submitted manuscript, 1996)]. Since Kiruna was located
near the terminator and edge of the polar vortex, quantitative
interpretation of the observed NO profile requires accurate
evaluation of the chemical and transport processes controlling
the abundance of active nitrogen. For this purpose, partition-
ing of NO v was calculated by models which take into account
the photochemistry and transport processes. The comparison
shows the capability of the current models in accurately pre-
dicting the NO which has a large spatial gradient and under-
goes rapid temporal changes due to large variations in the
photolysis rates controlling the NO levels.
2. Data Set
2.1. Balloon Data
Balloon-borne measurements of NO, 03, N20, aerosol, and
actinic solar flux up to 30.5 km (8.2 hPa) were made from
Kiruna (68øN, 20øE), Sweden, on January 22, 1992, as a part of
the European Arctic Stratospheric Ozone Experiment (EA-
SOE) JAimedieu et al., 1994; Bauer et al., 1994; Schiller et al.,
1994]. Similarly, NOv, 03, N20, and aerosol were measured on
January 31 [Kondo et al., 1994; Bauer et al., 1994]. The species
measured on these days are summarized in Table 1. Most of
the experimental data used in the present paper were already
given in the papers cited above. In this paper we focus on the
quantitative interpretation of the NO profile using these data.
Some additional data closely relevant o the present work are
given below, with brief summaries of the instruments and ac-
curacies of the measurements.
january 22. NO was measured with a chemiluminescence
detector [Kondo et al., 1988] during ascent between 0720 and
0930 UT and also during part of the descent, which was made
about 2.5 hours after the ascent. The solar zenith angle (SZA)
during ascent is given in Table 2. The SZA during descent was
87.8 ø _+ 0.3 ø. The accuracy and the detection limit for the NO
values averaged for 10 s are about 10% and 15 parts per trillion
by volume (pptv), respectively, between 20 and 32 km [Kondo
et al., 1988, 1989].
The NO values during descent were somewhat larger than
those during ascent probably in part due to the diurnal varia-
tion of NO. However, there is, of course, a rapid variation in
NO near the terminator. The measured NO values were gen-
erally lower than 50 pptv below 23 km. At 26 km the NO
mixing ratio was still as low as 100 pptv. These NO values were
an order of magnitude smaller than those measured at mid-
latitudes i n fall [e.g:, Kondo et al., 1989, also submitted manu-
script, 1996]. The NO mixing ratio started to increase rapidly at
around 26 km. A low NO mixing ratio was also observed below
27 km near 50øN in winter by Ridley et al. [i984, 1987].
03 was measured with an electrochemical concentration cell
(ECC) sonde during ascent with an uncertainty of 2-10% be-
tween 15 and 30 km [Komhyr et al., 1995] and is shown in
Figure 1. N20 measurements were made by collecting air sam-
ples during the parachute descent of the gondola, followed by
gas chromatographic analysis in the laboratory [Bauer et al.,
30
25
 20
15
CTM
ß
ß
920122
.... 920131
....... 920213(Deshler)
 920122
--o-- 920131
10
100
0 1 2 3 4 5 6
0 3 (ppmv)
Figure 1. The 03 profiles measured during ascent on Janu-
ary 22, January 31, and February 13, 1992. Bars indicate the
uncertainty of the measurements. 03 profiles calculated by the
chemistry and transport model (CTM) for January 22 and 31
are also shown.

KONDO ET AL.' NITRIC OXIDE IN JANUARY OVER KIRUNA 12,557
30
25
 2o
15
ß 92o122
o 920213(Deshler)
ß CTM
A TM1
....... BKG (wao)
\ ß
ß
0.1 1 10
Aerosol surface area density (gin 2CIT1-3)
lO
lOO
Figure 2. Profiles of aerosol surface area (SA) measured
from Kiruna on January 22, 1992 (solid circles), and February
13, 1992 (open circles). The SA used for CTM are shown as
squares. Typical background SA (BKG) is shown as a dashed
line.
1994]. The uncertainty of the measurement is _+ 10% above 24
km and less than +_5% at lower altitudes.
Aerosol concentrations were measured with an optical par-
ticle counter [Aimedieu e! al., 1994]. A detailed description of
the instrument is given by T. Sugita et al. (manuscript in prep-
aration, 1996). Concentrations of aerosols with radii larger
than 0.23, 0.28, 0.31, 0.41, 0.58, 2.0, and 3.3/m were measured.
The detection limit was 1.4 x 10 -3 cm -3. Concentrations of
aerosols with radii larger than 2.0/m were below the detection
limit in the stratosphere. SA was calculated from the measured
aerosol size distribution, and the ascent profile is shown in
Figure 2. The SA for a typical nonvolcanic period as given by
World Meteorological Organization (WMO) [1992, Table 8-8] is
also shown for reference. The observed SA was much larger
than the background value owing to the Pinatubo volcanic
aerosols.
The observed temperature shown by Aimedieu e! al. [1994]
was the lowest between 21 and 24 km, barely reaching the nitric
acid trihydrate (NAT) condensation temperature, which was
calculated following the formulation by Hanson and Mauers-
berger [1988]. Considering that in the Arctic, polar strato-
spheric clouds (PSC) are often only observed at relatively high
supersaturation [Kawa et al., 1992; Kondo et al., 1992], the
observed particles on January 22 can be considered to be
composed of sulfuric acid and water.
For reference, SA estimated from the aerosol concentra-
tions measured by Deshler [1994] over Kiruna on February 13,
1992, are also shown in Figure 2. Condensation nuclei and
aerosols with radii larger than 0.15, 0.25, 0.48, 1.08, 1.93, 2.76,
4.74, and 9.50/m were measured. Above 24 km, aerosols with
radii smaller than 0.15/m made the major contribution to SA
[Deshler, 1994]. Since the smallest size of the aerosols mea-
sured on January 22 was 0.23 /.rm, SA estimated from mea-
surements of aerosols larger than this size can be significantly
underestimated above 24 km and are therefore not shown
here.
SA observed on January 22 and February 13 agrees reason-
ably well below 23 km, as shown in Figure 2. Kiruna was
located outside the vortex at 475 K (20 km) and at the inner
edge at 550 K (23 km) on January 22, as discussed below. The
potential vorticity (PV) value over Kiruna at 550 K on Febru-
ary 13 was 126 x 10 -6 K m 2 kg- s-, which is similar or even
somewhat larger than the value on January 22 as given in Table
3. The similar aerosol profiles on these two days are consistent
with the similarities in the location of the vortex boundary
relative to Kiruna. The O 3 profile observed on February 13 is
shown in Figure 1 for comparison. The O3 profile on February
13 agrees well with that on January 22 above 22 km. It has been
found that correlation between SA and O 3 above 16 km
changed significantly depending on the location of measure-
ments relative to the vortex boundary [Borrmann e! al., 1995].
Considering the agreement of the PV and O_ values on Feb-
ruary 13 with those on January 22, SA from Deshler [1994] may
be considered to give a reasonable estimate for January 22
above 23 km.
NO2 photodissociation coefficient (J(NO2)) has been de-
rived from the UV flux measurements made simultaneously
(C. Schiller, unpublished data, 1996). The radiometer used to
measure the UV flux is described by Schiller e! al. [1994].
During ascent, SZA was larger than 89 ø below 22 km. In this
altitude region, J(NO2) was significantly reduced owing to in-
creased optical thickness caused by the enhanced aerosol load-
ing as shown in Figure 2. This gives no impact to the present
analysis because comparisons of model values with NO values
obtained during ascent are made only above 24 km.
Table 3. Potential Vorticity Values at Various Potential Temperatures at Kiruna
PT, Pressure, Approximate PV Jan. 22, PV Jan. 31,
K hPa Altitude, km 10 -6 K m 2 kg-  s-  In/Out 10 -6 K m 2 kg-  s-  In/Out
350 140 13 6 '\" 6 '\"
380 120 14 9 out 7 out
400 100 16 10 out 9 out
475 46 20 24 out 33 edge
550 25 23 102 in/edge 108 in/edge
650 16 26 144 in 129 in
700 13 27 190 in 170 in
800 10 29 284 in 250 in
Abbreviations are PV, potential vorticity, and PT, potential temperature.

12,558 KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA
,,
,,
/,
Figure 3a. Isentropic trajectories arriving over Kiruna on
January 22, 1992, on isentropic level of 475 K.
January 31. NOy was measured with a chemiluminescence
NO detector combined with a gold catalytic converter heated
at 300 ø +_ iøC, as described in detail by Kondo et al. [1990,
1992]. The overall error is estimated to be +_ 15% between 15
and 31 km. 03 was measured with an ECC sonde during ascent
and is shown in Figure 1.
2.2. Meteorological Data
Maps of Ertel's potential vorticity (PV) at six potential tem-
perature surfaces of 350 K (13 km), 380 K (14 km), 400 K (16
km), 475 K (20 km), 550 K (23 km), and 700 K (27 km) were
provided by the European Center for Medium-Range Weather
Forecasts (ECMWF). The PV values at 800 K were calculated
using the same meteorological data as used for the chemistry
transport model which is described below. The vortex bound-
ary at 475, 550, and 700 K was defined as the region with PV
values of 40 +_ 8 x 10 -6 90 + 12 x 10 -6, and 160 + 20 x 10 -6
K m 2 kg- s-2, respectively. PV values at these potential tem-
peratures over Kiruna on January 22 and 31 are given in Table
3. They were 24, 102, and 190 x 10 -6 K m 2 kg -] s -] at 475,
550, and 700 K, respectively, on January 22. The PV maps
indicate that Kiruna was located outside the vortex below 475
K, at the inner edge of it at 550 K, and inside it at 700 and 800
K. On January 31 the location of Kiruna relative to the vortex
boundary was similar to January 22.
Isentropic 30-day back trajectories were calculated from
United Kingdom Meteorological Office analyses. The trajecto-
ries used in this study were for January 22, 1992, and finished
over Kiruna. The trajectories a't the 475- and 550-K levels are
shown in Figures 3a and 3b, respectively. The air parcels below
475 K originate from lower latitudes, while they remain in high
latitudes at 550-1000 K, corresponding to the PV analysis as
mentioned above.
3. Models
3.1. Chemical Scheme
At night, NOx is oxidized into N20 5 via the following reac-
tions:
(R4) NO2 + 03 --> NO3 + 02
(R5) NO2 + NO3 + M--> N2Os + M
NOx is also sequestered as C1ONO 2 via (R1). N205 and
C1ONO 2 are photolized back to NOx with time constants of
from several hours to a few days, depending on altitude and
SZA.
(R6) N2Os + h v- NO2 + NO3
(R7) C1ONO2 + h v--> C1 + NO3
Further oxidation of N205 and CiONO 2 into HNO 3 occurs by
the heterogeneous reactions (R2) and (R3). Photolysis or the
reaction with OH reproduces NOx from HNO 3 with a time
constant of a few weeks or less.
(R8) HNO3 + hv--> NO2 + OH
(R9) HNO3 + OH - NO3 + H20
In order to quantitatively evaluate critical processes in con-
trolling NOx levels in the Arctic winter in combination with
transport processes, the NASA Goddard Space Flight Center
(GSFC) three-dimensional (3-D) chemistry and transport
model (CTM) and trajectory models from NASA GSFC and
the University of Cambridge were used. Descriptions of these
are given below.
3.2. Chemistry and Transport Model (CTM)
A meteorological data assimilation [Schubert et al., 1993] was
used to produce winds and temperatures that drove the off-line
3-D chemistry and transport model. In the assimilation, fields
for winds and temperatures which satisfy the equations of
motion are produced from combination of a general circula-
tion model and observations. Constituent fields calculated with
a CTM using these observation-based winds and temperatures
may be compared directly with COnstituent observations. Con-
stituents in the CTM were initialized November 15, 1991, with
550 K
t ...... / .......... -\" \",, .\"i - / '\":,;
i
Figure 3b. Same as Figure 3a but for the 550-K level.

KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA 12,559
500 K
8OO K
N20 920122
650 K .......... ...............
lOOO K
320
Plate la. Maps of N20 calculated by CTM for January 22, 1992, at isentropic levels of 500, 650, 800, and
1000 K. The vortex boundary is indicated as a white line. The location of Kiruna is denoted with a star.
03 data from microwave limb sounder (MLS), N20 from cryo-
genic limb array etalon spectrometer (CLAES) on upper at-
mosphere research satellite (UARS), and results from a two-
dimensional model. The CTM ran with a horizontal resolution
of 2 ø in latitude and 2.5 ø in longitude with 15-min time steps.
Mixing ratios of 34 species are calculated at 25 pressure levels
between 0.63 and 922 hPa.
Equatorward of 60øN, the Stratospheric Aerosol and Gas
Experiment II (SAGE II) satellite experiment SA data was
used [Thomason and Poole, 1993]. The SA north of 60øN was
chosen from the data given in chapter 8 of WMO [1992], and
the adopted profile is shown in Figure 2. This SA represents
the enhanced value. As can be seen from Figure 2, the SA
values used in this model are similar to those obtained by in
situ measurements at 24 and 27 km. The model SA values are
much lower than measured values below 24 km. However, it
has been found from simple calculation that the partitioning of
the reactive nitrogen depends little on the SA values exceeding
1 /xm 2 cm -3 under Arctic winter conditions due to saturation
of the heterogeneous reaction on sulfuric acid aerosol as ob-
served in midlatitudes [Fahey et al., 1993; Mills et al., 1993].
Photolysis in the CTM uses a look-up table based on the
radiative transfer model of Anderson and Lloyd [1990] and
Anderson et al. [1995] with molecular cross sections from De-
More et al. [1992]. The calculations assume clear sky and a
uniform surface albedo of 0.3. Absorption by overhead ozone
depends on the model 03 profile.
3.3. Trajectory Models
TM1. The 10-day back trajectories from Kiruna on January
22 were calculated on 20 isentropic levels between 330 and
1000 K using the same wind and temperature fields as used for
CTM. The vertical resolution is about 1.5 km above 20 km
Changes in the partitioning of reactive nitrogen were calcu-
lated along these trajectories including the same chemical re-
actions as used for CTM. The 03 and NOy mixing ratios
obtained by the balloon observations were used. The SA used
by the model was taken from the in situ measurements up to 27
km as shown in Figure 2.
TM2. The trajectory model used in this study was a version
of a new suite of models called AUTOcHEM. A total of 62
species were included. No family or photochemical equilibrium

12,560 KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA
500 K
NOy 920122
65O K
j.
8OO K -- 1000 K
2O
0.0
Plate lb. Same as Plate la but for NO v.
assumptions were made. Sixty species, including reactive nitro-
gen, hydrogen, chlorine, and bromine species, were integrated
separately with a 15-min time step. The version of AUTO-
CHEM used in this study contained a total of 292 reactions,
170 bimolecular eactions, 37 trimolecular eactions, 47 pho-
tochemical reactions, and 38 heterogeneous reactions on PSC
and sulfate aerosol. The rate constants for the reactions were
taken from Atkinson et al. [1992] and DeMore et al. [1994].
Photolysis rates were calculated using full spherical geometry
and multiple scattering as described by Lary and Pyle [1991]
after Meier et al. [1982], Nicolet et al. [1982], and Anderson
[19831.
The isentropic trajectory model calculations were performed
to simulate the chemical composition of air parcels arriving
over Kiruna on January 22. The values used for NOy, 03, and
SA are the same as for TM1. The partitioning of reactive
nitrogen at the start of the simulation, together with all the
other constituent concentrations, was taken from a January
simulation of the Cambridge two-dimensional model [Har-
wood and Pyle, 1977, 1980].
4. Comparison with Models
4.1. Chemical Fields
In order to understand the location of Kiruna in the chem-
ical fields, hemispheric plots of N20, NOy, NOx, and NO on
the 500, 650, 800, and 1000 K calculated from the CTM are
shown in Plates la-ld for January 22. For reference the
boundary of the polar vortex is shown as a white line. The
boundary was determined from PV gradients [Nash et al.,
19961.
The modeled N20 values inside the vortex are smaller than
those outside at 500 and 650 K. This is due mainly to stronger
descent by diabatic cooling inside the vortex. The gradient of
N20 through the vortex boundary is also caused by inefficient
horizontal transport of N20 from lower latitude to inside the
vortex. Corresponding to the N20 distribution, NOy, values
inside the vortex are larger than those outside it at 500 and 650
K. However, at 800 and 1000 K, NOy values inside the vortex
are slightly smaller than outside it. As shown below, NOy
decreases with altitude above 26 km due to net loss of NOy

KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA 12 561
500 K
NOx 92O122
65O K
,,.o
8OO K lOOO K
12
0.0
Plate lc. Same as Plate la but for NOx.
through the reaction N + NO --> N 2 + O in the tropical upper
stratosphere. Stronger descent inside the vortex brings down
air poorer in NOy from above, causing smaller NOy values at
800 and 1000 K.
In contrast, large gradients in NOx and NO are not defined
by the vortex boundary, as can be seen from Plates l c and ld.
The NO, c/NO v ratio is controlled by the conversion of NO to
N205 and further oxidation into HNO 3 and the photolysis of
these reservoir species. Here 3'3 calculated based on the results
by Hanson and Ravishankara [1991] reaches the maximum
value of 0.01 at 23 km where temperature was the lowest. Since
3'2 is 0.1, the 3'3/3'2 ratio does not exceed 0.1. Above 25 km and
below 20 km the 3'3/3'2 ratio is lower than 0.04. Therefore (R2)
plays a dominant role in oxidizing NOx.
The time constants of the photolysis of N205 and HNO3
vary from several days to a few weeks, depending on altitude
and SZA. Therefore the NO/NOy ratio depends on the time
history of air parcels over the last few days. On the other hand,
NO is in almost immediate photochemical equilibrium with
NO 2. Therefore the NO/NO ratio is controlled by the instan-
taneous solar illumination, temperature, and 03 concentra-
tion. Plate ld clearly shows strong control of insolation on NO.
Because of the complex distribution of the NO/NOy ratio,
interpretation of the NO profile measured at Kiruna requires
models which adequately take into account photochemical and
transport processes.
The profiles of 03 observed on January 22 and 31 are com-
pared with those constructed from the CTM for the same
location (68øN, 20øE) and dates in Figure 1. Modeled 0 3 val-
ues agree well with the measured values between 21 and 27 km.
However, below 20 km the model significantly overestimates
the O3 abundance. On the other hand, at 31 km the calculated
03 value is 20-30% lower than that observed. The discrepancy
below 20 km is not relevant to the present discussion, since
detailed comparison of the observed and calculated NO values
are made above 20 km. The effect of the 03 difference at 31 km
is discussed below.
The profiles of N20 observed on January 22 and 31 are
compared with those calculated by the model in Figure 4.
Modeled N20 values agree well with the measured values up
to 22 km. However, at higher altitudes the model overesti-
mates the N20 abundance. The observed N20 mixing ratios as
low as 10-20 parts per billion by volume (ppbv) are caused by
strong descending motion inside the vortex induced by diabatic

12,562 KONDO ET AL.' NITRIC OXIDE IN JANUARY OVER KIRUNA
500 K
NO 92O122
650 K
800 K lOOO K
Plate ld. Same as Plate la but for NO.
8
0.0
cooling [Batter et al., 1994]. It is indicated that the descending
process is not well reproduced in the assimilation wind fields.
The model NOy profiles at 66øN and 70øN at 20øE are com-
pared with that observed in Figure 5. The model NOy profile at
68øN almost overlaps with that at 70øN above 17 km and hence
is not plotted here. For reference, the model NOy profile at
68øN on January 22 is also shown. Generally, the model NOy
for 68øN and January 31 agrees well with the observed one.
The agreement is especially good below 23 km. This is consis-
tent with the comparison of N20 as discussed above, consid-
ering that N20 and NOy are tightly correlated in the lower
stratosphere [Fahey et al., 1990; Loewenstein et al., 1993; Kondo
et al., 1994]. Even above 22 km, difference between the model
and measurements is not so large as for N20. This is because
NOy mixing ratios do not increase linearly with the decrease in
N20 for N20 mixing ratio lower than 100 ppbv [Loewenstein et
al., 1993; Kondo et al., 1994], as NOy is subject to photochem-
ical loss at these altitudes. Because of this, the lack of the
strong descent does not lead to significant underestimation of
the NOy mixing ratio by the model.
It can be seen from Figure 5 that the profiles of the modeled
NOy mixing ratio vary little with latitude between 66øN and
70øN. Correspondingly, the N20 mixing ratios vary little with
latitude, although they are not shown here. The model also
predicts that the changes in the NOy values between January
22 and 31 are small. Similarity in the N20 profiles on these
days can be seen in Figure 4.
4.2. NO Profile
The model NO profiles by CTM for 0900 and 1200 UT on
January 22 are compared with the observed profiles in Figures
6 and 7. In the model the SZA at 0900 and 1200 UT are 89.1 ø
and 88.8 ø , respectively. In addition, profiles of all reactive ni-
trogen species calculated for 1200 UT on January 22 are shown
in Figure 8. The partitioning of NOy at each level is given in
Table 4. These show that NO + NO 2 is a minor component of
NOy. Considering the SZA given in Table 3, ascent data which
can be used for comparison with the CTM are limited to those
obtained above 24 km, where SZA was smaller than 89 ø. All
the descent data can be used for the comparison, although the
number of data points below 26 km is limited due to the high
descent speed. For these data, J(NO2) was disturbed little by
the Pinatubo aerosol, as discussed above.
The NO mixing ratios were calculated by TM1 on the tra-

KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA 12,563
jectories reaching Kiruna on January 22. The calculated NO
profile for 0930 UT is shown in Figures 6 and 7.
The changes in the NO/NOy ratio simulated by TM2 along
the back trajectories are shown in Figure 9. The NO/NOy
ratios were also calculated including only gas phase chemistry,
and these results are also shown as a reference. The NO/NO
ratio at the start of the simulation has little influence on the
results shown. Air parcels arriving at Kiruna during sunrise on
January 22 on the isentropic level at 550 K had seen less
sunlight than those at the other levels during the previous 10
days. This is clearly evident when examining the 10-day time
series of the NO/NOy ratio shown in Figure 9. On the 550-K
isentropic surface, the most sunlight was seen on January 16
and 17. The NO profile calculated by TM2 is compared with
the observations in Figures 6 and 7. in Figure 7 the NO values
calculated including only gas phase chemistry are compared
with those including heterogeneous chemistry.
Both CTM, TM1, and TM2 well reproduce the observed NO
profile between 20 and 27 km. The good agreement between
the observations and TM1 results extends up to 30 km. The
calculated NO values between 19 and 22 km are lower than 40
pptv in agreement with the descent data. At this level, NOx/
NOy and 2(N2Os/NOy) ratios calculated by CTM are about
0.01 and 0.001, respectively (Table 4). Reaction (R2) reduces
the NO/level by more than an order of magnitude, as can be
seen from Figures 7 and 9. Agreement of TM1 and TM2 with
CTM, which uses less SA than TM1 and TM2, indicates satu-
ration at low SA amounts for these trajectories. NO mixing
ratio at 24 km has been calculated as a function of SA by using
TM2. The NO values relative to those including only gas phase
chemistry are 0.18, 0.13, 0.11, and 0.09 for SA = 0.40, 0.60, 1.0,
30 '\"',,,,,,,,,,,,,,,,,,, 68øN
25
0 ß ,'
,,
 20
,, o
Observed(
15 ß 920122
o 920131 CTM _.__ 920122
--c-- 9201
0 50 100 250 300
10
lOO
15o 200
N20 (ppbv)
350
Figure 4. Comparison of the N20 profiles observed from
Kiruna on January 22 (solid circles) and January 31, 1992
(open circles), with those calculated by CTM (solid and open
squares).
30
25
 20
15
ß Observed (920131)
-o-. 70øN(920!31)
.... A. . .. 66ON(920131)
-- 68ON(920122)
10
lOO
0 2 4 6 8 10 12 14 16
NOy (ppbv)
Figure 5. Comparison of the NOy profile observed from
Kiruna on January 31, 1992 (solid circles), with those calcu-
lated by CTM for 66øN (triangles) and 70øN (diamonds). The
CTM results for January 22 are also shown (solid squares).
and 2.0 m 2 cm -3, respectively. Since background SA is 0.4-
0.6 m 2 cm -3 at 20-24 km (Figure 2), predominant reduction
in NO below 24 km is predicted to occur by heterogeneous
chemistry under nonvolcanic conditions. Volcanic aerosol re-
duces NO further by 50%.
The rapid increase in the NO values between 24 and 29 km
is well reproduced in the models. As can be seen from Table 4,
the increase in NO in this altitude region is due to an increase
in the NOx/NOy ratio resulting from the reduced SA (Figure 2)
and increased rate of HNO3 photolysis. The increase in the
NO values above 24 km is associated with the decrease in the
HNO3 values.
The CTM model NO values at 0900 UT are somewhat
smaller than those at 1200 UT between 20 and 24 km, as can
be seen from Figure 7. At 24 km the NOx mixing ratio is 80
pptv at 0900 UT, while it is 270 pptv at 1200 UT. At 0900 UT
the model values of N205, HNO3, C1ONO2, and HO2NO 2 are
larger than those at 1200 UT. Most of the difference in the
NOx value is balanced by the difference in the values of HNO3
and C1ONO2, which are the dominant components of NOy
below 24 km, as can be seen from Figure 8. Considering the
time constant of the photolysis of these reservoir species, the
difference in the NO x values between 0900 and 1200 UT is
primarily caused by the diurnal variation of NOx resulting from
C1ONO 2 photolysis. This emphasizes the importance of using
a model that accurately simulates the diurnal cycle in compar-
ing a single NO profile at this critical season and latitude.
At 31 km the CTM NO mixing ratio is 3.9 ppbv at 1200 UT.
Similarly, TM2 NO values are 2.5-3.6 ppbv at 30.5 km. If these
values are used to linearly interpolate model values between 27
and 31 km, the calculated NO values are considerably higher
than those observed between 29 and 30.5 km, as can be seen

12,564 KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA
3O
25
-o
: 20
15
92O122
ß Ascent
o Descent
ß TM1
ß TM2 (Het)
--e-- CTM (9UT)
-O-- CTM (12UT)
10
100
0 1 2 3 4 5
NO (ppbv)
Figure 6. Comparison of the NO profile observed from
Kiruna on January 22, 1992 (solid and open circles), with those
calculated by CTM for 1200 UT (open diamonds) and 0900 UT
(solid diamonds). The NO profile calculated by TM1 is shown
as closed triangles. The TM2 values including heterogeneous
chemistry (TM2(Het)) are shown as solid squares.
from Figure 6. In contrast, the TM1 values in this altitude
region are in better agreement with the observations as men-
tioned above. The shape of the NO vertical profile is better
simulated by TM1 owing to higher vertical resolution. The 0 3
mixing ratio used for TM1 around 30 km is 20-30% higher
than that used by CTM. Specifying observed 0 3 in TM1 de-
creases NO by about 15% from that using CTM 03, indicating
that the difference in the 03 value is not the major cause of the
difference between CTM and TM1.
Garcia and Solomon [1994] examined the possible ffect of
(R2) in the upper stratosphere, where SA and composition of
aerosol are poorly known. They used SA given by WMO [1992]
at 32 km and extrapolated it exponentially above 32 km. It has
been found that in the upper stratosphere, HNO 3 mixing ratios
calculated including (R2) are considerably larger than those
for the gas phase case in high-latitude winter when the effect of
(R2) is maximized. This effect corresponding to the conditions
for our observations was calculated using CTM. The SA shown
in Figure 2 was extended up logarithmically, leading to the
value of 0.041/m 2 cm -3 at 7 hPa. Over 20 day sfrom January
2 to January 22, this reduced NO x and N20 5 mixing ratios at 7
hPa by 0.8 ppbv (20%) and 0.5 ppbv (25%), respectively, from
the gas phase case. On the other hand, HNO 3 increased by 1.9
ppbv (nearly 70%), and C1ONO2 was not significantly affected.
Since the HNO 3 was still increasing after 20 days, the full effect
of (R2) may be greater than these values. Since the present
models do not include the effect of (R2) above 28 km, the
calculated NO may be overestimated to some extent.
It can be concluded that the observations of NO made over
Kiruna during sunrise on January 22 are consistent with cal-
920122 30 ß Ascent ro .,\"/
o Descent =IZ 4
ß TM1 'e ß TM2 (Het)
[] TM2 (Gas) A..,;> ---- CTM (9UT) m ) []
-O-- CTM (12UT)
 _
E
(t) 25  .
o ,,,I ß / []
O 0;
20 ,' ! ,
,,,i / , , , , , ,,,i ! i , , , ,,,1 , , ,
0.01 0.1 1
NO (ppbv)
lO
50
Figure 7. Same as Figure 6 but for above 19 km. The TM2
values including only gas phase dhemistry (TM2(Gas)) are also
shown as open squares.
culations made by the CTM, TM1, and TM2. These models
reproduce the observed very low NO/NOy ratio and identify
the two main reasons for this: (1) There was very little sunlight
to generate NOx from HNO3, and (2) there was oxidation of
4O
35
''30
 25
2O
15
\"- ..... '-, 920122 (68øN) \".,,
\, \"x,,x
1013=
100
0.001 0.01 0.1 1 10
Mixing ratio (ppbv)
Figure 8. Profiles of each NOy species calculated by CTM
for Kiruna at 1200 UT on January 22, 1992.

KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA 12,565
Table 4. Partitioning of NOy Calculated by the Chemistry and Transport Model for January 22, 1992
Pressure,
hea NO/NOy 2(N2Os)/NOy HNO3/NOy HO2NO2/NOy C1ONO2/NOy
92 0.008 0.0006 0.905 0.009 0.077
77 0.0.08 0.0005 0.893 0.010 0.088
65 0.009 0.0006 0.884 0.009 0.098
55 0.011 0.0007 0.884 0.009 0.096
45 0.013 0.0011 0.889 0.008 0.089
35 0.013 0.0009 0.891 0.006 0.089
24 0.019 0.0030 0.879 0.004 0.095
14 0.059 0.0517 0.783 0.003 0.103
7 0.514 0.264 0.147 0.002 0.073
NO x due to hydrolysis of N205 which was particularly notice-
able between 20 and 23 km.
5. Summary
CTM, TM1, and TM2 were run to simulate the profile of the
mixing ratio of NO up to 30.5 km obtained by balloon-borne
measurements made at 68øN in January 1992. Hemispheric
maps of NOy on isentropic surfaces generated from CTM
predict he NOy mixing ratio to be rather uniform near Kiruna.
On the other hand, NOx and NO mixing ratios have large
0.3
0.2
850K (30km)
700K (27km)
o,
....
;',
I
0.1
550K (23km)
 .,\"i
, , ,
475K (20km)
0.1
...... TM(Gas)
-- TM(Het)
12 14 16
Day of January 1992
Figure 9. Temporal variation of the NO/NOy ratio calcu-
lated by TM2 on four isentropic levels: 850, 700, 550, and
475 K.
spatial variability near the terminator, where Kiruna was lo-
cated.
The N20 profiles calculated by CTM for January 22 and 31
are in good agreement with those observed below 22 km. At
higher altitudes the model N20 values are considerably higher
than those measured, since the effect of strong descent by
diabatic cooling in the Arctic winter is not well reproduced in
the assimilation wind fields. On the other hand, the model NOy
values are in good agreement with those measured up to 30.5
km, giving a good basis for comparing the model and measure-
ment for NO, since NO is determined by the partitioning
among NOy species.
The NO profiles calculated by CTM, TM1, and TM2 are in
agreement with the observed profile between 20 and 27 km.
The calculated NO values between 20 and 22 km are lower
than 40 pptv, in agreement with the descent data. At this level,
NOx constitutes a very small fraction of NOy. Reaction (R2),
which is saturated already at SA of 1-2 txm 2 cm -3, reduces the
NOx level by more than an order of magnitude. The increase in
NO and the NOx/NOy ratio between 24 and 29 km results from
the reduced SA and increased rate of generation of NO x from
HNO3. The good agreement between the observations and
TM1 results extends up to 30 km, owing mainly to the higher
vertical resolution of TM1.
The CTM model predicts some diurnal variation of NO and
NO x in the morning. At 24 km the NOx mixing ratio is 80 pptv
at 0900 UT, while it is 270 pptv at 1200 UT. Considering the
time constant of the photolysis of the reservoir species, the
diurnal variation of NO x results primarily from the C1ONO2
photolysis.
It has been found that the heterogeneous reaction (R2) can
play a role in reducing NO at this altitude if SA of 0.04 txm 
cm -3 is assumed. The present calculations, which do not in-
clude (R2) above 28 km, may overestimate NO values to some
extent.
It can be concluded that CTM, TM1, and TM2 can generate
the very low NO/NOy ratios observed near the terminator
where NO/NOy ratios have large spatial variations. These
models indicate that the NO/NOy ratios were low for two main
reasons: (1) There was very little sunlight to generate NO x
from HNO3, and (2) there was oxidation of NO x due to hy-
drolysis of N20 5 which was particularly noticeable between 20
and 23 km. The quantitative agreement between observed and
model NO values indicates that photochemical and transport
processes near the terminator and the vortex edge are accu-
rately calculated by the models.
Acknowledgments. Partial funding by CEC and the Japanese
MESC are gratefully acknowledged. Thanks are due to C. Schiller,

12,566 KONDO ET AL.: NITRIC OXIDE IN JANUARY OVER KIRUNA
who showed us his data before publication. Potential vorticity maps
were calculated by ECMWF and provided through NILU.
References
Aimedieu, P., Y. Kondo, T. Sugita, M. Koike, and W. R. Sheldon, In
situ measurements of nitric oxide, ozone, and aerosol in the Scan-
dinavian Arctic stratosphere in January 1992, Geophys. Res. Lett., 21,
1243-1246, 1994.
Anderson, D. E., The troposphere to stratosphere radiation field at
twilight: A spherical model, Planet. Space Sci., 31, 1517-1523, 1983.
Anderson, D. E., Jr., and S. A. Lloyd, Polar twilight UV-visible radi-
ation field: Perturbations due to multiple scattering, ozone deple-
tion, stratospheric louds, and surface albedo, J. Geophys. Res., 95,
7429-7434, 1990.
Anderson, D. E., Jr., R. DeMajistre, and S. A. Lloyd, Impact of
aerosols and clouds on the troposphere and stratosphere radiation
field with application to twilight photochemistry at 20 km, J. Geo-
phys. Res., 100, 7135-7145, 1995.
Atkinson, R., D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr,
J. Troe, Evaluated kinetic and photochemical data for atmospheric
chemistry, suppl. IV, IUPAC subcommittee on gas kinetic data
evaluation for atmospheric chemistry, J. Phys. Chem. Ref. Data, 21,
1125-1568, 1992.
Bauer, R., et al., Monitoring the vertical structure of the Arctic polar
vortex over northern Scandinavia during EASOE: Regular N20
profile observations, Geophys. Res. Lett., 21, 1211-1214, 1994.
Borrmann, S., et al., Aerosols as dynamical tracers in the lower strato-
sphere: Ozone versus aerosol correlation after the Mount Pinatubo
eruption, J. Geophys. Res., 100, 11,147-11,156, 1995.
DeMore, W. B., et al., Chemical kinetics and photochemical data for
use in stratospheric modelling, in Evaluation 10, Tech. Rep. 92-20,
Jet Propul. Lab., Pasadena, Calif., 1992.
DeMore, W. B., et al., Chemical kinetics and photochemical data for
use in stratospheric modelling, in Evaluation 11, Tech. Rep. 94-26,
Jet Propul. Lab., Pasadena, Calif., 1994.
Deshler, T., In situ measurements of Pinatubo aerosol over Kiruna on
four days between January 18 and February 13, 1992, Geophys. Res.
Lett., 21, 1323-1326, 1994.
Evans, W. F. J., C. T. McElroy, and I. E. Galbally, The conversion of
N20 5 to HNO 3 at high latitudes in winter, Geophys. Res. Lett., 12,
825-828, 1985.
Fahey, D. W., S. Solomon, S. R. Kawa, M. Loewenstein, J. R. Podol-
ske, S. E. Strahan, and K. R. Chan, A diagnostic for denitrification
in the winter polar stratosphere, Nature, 21,698-702, 1990.
Fahey, D. W., et al., In situ measurements constraining the role of
sulphate aerosols in mid-latitude ozone depletion, Nature, 363, 509-
514, 1993.
Fried, A., B. E. Henry, and J. G. Calvert, The reaction probability of
N20 5 with sulfuric acid aerosols at stratospheric temperatures and
compositions, J. Geophys. Res., 99, 3517-3532, 1994.
Garcia, R. R., and S. Solomon, A new numerical model of the middle
atmosphere, 2, Ozone and related species, J. Geophys. Res., 99,
12,937-12,951, 1994.
Hanson, D. R., and K. Mauersberger, Laboratory studies of the nitric
acid trihydrate: Implications for the south polar stratosphere, Geo-
phys. Res. Lett., 15, 855-858, 1988.
Hanson, D. R., and A. R. Ravishankara, The reaction probabilities of
CIONO 2 and N20 5 on 40-75% sulfuric acid solutions, J. Geophys.
Res., 96, 17,307-17,314, 1991.
Hanson, D. R., A. R. Ravishankara, and S. Solomon, Heterogeneous
reactions in sulfuric acid aerosols: A framework for model calcula-
tions, J. Geophys. Res., 99, 3615-3629, 1994.
Harwood, R. S., and J. A. Pyle, Studies of the ozone budget using a
zonal mean circulation model and linearised photochemistry, Q. J.
R. Meteorol. Soc., 103, 319-343, 1977.
Harwood, R. S., and J. A. Pyle, The dynamical behaviour of a two-
dimensional model of the stratosphere, Q. J. R. Meteorol. Soc., 106,
395-420, 1980.
Kawa, S. R., D. W. Fahey, K. K. Kelly, J. E. Dye, D. Baumgardener,
B. W. Gandrud, M. Loewenstein, G. V. Ferry, and K. R. Chan, The
Arctic polar stratospheric cloud aerosol: Aircraft measurements of
reactive nitrogen, total water, and particles, J. Geophys. Res., 97,
7925-7938, 1992.
Kawa, S. R., et al., Interpretation of NOx/NOy observations from
AASE II using a model of chemistry along trajectories, Geophys.
Res. Lett., 20, 2507-2510, 1993.
Komhyr, W. D., R. A. Barnes, G. B. Brothers, J. A. Lathrop, and D. P.
Opperman, Electrochemical concentration cell ozonesonde perfor-
mance evaluation during STOIC 1989, J. Geophys. Res., 100, 9231-
9244, 1995.
Kondo, Y., W. A. Matthews, P. Aimedieu, and D. E. Robbins, Diurnal
variation of nitric oxide at 32 km: Measurements and interpretation,
J. Geophys. Res., 93, 2451-2460, 1988.
Kondo, Y., N. Toriyama, W. A. Matthews, and P. Aimedieu, Calibra-
tion of the balloon-borne NO instrument, J. Geomagn. Geoelectr.,
41,507-523, 1989.
Kondo, Y., P. Aimedieu, W. A. Matthews, W. R. Sheldon, and J. R.
Benbrook, A midlatitude balloon-borne observation of total odd
nitrogen, Geophys. Res. Lett., 17, 73-76, 1990.
Kondo, Y., P. Aimedieu, M. Koike, Y. Iwasaka, P. A. Newman,
U. Schmidt, W. A. Matthews, M. Hayashi, and W. R. Sheldon,
Reactive nitrogen, ozone, and nitrate aerosols observed in the Arctic
stratosphere in January 1990, J. Geophys. Res., 97, 13,025-13,038,
1992.
Kondo, Y., U. Schmidt, T. Sugita, P. Aimedieu, M. Koike, H. Ziereis,
and Y. !wasaka, Total reactive nitrogen, N20 , and ozone in the
winter Arctic stratosphere, Geophys. Res. Lett., 21, 1247-1250, 1994.
Lary, D. J., and J. A. Pyle, Diffuse radiation, twilight and photochem-
istry, J. Atmos. Chem., 13, 373-392, 1991.
Loewenste!n, M., et al., New observations of the NOy/N20 correlation
in the lower stratosphere, Geophys. Res. Lett., 20, 2531-2534, 1993.
Meier, R. R., D. E. Anderson, and M. Nicolet, The radiation field in
the troposphere and stratosphere from 240 and 1000 nm: General
analysis, Planet. Space Sci., 30, 923-933, 1982.
Mills, M. J., A. O. Langford, T. J. O'Leary, K. Arpag, H. L. Miller,
M. H. Proffitt, R. W. Sanders, and S. Solomon, Geophys. Res. Lett.,
20, 1187-1190, 1993.
Nash, E. R., P. A. Newman, J. R. Rosenfield, and M. R. Schoeberl, An
objective determination of the polar vortex using Ertel's potential
vorticity, J. Geophys. Res., in press, 1996.
Nicolet, M., R. R. Meier, and D. E. Anderson, The radiation field in
the troposphere and stratosphere from 240 to 1000 nm: Numerical
analysis, Planet. Space Sci., 30, 935-983, 1982.
Ridley, B. A., et al., Stratospheric odd nitrogen: Measurements of
HNO3, NO, NO2, and 0 3 near 54øN in winter, J. Geophys. Res., 89,
4797-4820, 1984.
Ridley, B. A., M. McFar|and, A. L. Schmeltekopf, M. H. Proffitt, D. L.
Albritton, R. H. Winkler, and T. L. Thompson, Seasonal differences
in the vertical distributions of NO, NO2, and 03 in the stratosphere
near 50øN, J. Geophys. Res., 92, 11,919-11,929, 1987.
Schiller, C., A. Hofzumahaus, M. M/iller, E. Klein, E.-P. R6th, and U.
Schmidt, Ultraviolet actinic flux in the stratosphere: An overview of
balloon-borne measurements during EASOE, 1991/1992, Geophys.
Res. Lett., 21, 1239-1242, 1994.
Schubert, S. D., J. Pfaendtner, and R. Rood, An assimilated data set
for Earth Science applications, Bull. Am. Meteorol. Soc., 74, 2331-
2342, 1993.
Thomason, L. W., and L. R. Poole, Use of stratospheric aerosol prop-
erties as diagnostics of Antarctic vortex processes, J. Geophys. Res.,
98, 23,003-23,012, 1993.
World Meteorological Organization, Scientific Assessment of Ozone
Depletion: 1991, Global Ozone Research and Monitoring Project,
Rep. 25, Geneva, 1992.
T. Deshler, Department of Atmospheric Science, University of Wy-
oming, Laramie, WY 82071.
A. R. Douglass and S. R. Kawa, NASA Goddard Space Flight
Center, Code 916, Greenbelt, MD 20771.
M. Koike, Y. Kondo, and T. Sugita, Solar-Terrestrial Environment
Laboratory, Nagoya University, Toyokawa, Aichi 442, Japan.
D. Lary and E. Lutman, Department of Chemistry, University of
Cambridge, Lensfield Road, Cambridge CB2 1EW, England.
(Received September 12, 1995; revised January 31, 1996;
accepted February 1, 1996.)