5. Conclusions
In terms of reactivities, by extending the ATom-1 (G2021) to the four
seasons of ATom-1234, we find:
- The tropics dominate production of O3, over at least
the Pacific and Atlantic Ocean basins, with average P-O3 greater than
2 ppb/day throughout 0-12 km. Loss of O3 and
CH4 are also tropical, but primarily at lower
altitudes (0-6 km), and show similar seasonal shifts as P-O3,
following the overhead sun. The northern mid-latitudes add a hot
reactive season in the summer that is comparable to the tropics, but
high reactivities are limited to below 8 km. There is no corresponding
hot season for the southern mid-latitudes.
- Elevated levels of O3 in the tropical Atlantic
O3 are known to be influenced by continental outflow
of biomass burning from primarily Africa and also South America
(Fishman et al., 1990), and here we find that the outflow is still
actively producing O3. High levels of L-O3 occur in a
more tightly constrained region than P-O3, both in latitude and
altitude (0-6 km), indicating a different process.
- The Eastern Pacific stands out: there are large, coherent reactive
regions (20° in latitude by 2-3 km thick) associated with well mixed
convective outflow from North America having extreme reactivities in
both August (monsoonal) and May (biomass burning).
- The Arctic and Antarctic have much lower overall reactivity; and,
given their small area, they play little role in the global
O3 and CH4 budgets.
Our sensitivity analyses using ATom measurements has identified some
clear direction and pitfalls, as well as providing a more robust view of
chemical feedbacks:
- The critical species (S > 0.1) are NOx,
O3, CH4, CO, and H2O.
- While the sensitivities are linear for 0-20% perturbations, there are
large 2nd-order effects for coupled perturbations,
comparable to 1st-order sensitivities. Thus, coupled
perturbations cannot be calculated simply from the sum of the linear
sensitivities.
- Our 24-hour CH4 feedback factor derived from ATom
flights is similar to the global, steady-state feedback factors
derived from global models over the past two decades.
- Our analysis presents quite a different picture of the lifetime of
O3 than is often assumed: if O3increases by 10%, then L-O3 increases by only 2.5% but P-O3
decreases by 5.4% and thus net loss increases by 7.9%. This
lengthens the effective timescale of a O3perturbation, especially since the regions of high P-O3 and L-O3 do
not often coincide.
Probability densities for these critical species from ATom are presented
as possible performance metrics for CCMs.
Comparing H2O with mean profiles is difficult because
of the 3 orders of magnitude change over the troposphere, and thus, we
recommend that relative humidity over liquid water (RHw) be used to
detect model bias. Here we present a clear bimodal distribution of RHw
in the tropics as measured by ATom.
The full ATom data set, including reactive species and the derived
reactivities along with other atmospheric components describing the
origins and processing of the air masses, provides the most extensive
sampling of tropospheric chemistry over the remote ocean basins to date.
The objective flight planning and near-continuous climb/descent
profiling provided full sampling of the 0-12 km troposphere over the
oceans. The statistics, including the co-variations of critical species
(2D probability densities, in G2021), provide an excellent measurement
metric for CCMs. The model-derived reactivities provide a testbed for
the chemistry modules used in CCMs and also for independent analyses of
the origins and chemical evolution of the air that matters, those
chemically hot air masses clearly seen in the ATom flights. The
sensitivity analysis of the 24-hour reactivities provides some core data
that we feel should become a standard part of CCM evaluations and
inter-comparisons. With this analysis, based on 10 s (2 km) air parcels,
we believe we have partially deconstructed the spatial scales and
variability that defines tropospheric chemistry.
Acknowledgements. The authors are indebted to the entire
ATom Science Team including the managers, pilots and crew, who made this
mission possible. We thank the instrument teams (co-authors on the first
paper, Guo et al., 2021) for this valuable data set. Primary funding of
the preparation of this paper at UC Irvine was through NASA grants
NNX15AG57A and 80NSSC21K1454.
Data Availability. The full ATom data set as well as the
derived MDS-2 and RDS*-2 data for ATom 1, 2, 3 and 4 are posted on the
NASA ESPO ATom website (https://espo.nasa.gov/atom/content/ATom). The
final archive for ATom data will be at Oak Ridge National Laboratory
(ORNL), see https://daac.ornl.gov/ATOM/guides/ATom_merge.html. The
MATLAB codes and data sets used in the analysis here are posted on Dryad
(https://doi.org/10.7280/D1Q699), which has been expanded from Guo et
al., 2021 to include this paper’s data and codes.