The Tropospheric Ozone Production about the Spring Equinox
(TOPSE) Experiment: Highlights and
Preliminary Results
One of the major
research emphases in atmospheric chemistry continues to be the study of the
photochemical and dynamic processes that determine the rates of formation and
loss of oxidants throughout the atmosphere.
Oxidation processes shape the chemical composition of the atmosphere
through transformations that act on the large variety of trace gases emitted
from natural and anthropogenic sources.
Since these processes affect the distribution and trends of radiatively
important trace gases in the atmosphere, there is a potential feedback between
chemistry and climate that has implications for future global change.
Tropospheric
ozone plays a central role in the oxidative chemistry of the troposphere, has
an important impact on the radiative balance of the atmosphere, and is known to
have detrimental effects on human health and agricultural crop production. For these
reasons, understanding the processes that control the origin, trends,
distribution, and effects of tropospheric ozone have been high priority themes
to atmospheric chemistry research over the last few decades.
The processes that
control the concentrations and distribution of tropospheric ozone are a complex
interplay of chemistry and dynamics.
The sources, transformations, and interactions of radical and reservoir
species in the HOx, NOx, and organic carbon families, as
well as ozone itself, play a role in determining local rates of ozone
production and loss. In addition to
strictly chemical and photochemical processes, ozone distributions are affected
at least as significantly by atmospheric dynamics. Ozone distributions are affected by direct advective transport of
ozone and reactive precursors through the troposphere, by
stratosphere/troposphere exchange, and through convective processes which
rapidly transport ozone and precursors to different environments of radiation,
temperature, humidity, aerosol loading, etc.
A feature of
tropospheric ozone that is thought to have significant contributions from both
chemical and dynamic origins is the widely observed springtime maximum of ozone
in the Northern Hemisphere mid latitudes.
The origin of this maximum has been attributed to increased input of
ozone from the stratosphere to the troposphere, but it has also been suggested
that tropospheric photochemical processes may also play a significant role in
producing the maximum.
The TOPSE
investigation combined model studies and simulations with a set of chemical and
photochemical measurements taken over the critical winter-spring transition in
the northern mid-to-high latitude troposphere. The overall goal of the
experiment was to investigate the chemical and dynamic evolution of
tropospheric chemical composition over mid- to high-latitude continental North
America during the winter/spring transition; a particular emphasis was placed
on the springtime ozone maximum in the troposphere. As the experiment
progressed additional effort was used to investigate dramatic ozone depletion
events found over large regions of the Arctic surface layer.


Figure 1.
Interior of NCAR/NSF C-130 aircraft showing TOPSE chemical instrumentations and
exterior at airfield at Churchill, Manitoba, Canada (Feb, 2000).
The experiment took
place from February to May, 2000 in a series of 7 round trip missions from
Colorado to northern latitudes. Using the NCAR C-130 flying laboratory (Figure
1), the TOPSE experiment covered a latitude range from 40 N to near the North
Pole, and from 100 ft
to 25,000 ft. in altitude. The experimental design used in-situ and remote
sensing observations to evaluate the rapid onset and early stages of active
photochemistry in the Northern Hemisphere. These observations provided a unique
characterization of the temporal and spatial distribution of ozone and ozone
precursors over continental North America, and, more importantly, a better
understanding of the primary photochemical and dynamic processes that control
the budgets of radicals and reservoirs in the free troposphere. Measurements of
ozone chemistry in the Arctic boundary layer provided important new data to
document ozone loss processes near the surface. Furthermore, through appropriate comparisons of measurements and
model predictions, the TOPSE experiment is providing important tests and
constraints of photochemical models that will advance the development of global
and regional simulations of tropospheric ozone. Investigators who contribute to experimental and modeling
research activities in TOPSE are listed in Table 1.
Table
1. List of investigators and collaborators in the TOPSE Program.
Name Institution Activity
D.
Blake/N. Blake UC Irvine NMHC/Halocarbons
E.
Browell/W.Grant NASA-Langley Ozone/Aerosol LIDAR
R.
Cohen/J. Thornton UC Berkeley NO2 (LIF)
J.
Dibb UNH Be-7/Be-10/Pb-210
B.
Heikes/J. Snow URI Speciated
Peroxides
R.
Talbot UNH Soluble
Acids/Aerosols
R.
Weber GIT Ultrafine Aerosol
J.
Moody U.Va Traj./remote sensing
J.
Merrill URI Meterology/O3-sondes
D.
Allen/K. Pickering UMd Mesoscale/ 3-D
Y.Wang Rutgers Univ. Modeling
D.
Jacob/M. Evans Harvard 3-D Modeling
B.
Ridley NCAR/ACD NOx,y; in situ O3, Co-PI
F.
Flocke/A. Weinheimer NCAR/ACD PAN/PPN
A.
Fried NCAR/ACD Formaldehyde/Peroxide
L.
Mauldin/F. Eisele NCAR/ACD OH, H2SO4, other (CIMS)
M.
Zondlo NCAR/ACD HNO3 (CIMS)
C.
Cantrell NCAR/ACD HO2/RO2; Mission Co-PI
M.
Coffey/J. Hannigan NCAR/ACD CO, N2O
R.
Shetter NCAR/ACD Photolysis Rates
E.
Atlas NCAR/ACD Mission Scientist, Co-PI
S.
Madronich NCAR/ACD Radiation/Process models
A.
Klonecki/P. Hess NCAR/ACD/ASP Mesoscale Model and Forecast
L.
Emmons/X. Tie/D. Kinnison NCAR/ACD Global Modeling
J.
Bottenheim Environ.
Canada Polar Sunrise 2000
J.
D. Fast Pac.
NW Lab Strat. Folds
S.
Penkett/K. Law/C. Reeves UEA MAXOX/UTLS Ozone
P.
Shepson Purdue Polar Sunrise 2000
E.
Schuepbach U. Berne Jungfraujoch Ozone
J.
Dibb et al. UNH Summit Greenland
Obs.
P.
Newman NASA
GSFC Forecasting/Strat. Ozone
A.
Richter/J. Burrows U. Bremen GOME BrO
G.
Brasseur MPI/NCAR Global Modeling
The analysis and
interpretation of the extremely large and rich data set from the TOPSE missions
is in its very early stage. Some of the
significant observations and inferences are given in the text below, but it
should be appreciated that the full scientific story of the TOPSE experiment
will be based on more complete and detailed analyses than have been done to
date.
Large-scale changes
in lower tropospheric ozone, aerosols and in the Arctic were observed for the
first time over the entire winter-to-spring transition during TOPSE. Several significant features of the ozone
distribution were noted from the measurements: 1) a broad scale ozone increase
in the mid troposphere, including plumes with elevated ozone produced by
photochemical processes in the troposphere; 2) stratospheric intrusions of
containing high levels of ozone (mostly at the highest altitudes sampled by the
aircraft); and 3) extensive areas in the boundary layer where ozone was
depleted to near zero levels. One
example showing all of these features is shown in Figure 2 below.
During TOPSE, ozone
in the Arctic lower troposphere was found to increase at a rate greater than 5
ppbv/mo from early February to mid-May, rising to over 60 ppbv by mid-May
(tropospheric ozone from 60-85o N, 2-5 km increased 5 ppbv/mo
according to the UV Differential Absorption Lidar (DIAL) system data and 6
ppbv/mo using the in situ ozone data.).
Aerosols were found in numerous tropospheric layers
throughout the TOPSE mission, but the thickness and horizontal extent of the
layers increased as the season progressed.
These layers also had photochemically enhanced ozone in them, and the
ozone levels in the layers also increased with the advance of the season. In
situ measurements (by the UHH group) of fine aerosol sulfate mixing ratios were
similar to previous values reported for the springtime Arctic troposphere at ground
level. Mixing ratios of 250 – 600 pptv were widespread throughout the Arctic at
altitudes up to ~7 km. In addition, the influence of Arctic haze on aerosol
sulfate was not confined to high latitudes, contrary to expectations from
previous measurements. The correlation
between fine aerosol sulfate and condensation nuclei (CN) was variable in the
TOPSE region; the correlation ranged from very high in some air parcels to no
correlation in others. Both of these
data sets show complex geographic and vertical distributions that will require
detailed analyses to understand them.
Measurements of
large-scale ozone and aerosol layers with the UV DIAL system provided a unique
opportunity to document the pervasive nature and important role of transport
from urban/industrial regions into the Arctic as a leading mechanism whereby
ozone levels are increased during the spring.
While all northern hemisphere continents are likely source regions for
the pollution, the specific places of origin will have to be determined from
back trajectory calculations.
Stratospheric
intrusions/tropospheric folds (SI/TF) were found to play a relatively minor
role in affecting tropospheric ozone levels during the mission. While SI/TF were observed on 14 of 32 flight
days, they were less frequent than urban/industrial plumes and seemed to
transport less ozone irreversibly into the troposphere than was produced
photochemically in the plumes. (The
months with the most observed events were March and April.)

Figure 2. TOPSE Flight 41.
This image of ozone from the surface to 10.5 km obtained on a flight
from Thule to Winnipeg on May 22, 2000 shows different impacts on tropospheric
ozone. Note a stratospheric intrusion just before 1500 and a pollution plume
between 1630 and 1800. An ozone
depleted layer is measured in the lower 2 km (1400 – 1500) over Baffin Bay.
The first large-scale airborne observations of very low
ozone (<5-10 ppbv) regions just above the surface (0 to 500-2500 m) in the
Arctic were made in March through May over Hudson Bay, Baffin Bay, Davis
Strait, and over the Arctic Ocean. A
virtual surface ozone hole was observed for the first time over much of Hudson
Bay, and over the Arctic Ocean. Surface
ozone loss was observed to occur over a larger region of the arctic than had
been previously recognized.
Ozone loss in these layers is thought to occur from
destruction by natural halogen compounds that become photochemically reactive
during polar sunrise. It is believed that the mechanisms leading to the
destruction of the ozone involves interplay between gas, aerosol and snow pack
chemistry centering on bromine and chlorine. Indeed, measurements of enhanced
levels of gaseous and particulate Br by the UNH group and of NMHC by the UCI
research team demonstrate the presence of reactive halogen processes associated
with the surface ozone depletion. Efforts to understand this process will be
enhanced by the comprehensive suite of chemical and radiation measurements
during TOPSE.
Major differences
between the composition of the boundary layer during Low Ozone Events (LOEs)
and the normal boundary layer were found during TOPSE. The total reactive
oxidized nitrogen (NOy) concentrations are elevated during LOEs due
to increases in the PAN (peroxyacetyl nitrate) and PPN (peroxyproprionyl
nitrate) concentrations. The NOx concentration seems similar in both
normal and low ozone air masses.
Hydrocarbons are significantly depleted during LOEs with the higher
carbon alkanes being relatively more depleted. Ethyne is especially depleted,
and this indicates the influence of reactive bromine. The NO/NO2
ratio is significantly perturbed from a simple steady state involving
photolysis of NO2 and oxidation of NO by O3. H2O2,
CH3OOH and CH2O concentrations vary little between these
cases
To understand this
process, TOPSE investigators at Harvard are employing a model that considers
the gas and aerosol phase chemistry important in the arctic and includes fluxes
from both the ice pack and the free troposphere. The model reproduces some of these
features such as the rapid loss of ozone and the decrease in the hydrocarbon
concentrations. Increases in NOy due to the change of the speciation towards
increased concentrations of PANs is also calculated by the model. The model
also has a perturbed NO/NO2 ratio and this is due to the existence
of BrO and large radical concentrations. The model fails to reproduce some
features. The NOx concentration in the model reduces dramatically as production
of BrNO3 and PANs removes NOx. This is not seen in the
measurements. The model also calculates less H2O2 and CH2O
and more CH3OOH during the LOE than in the normal boundary layer,
again a features not observed in the measurements.
The chemistry of
ozone is driven by a series of photochemical reactions involving nitrogen
oxides, carbon monoxide, and volatile organic carbon compounds (hydrocarbons
and other compounds). TOPSE has
provided a new and detailed view of the distribution and evolution of these
critical ozone precursors that is necessary to fully understand the chemical
evolution. For example, the family of
reactive nitrogen oxides includes compounds that interchange between active
(NO/NO2) and reservoir (e.g., HNO3, PAN) species. A major unexpected finding was the
predominance of PAN as the most abundant NOy species in TOPSE
region. PAN (and related) compounds
account for 50 – 90% of the total oxidized nitrogen reservoir in the
winter/spring atmosphere. Further, the
correlation of PAN with ozone during the spring build-up of ozone tends to
confirm that in-situ photochemical production of ozone is the major process
contributing to the spring maximum in the mid-troposphere.
In addition to the
direct measurement of PAN, TOPSE investigators from UC-Berkeley have used a
newly developed instrument to make high time resolution observations of the sum
of PAN, PPN, N2O5, HNO4 and other NOy
compounds of the same bond energy. Intercomparison of these measurements with
those of PAN and PPN (NCAR/ACD) show that the new approach is reasonably
accurate and precise. Details await a more thorough study. The combination of
these two appraoches to PANs will provide constraints on the abundance of HNO4
and of N2O5, two compounds that play an important role in
the nitrogen oxide budget but have been left as free parameters in most models
because of a lack of observational evidence constraining their abundance.
Investigators at UNH have examined the preliminary NOy
budget for each of the 42 TOPSE flights.
In general, the agreement between measured NOy and NOy
determined from the sum of individual species measurements was quite good (NOy
sum = 1.005 NOy meas. + 32).
The latitudinal distribution of the ratio of individual species to
measured NOy showed trends not captured well by the initial model simulations. The data show the following trends: NOx/NOy was ~ 5% at
high latitudes increasing to 15% at 40 N; HNO3/NOy was
~ 10% at high latitudes increasing to
40% at 40 N; PAN/NOy was ~ 80% at high latitudes decreasing to 50%
at 40 N; and alkyl nitrates/NOy ~ 2% at all latitudes. The large fraction existing as PAN was
unexpected but consistent with its long lifetime at the cold temperatures of
the Arctic troposphere.
Measurements of
non-methane hydrocarbons also provided insight into the evolution of
photochemical oxidation from winter to spring.
Consistent with other measurements, the levels of hydrocarbons decreased
from their wintertime maximum as increase in oxidation rates proceeded into
spring. Details of the rates of hydrocarbon loss will be used to make evaluate
integrated radical abundances over the seasonal transition. Also, the NMHC will
be valuable as airmass tracers to diagnose source characteristics and
long-range transport as more detailed diagnostic studies proceed.
Stratospheric/Troposphere Exchange (STE)
While some of the
TOPSE measurements suggest a minor role for stratospheric ozone in producing
the spring maximum, other observations indicated a more significant role for
STE. Investigators from UNH identified ~ 45 individual samplings of
stratospherically influenced air parcels where O3 and HNO3 were
both significantly elevated during the TOPSE missions. The 7Be aerosol data also
indicated a stratospheric influence in nearly all of these cases, supporting
our inference that the enhanced O3 and HNO3 were not of
anthropogenic origin. These events
occurred over the entire suite of flight profiles during TOPSE, with no bias
toward any sampled geographic region.
Overall 57% of the flights (i.e., 24 of 42) were observed to have
stratospheric influence in the sampled air parcels. The correlation between HNO3 and O3 in the
stratospherically influenced air parcels showed a slope of 3.5 x 10-3
pptv/ppbv (r2 = 0.75), which
is very similar to values previously reported for stratospheric air. These results suggest potentially
significant inputs of O3 and NOy to the middle
troposphere in the middle-to-high latitudes from the stratosphere in late
winter and springtime. However, more
detailed meteorological and computer model analyses is required to quantify
this input.
An essential part of
the TOPSE project has been the interaction between measurements and chemical
transport models used to expand the TOPSE data into a broader, hemispheric
context. This interaction has led to
some significant findings even with models and data in their preliminary
form. Further modeling efforts are
planned. Thus far, two large scale
models have been applied to the TOPSE experiment. – MOZART and HANK. These models are used to assess the
importance of the two processes that are responsible for the build-up of ozone
in the mid-troposphere during winter and spring: transport from the
stratosphere, where ozone mixing ratios are much higher; and production in the
background troposphere from ozone precursors coming from both natural and
anthropogenic sources.
Models that
incorporate the most recent findings relating to atmospheric chemistry and transport
are very useful in testing our understanding of processes taking place in the
atmosphere, with the discrepancies between measured and modeled values often
pointing to areas where more work needs to be done.
MOZART, Model for
OZone and Related chemical Tracers, a global 3D chemical transport model was
run in near-real-time using observed (ECMWF) winds, allowing for immediate
preliminary data-model comparisons at the completion of each mission. These comparisons provided an indication of
the influences on the sampled air masses, such as the long-range transport of
pollutants from Europe and Asia, or clean air from the Pacific. The TOPSE data
set has also provided a unique set of measurements for evaluating the model at
northern high latitudes in winter and spring, and shown that the model can
generally reproduce long-lived pollutants such as CO and C2H6. Comparisons with the data have also helped
uncover some deficiencies in the model.
For example, the dry deposition of O3 over ice was too high
in the model, leading to calculated values of O3 much lower than
observed throughout the free troposphere at high latitudes. The model also underestimates NOx
while overestimating HNO3, leading us to investigate possible
missing processes in the model. Further
analysis of the model results will provide additional understanding of the
reactive nitrogen and ozone budgets during the winter-spring transition;
identifying the influence of the stratosphere, long-range transport, and in
situ chemistry on the observed distributions during TOPSE.
Results from MOZART
show that most of the ozone at high northern latitudes during winter is
transported from the stratosphere. The
increase of ozone through the spring was caused by a combination of increased
transport from the stratosphere in the upper troposphere and, in the lower
troposphere, increased photochemical production from anthropogenic
precursors. These findings are in broad
agreement with inferences based on measurements from the C-130 and from
ozonesondes.
The HANK regional
chemical transport model was designed to run for the conditions and domain
covering the Arctic region. In a novel approach the model simulated the
chemical composition of the troposphere in a "real-time" mode during
the campaign, with the model dynamics being driven by the analyzed winds
obtained from the weather forecast models from the National Center of
Environmental Prediction.
A comparison of the
tracer concentrations simulated by the HANK model against measured values shows
that the model simulated levels of NOx, which is a necessary precursor for
ozone production in the troposphere, are significantly lower than the measured
ones. The sensitivity studies conducted so far indicate that the heterogeneous
removal of NO3 and N2O5 on the surface of
aerosols is the major sink for nitrogen oxides. It is possible, that the rate
of this removal in the model, which reflects our current understanding of this
process, is too high in the Arctic region.
The model was used to
test earlier hypothesis relating to the transport of pollutants into the Arctic
region. In agreement with these earlier studies (Arctic Haze), it was found
that the lower troposphere over the Arctic is exposed to significant levels of
anthropogenic pollutants in the winter and the major transport pathways are
from Europe and Siberia. In the winter, this transport pathway is especially
efficient with the pollutants being transported towards and over the north
pole, and on one occasion, reaching as far south as the Hudson Bay. The model
indicates that during the winter the pollution from the European and Siberian
sources are often trapped in the cold continental boundary layer and can be
transported over large distances into the Arctic. In the Spring, different
meteorological circulation patterns and warming up of the boundary layer, lead
to weaker transport into the Arctic and enhanced mixing of pollutants with the
free troposphere.
Preliminary analysis
of the model results indicates that the tropospheric values simulated by the
model are very sensitive to the ozone mixing ratios in the lower stratosphere,
and therefore to the cross-tropopause transport of ozone. The sensitivity of
tropospheric ozone to the levels of stratospheric ozone actually observed in
the TOPSE region will be examined in more detail through a variety of different
model approaches.