ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-16-4369-2016Sensitivity to grid resolution in the ability of a chemical transport model to simulate observed oxidant chemistry under high-isoprene conditionsYuKarenkyu@seas.harvard.eduhttps://orcid.org/0000-0003-1307-3738JacobDaniel J.FisherJenny A.https://orcid.org/0000-0002-2921-1691KimPatrick S.MaraisEloise A.MillerChristopher C.TravisKatherine R.https://orcid.org/0000-0003-1628-0353ZhuLeihttps://orcid.org/0000-0002-3919-3095YantoscaRobert M.https://orcid.org/0000-0003-3781-1870SulprizioMelissa P.CohenRon C.https://orcid.org/0000-0001-6617-7691DibbJack E.FriedAlanMikovinyTomasRyersonThomas B.WennbergPaul O.https://orcid.org/0000-0002-6126-3854WisthalerArminhttps://orcid.org/0000-0001-5050-3018School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USADepartment of Earth and Planetary Sciences, Harvard University, Cambridge, MA, USACentre for Atmospheric Chemistry, School of Chemistry, University of Wollongong, Wollongong, NSW, AustraliaSchool of Earth and Environmental Sciences, University of Wollongong, Wollongong, NSW, AustraliaDepartment of Chemistry, University of California, Berkeley, CA, USAEarth System Research Center, University of New Hampshire, Durham, NH, USAInstitute for Arctic and Alpine Research, University of Colorado, Boulder, CO, USADepartment of Chemistry, University of Oslo, Oslo, NorwayEarth System Research Laboratory, National Oceanic and Atmospheric Administration, Boulder, CO, USADivision of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA, USADivision of Engineering and Applied Sciences, California Institute of Technology, Pasadena, CA, USAInstitute for Ion Physics and Applied Physics, University of Innsbruck, Innsbruck, AustriaKaren Yu (kyu@seas.harvard.edu)7April2016167436943785December201518January201626March201629March2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/16/4369/2016/acp-16-4369-2016.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/16/4369/2016/acp-16-4369-2016.pdf
Formation of ozone and organic aerosol in continental atmospheres depends on
whether isoprene emitted by vegetation is oxidized by the high-NOx pathway
(where peroxy radicals react with NO) or by low-NOx pathways (where peroxy
radicals react by alternate channels, mostly with HO2). We used mixed
layer observations from the SEAC4RS aircraft campaign over the Southeast
US to test the ability of the GEOS-Chem chemical transport model at different
grid resolutions (0.25∘× 0.3125∘,
2∘× 2.5∘, 4∘× 5∘) to
simulate this chemistry under high-isoprene, variable-NOx conditions.
Observations of isoprene and NOx over the Southeast US show a negative
correlation, reflecting the spatial segregation of emissions; this negative
correlation is captured in the model at
0.25∘× 0.3125∘ resolution but not at coarser
resolutions. As a result, less isoprene oxidation takes place by the
high-NOx pathway in the model at 0.25∘× 0.3125∘
resolution (54 %) than at coarser resolution (59 %). The cumulative
probability distribution functions (CDFs) of NOx, isoprene, and ozone
concentrations show little difference across model resolutions and good
agreement with observations, while formaldehyde is overestimated at coarse
resolution because excessive isoprene oxidation takes place by the
high-NOx pathway with high formaldehyde yield. The good agreement of
simulated and observed concentration variances implies that smaller-scale
non-linearities (urban and power plant plumes) are not important on the
regional scale. Correlations of simulated vs. observed concentrations do not
improve with grid resolution because finer modes of variability are
intrinsically more difficult to capture. Higher model resolution leads to
decreased conversion of NOx to organic nitrates and increased conversion
to nitric acid, with total reactive nitrogen oxides (NOy) changing little
across model resolutions. Model concentrations in the lower free troposphere
are also insensitive to grid resolution. The overall low sensitivity of
modeled concentrations to grid resolution implies that coarse resolution is
adequate when modeling continental boundary layer chemistry for global
applications.
Introduction
Global simulations of tropospheric chemistry present a major computational
challenge. Chemical mechanisms typically include over 100 coupled species,
with lifetimes ranging from less than a second to more than a year,
interacting with transport on all scales from concentrated emission plumes to
the remote troposphere. This complexity has hindered the inclusion of
detailed tropospheric chemistry in climate models and
generated considerable discussion over the importance of spatial resolution.
Global models typically have a horizontal resolution of ∼ 100 km,
which does not properly resolve chemical gradients and may lead to large
errors from non-linear chemistry and coupling to transport. On the other
hand, increasing resolution is computationally expensive and may require
trade-offs in other aspects of the model. Here we apply a global chemical
transport model (GEOS-Chem CTM) to simulate extensive boundary layer
observations of ozone and related species over the Southeast US from the
SEAC4RS aircraft campaign . We explore how
varying grid resolution from 4∘× 5∘
(≈ 400 × 400 km2) to
0.25∘× 0.3125∘
(≈ 28 × 28 km2) affects model results and the ability
to simulate observations.
Ozone (O3) production is central to driving the complexity of chemical
mechanisms. It is controlled by reaction chains involving hydrogen oxide
radicals (HOx≡ OH + peroxy radicals), nitrogen oxide
radicals (NOx≡ NO + NO2), halogen radicals, volatile
organic compounds (VOCs), and a large ensemble of reservoir and product
species. NO and VOCs are emitted by a wide range of sources, both natural and
anthropogenic. NOx typically has a lifetime of hours while VOCs have
lifetimes ranging from minutes to years . Their
interactions result in a diversity of chemical regimes.
A number of studies have examined the effect of model resolution on ozone
production in urban regions . A typical result is that dilution
from grid averaging causes positive bias in the ozone production efficiency
(OPE) per unit NOx emitted .
showed that premature mixing of urban and background air masses can lead to
either overestimates or underestimates of OPE. Model simulations of ship
plumes indicate particularly large ozone overestimates when mixing NOx
from the plumes into otherwise clean grid cells . On a global scale, found from an
asymptotic error convergence method that grid averaging in a
2.8∘× 2.8∘ model caused a +4 % bias in the
global tropospheric ozone burden, with larger errors on regional scales.
compared global models of varying resolutions and concluded
that artificial mixing of biogenic VOC emissions into coarse grid cells
drives excessive conversion of NOx to organic nitrate reservoirs, leading
to release of NOx downwind under higher OPE conditions and thereby causing
excessive ozone production.
The interaction between NO2 and OH, the dominant NO2 oxidant, can also
lead to biases in model estimates of NO2 tropospheric columns for
comparison to satellite observations. found that large
point sources of NOx titrate OH, producing long-lived NOx plumes.
Dilution of these plumes into coarse resolution grid cells shortens the
NOx lifetime, leading to underestimation of NO2 columns.
found that coarse resolution CTMs tend to underestimate
retrievals of NO2 columns over industrial regions.
GEOS-Chem surface emissions of NOx and isoprene in August 2013 at
0.25∘× 0.3125∘ resolution. See text for emission
inventory references. NOx emissions shown here include contributions from
combustion and soils but not lightning. Gray lines on isoprene panel indicate
flight tracks of the DC-8 during the SEAC4RS campaign. The black line
delineates the Southeast US as used for regional budget analyses in the
text.
The SEAC4RS observations provide an opportunity to investigate the effect
of chemical non-linearity in an environment with very high biogenic VOC
emissions (mainly isoprene) and variable levels of NOx (mostly from
combustion). The interactions between these species define oxidant (ozone and
OH) chemistry. Isoprene is oxidized by the OH radical on a timescale of an
hour to produce isoprene peroxy radicals (ISOPO2). These ISOPO2
radicals may either react with NO to produce ozone (high-NOx pathway) or
react by other channels (low-NOx pathways). Grid averaging may affect the
ability to resolve the different pathways with implications for ozone,
oxidant chemistry in general, and the formation of secondary organic aerosol
(SOA) . It may also affect the ability to simulate
variability in chemical concentrations, especially events at the high tails
of the probability distributions. Higher resolution enables better simulation
of chemical gradients, with some benefit for capturing extreme values
. However, higher model resolution does not always improve
the correlations with observed concentrations because finer modes of variability are
harder to capture than coarser modes .
Methods
The GEOS-Chem simulation for the SEAC4RS period (August–September 2013)
uses GEOS-5 assimilated meteorological data produced by the NASA Global
Modeling and Assimilation Office (GMAO) at
0.25∘× 0.3125∘ horizontal resolution with 72
vertical layers and 3 h temporal resolution (1 h for surface quantities and
mixing depths). Here we use that native resolution as reference for
comparison to the 2∘× 2.5∘ and
4∘× 5∘ resolutions routinely adopted by the
GEOS-Chem user community (http://geos-chem.org) for global
applications. Companion GEOS-Chem studies apply the model with
0.25∘× 0.3125∘ resolution to interpret SEAC4RS
observations of organic nitrates , ozone
, aerosols , and formaldehyde
. uses
2∘× 2.5∘ resolution to simulate isoprene secondary
organic aerosol in SEAC4RS. These references present relevant model
descriptions and evaluation.
The 0.25∘× 0.3125∘ simulation extends over a
continental-scale North America window (130–60∘ W,
9.75–60∘ N). Initial and dynamic boundary conditions are from a
global simulation with 4∘× 5∘ resolution serving as
outer nest . The chemical time step is 10 min at
0.25∘× 0.3125∘ resolution, 30 min at
2∘× 2.5∘, and 60 min at
4∘× 5∘. The transport time step is half the
chemical time step.
The SEAC4RS simulation is based on GEOS-Chem version 9-02
(http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v9-02)
including a detailed mechanism for HOx-NOx-VOC-O3-bromine-aerosol
tropospheric chemistry with 196 chemical species . Isoprene chemistry is updated as described by
. US anthropogenic emissions are from the 2011 National
Emissions Inventory (NEI), with national scaling for 2013 and 60 % downward
correction for NOx as described in . Biogenic
emissions are from the Model of Emissions of Gases and Aerosols from Nature
(MEGANv2.1) with 15 % downward correction for isoprene
emission . Lightning NOx emissions are constrained by
satellite observations as described in . Soil NOx
emissions, including fertilizer, are from with
canopy reduction factors.
Figure 1 shows the emissions of NOx and isoprene over the Southeast US
domain at 0.25∘× 0.3125∘ resolution. Isoprene
emission is mainly from forests and NOx emission is mostly from mobile
sources and power plants . The resulting spatial
segregation between isoprene and NOx emissions has important implications
for whether isoprene oxidation takes place by high-NOx or low-NOx
pathways. We ensure that emission totals are the same at all model
resolutions so that they are not a factor of differences in results. This
required minor scaling of biogenic VOC emissions that depend on environmental
variables.
Comparisons between model and aircraft observations use model output sampled
along the DC-8 flight tracks (Fig. 1), and observations from a 60 s merged
data set. For computing aggregate statistics, such as probability
distributions, we use observations directly from the 60 s data set. For
computing correlations between model and observations, we average data over
model grid cells along individual flight tracks. We focus on daytime
continental data over the Southeast US domain (94.5–75∘ W,
29.5–40∘ N, box in Fig. 1) and within the mixed layer as determined
from lidar measurements aboard the aircraft . Typical mixed
layer heights (10th and 90th percentiles) during the campaign ranged from 600
to 2200 m, with a mean value of 1500 m. The aircraft occasionally targeted
fire plumes and we remove those as diagnosed by measured acetonitrile
concentrations exceeding 225 ppt. Our analysis domain also excludes the
Houston plume, targeted on two SEAC4RS flights (16 and 18 September) and
for which our model would not be expected to give a representative
simulation.
Relationship between NOx and isoprene concentrations in the mixed
layer in SEAC4RS. Observations (60-s average) are compared to the
GEOS-Chem (GC) simulations at different resolutions sampled along the flight
tracks over the Southeast US domain of Fig. 1. Pearson's correlation
coefficients (r) are inset. The top model panels are colored by the
fraction of the first-generation isoprene peroxy radical (ISOPO2) reacting
with NO; a fraction less than 0.5 indicates that low-NOx pathways dominate
for isoprene oxidation. The bottom model panels are colored by the OH
concentration, restricted to midday points (10:30–14:30 local time).
Regional mean values computed along the flight tracks (primarily daytime) are
given inset.
Segregation of high-NOx and low-NOx isoprene oxidation pathways
Oxidant chemistry and SOA formation over the Southeast US in summer is
largely determined by whether isoprene oxidation follows the high-NOx or
low-NOx pathways .
diagnosed the prevalence of different isoprene
oxidation pathways in the GEOS-Chem
(0.25∘× 0.3125∘) simulation of the Southeast US
during SEAC4RS on the basis of the fate of the first-generation ISOPO2
radicals. They found on average that 56 % of ISOPO2 radicals reacted
with NO (high-NOx pathway). The low-NOx pathways mostly involved
reaction of ISOPO2 with HO2 (25 %), ISOPO2 isomerization (15 %),
and reactions with other organic peroxy radicals (4 %). The transition from
high- to low-NOx pathways occurred at an NO concentration of about
60 ppt, corresponding to a NOx
concentration of about 300 ppt.
The effect of the geographical segregation between NOx and isoprene
emissions is illustrated in Fig. 2, which shows the relationship between
NOx and isoprene concentrations along the flight tracks. NOx was
measured by chemiluminescence and isoprene by
proton-transfer-reaction mass spectrometry (PTR-MS) . The
observations show a negative correlation that is captured in the simulation
at 0.25∘× 0.3125∘ resolution but not at coarser
resolutions. The 4∘× 5∘ simulation actually shows
positive correlations, likely reflecting a common sensitivity to stagnation
and resulting high temperatures. We find in the model that NOx
concentrations are strongly correlated with surface air temperature on the
4∘× 5∘ (r=0.66) but not on the
0.25∘× 0.3125∘ scale (r=0.17).
For the SEAC4RS data set sampled along the flight tracks, the
0.25∘× 0.3125∘ simulation finds that 42 % of the
ISOPO2 radicals react with NO. That fraction increases to 44 % for the
2∘× 2.5∘ simulation and 49 % for the
4∘× 5∘ simulation. Over the whole Southeast US
domain (box in Fig. 1), the fraction of ISOPO2 radicals reacting by the
high-NOx pathway is 54 % at 0.25∘× 0.3125∘
resolution and 59 % at the 2∘× 2.5∘ and
4∘× 5∘ resolutions. Our values differ slightly from
because we did not exclude ocean grid cells in the
domain in order to keep comparisons between different resolutions consistent.
Cumulative probability distribution functions (CDFs) of the
[ISOPN][ISOPN]+[ISOPOOH] ratio
measuring the relative importance of the high-NOx pathway vs. low-NOx
pathways for isoprene oxidation. High values of the ratio indicate a dominant
high-NOx pathway. Mixed layer observations from the SEAC4RS aircraft
over the Southeast US are compared to GEOS-Chem output along the flight
tracks at different resolutions. The x axis is a normal probability scale
such that a normal distribution would plot on a straight line.
Cumulative probability distribution functions (CDFs) of NOx,
isoprene, formaldehyde (CH2O), and ozone (O3) concentrations in the
Southeast US mixed layer during the SEAC4RS aircraft campaign.
Observations are compared to GEOS-Chem model values at different resolutions
(0.25∘× 0.3125∘,
2∘× 2.5∘, 4∘× 5∘) sampled
along the flight tracks over the Southeast US domain of Fig. 1. The x axis
is a normal probability scale such that a normal distribution (formaldehyde,
ozone) or lognormal distribution (NOx, isoprene) would plot on a straight
line.
We see that the branching ratio between high-NOx and low-NOx pathways
for isoprene oxidation is much less sensitive to model resolution than would
be expected from the segregation of isoprene and NOx emissions. This is
due to OH depletion under low-NOx conditions prolonging the lifetime of
isoprene and its oxidation products, allowing them to travel to higher-NOx
environments before being oxidized. The bottom panels of Fig. 2 show the
dependence of midday, mixed layer model OH concentrations at different
resolutions on NOx and isoprene concentrations along the SEAC4RS flight
tracks. The depletion of OH under low-NOx, high-isoprene conditions is
apparent and is better represented at the finer resolution: the implication
is that isoprene is more likely to be oxidized under higher NOx conditions
than at its point of emission. Although regional mean OH concentration
(computed along flight tracks) varies with grid resolution, the change is not
monotononic as shown inset in the figure.
The SEAC4RS aircraft payload included measurements of first-generation
isoprene nitrates (ISOPN), a product of the high-NOx pathway, and ISOPOOH,
the principal product of the HO2 pathway . Figure 3
shows a cumulative probability distribution function (CDF) plot of the
[ISOPN] / ([ISOPOOH] + [ISOPN]) ratio in the Southeast US mixed
layer, providing a diagnostic of whether isoprene oxidation proceeds by the
high-NOx pathway (high ratio) or the low-NOx pathways (low ratio). We
see that the 0.25∘× 0.3125∘ simulation is more
consistent with observations than the 2∘× 2.5∘ and
4∘× 5∘ simulations, although the difference is
mainly for the high tail of the distribution.
Chemical variability and bias
Figure 4 shows the simulated and observed CDFs of NOx, isoprene,
formaldehyde, and ozone concentrations sampled along the flight tracks in the
Southeast US mixed layer. Formaldehyde was measured by the Compact
Atmospheric Multispecies Spectrometer (CAMS) and ozone by
chemiluminescence . NOx and isoprene are primary
(directly emitted) and have mean lifetimes against chemical loss of a few
hours and less than an hour, respectively. The NOx distribution is
approximately lognormal while isoprene is better described by a Weibull
distribution. Formaldehyde and ozone are secondary (chemically produced) and
their distributions are more normal. Most of the formaldehyde in SEAC4RS
originated from isoprene oxidation .
We find that, with the exception of formaldehyde, differences between the
different model resolutions are mainly at the high tails of the
distributions. In the case of NOx, the highest resolution model better
captures the high tail in the observations due to urban plumes. In the case
of isoprene the highest resolution model over-predicts the high tail in the
observations, which could reflect errors in the fine structure of MEGAN
emissions or excessive local depletion of OH (the main isoprene sink) as a
result of high isoprene. In the case of ozone, the highest resolution model
only marginally improves the simulation of the high tail in the observations
(industrial plumes downwind of Port Arthur, Texas), for which even
0.25∘× 0.3125∘ may not provide adequate resolution.
In the case of formaldehyde, the highest resolution model actually produces
weaker maxima that are more consistent with observations. Production of
formaldehyde from isoprene oxidation has a higher yield in the high-NOx
pathway than the low-NOx pathways . High isoprene is
associated with low NOx in the observations (Fig. 2) and therefore with a
low formaldehyde yield but this is captured only at the highest model
resolution. This result has important implications for the use of
formaldehyde–isoprene model relationships to infer isoprene emissions from
satellite measurements of formaldehyde columns .
Relationships derived from coarse-resolution models overestimate the yield of
formaldehyde from isoprene emission.
Taylor Diagram for NOx, isoprene, formaldehyde, and ozone
concentration statistics sampled in the Southeast US mixed layer along
SEAC4RS flight tracks (Figs. 1 and 4). GEOS-Chem model results at
different grid resolutions are compared to observations at the corresponding
resolutions. Standard deviation is plotted along the radial coordinate, while
the angular coordinate denotes the Pearson's correlation coefficient between
model and observations. Model standard deviation is normalized to the
observations, so that a value above 1 indicates greater variance than
observed. The open circle located at (1, 1) represents the observations. The
grey arcs represent the root-mean-squared error between model and
observations after the mean bias has been removed.
Except for formaldehyde, the bulk of the distributions shows very little
difference between model results at different resolutions, despite NOx
levels spanning 4 orders of magnitude. This suggests that the coarser
resolutions are adequate for simulating regional averages. Non-linear effects
have been reported in previous model studies at the kilometer scale of urban areas
or pollution
plumes , and these small-scale effects are
not resolved here even at 0.25∘× 0.3125∘
resolution. Nevertheless, the general ability of the GEOS-Chem simulation to
capture the variance in the SEAC4RS observations implies that small-scale
effects are not of major importance at least for the Southeast US.
Figure 5 presents a Taylor Diagram of modeled NOx, isoprene, formaldehyde,
and ozone concentration statistics compared to observations averaged to the
corresponding model grid and timestep. The Taylor Diagram is a concise
graphical summary of how well two patterns match each other in terms of their
correlation, root mean squared difference, and variances
. Comparison of variances for different resolutions is
consistent with the previously discussed information from Fig. 4.
Correlations with observations improve when resolution is increased from
4∘× 5∘ to 2∘× 2.5∘, but
then generally degrade at 0.25∘× 0.3125∘ (except
for NOx). This is because finer-scale features are more difficult to model
. A higher
resolution model will be penalized for placing these features in the wrong
place or time, while a coarse resolution model does not attempt to resolve
them.
Implications for global models
From a global modeling perspective, the ability to simulate local extrema in
a chemical source region is generally not critical and the focus instead is
on simulation of regional means and export to the global atmosphere. From
that standpoint, the general insensitivity of concentrations to model
resolution over the bulk of the distributions suggests that coarse resolution
is adequate for global modeling purposes. We confirm this result by
examination of the regional budgets of total reactive nitrogen oxides
(NOy≡ NOx+ oxidation products) and ozone over the
Southeast US at different model resolutions.
Mean concentrations of reactive nitrogen oxides (NOy) species
over the Southeast US during SEAC4RS. Aircraft observations in the mixed
layer over the Southeast US domain of Fig. 1 are compared to GEOS-Chem model
values at different resolutions. ΣANs refers to the sum of alkyl and
multifunctional nitrates and PAN to peroxyacetylnitrate. Observations are
from T. Ryerson for NOx, R. Cohen for ΣANs, G. Huey for PAN, and
J. Dibb for nitric acid (HNO3) .
Cumulative probability distribution functions (CDFs) of NOx and
ozone (O3) concentrations in the lower free troposphere over the Southeast
US (from mixed layer top to 4 km altitude) during the SEAC4RS aircraft
campaign. Observations are compared to GEOS-Chem model values at different
resolutions sampled along the flight tracks over the Southeast US domain of
Fig. 1. The x axis is a normal probability scale such that a normal
distribution (ozone) or lognormal distribution (NOx) would plot as a
straight line.
Figure 6 shows mean simulated and observed daytime concentrations of NOy
species in the Southeast US mixed layer during SEAC4RS. Total NOy in
the model is within 10 % of observations, at all resolutions, but there is
somewhat more difference in the speciation of NOx oxidation products. The
0.25∘× 0.3125∘ simulation has lower
peroxyacetylnitrate (PAN) than the coarser resolutions, in better agreement
with observations, due to less artificial mixing of NOx and isoprene
emissions as previously pointed out by . All model resolutions
show similar low bias relative to observed total alkyl nitrates
(ΣANs). This low bias is discussed in . The model
has larger nitric acid concentrations at higher resolution because it
resolves better the positive correlation between NOx and OH concentrations
(Fig. 2). We find that the regional mean NOx lifetime with respect to
conversion to nitric acid is 0.66 days for the
4∘× 5∘ model, 0.62 days for the
2∘× 2.5∘ model, and 0.51 days for the
0.25∘× 0.3125∘ model. Although simulated regional
mean OH concentrations do not change monotonically with resolution, the
higher resolution model captures better the association of elevated OH with
NOx.
From a global model perspective, chemical venting from continental source
regions such as the Southeast US is of paramount importance. Figure 7 shows
CDF plots for ozone and NOx in the lower free troposphere (from mixed
layer top to 4 km altitude). Model results are insensitive to resolution,
with differences even smaller than in the mixed layer. As in the mixed layer,
all model resolutions simulate the median of the distributions well, and
differences across model resolution occur mainly at the tails of the
distributions. Overall, we find that grid resolution has little effect on the
export of ozone and its precursors out of the mixed layer.
Conclusions
Production of ozone and organic aerosol in continental atmospheres is highly
sensitive to whether isoprene emitted by vegetation is oxidized by the
high-NOx pathway (where peroxy radicals react with NO) or by the
low-NOx pathways (where peroxy radicals react mostly with HO2). This
distinction between pathways is becoming increasingly relevant in the US as
anthropogenic NOx emissions decrease. In this work, we used SEAC4RS
aircraft observations in the mixed layer over the Southeast US to test the
ability of the GEOS-Chem chemical transport model (CTM) at different
horizontal resolutions (0.25∘× 0.3125∘,
2∘× 2.5∘, 4∘× 5∘) to
simulate the different pathways and the resulting variability in
concentrations. 2∘× 2.5∘ and
4∘× 5∘ are the standard resolutions used in global
GEOS-Chem simulations of tropospheric chemistry while the
0.25∘× 0.3125∘ continental-scale resolution is a
new GEOS-Chem capability.
Emissions of NOx and isoprene are spatially segregated (NOx in urban
centers, isoprene in forests). SEAC4RS observations in the mixed layer
show a negative correlation between isoprene and NOx concentrations that
is captured in the model at 0.25∘× 0.3125∘
resolution but not at coarser resolution. 54 % of isoprene oxidation in the
Southeast US takes place by the high-NOx pathway in the
0.25∘× 0.3125∘ resolution model, as compared to
59 % at the coarser resolutions. Observed ratios of isoprene nitrates
(ISOPN) to isoprene hydroperoxides (ISOPOOH) show segregation between high-
and low-NOx pathways that is better captured at
0.25∘× 0.3125∘ than at coarser resolutions. The
segregation between high- and low-NOx pathways is less than would be
expected from the segregation of NOx and isoprene emissions because OH
depletion in low-NOx environments allows isoprene to travel to
higher-NOx environments to become oxidized.
We examined the ability of the model at different resolutions to simulate the
observed probability distributions of NOx, isoprene, formaldehyde, and
ozone concentrations across the Southeast US. Differences between model
resolutions are mainly at the high tails of the distributions. There is
remarkably little difference for the bulk of the distributions. An exception
is formaldehyde, which is overestimated at coarser model resolution because
more isoprene is oxidized by the high-NOx pathway, with a higher
formaldehyde yield. Inference of isoprene emissions from satellite
observations of formaldehyde columns may be biased high if a
coarse-resolution model is used for the formaldehyde-isoprene relationship.
Spatial correlations between model and observations improve as resolution
increases from 4∘× 5∘ to
2∘× 2.5∘ but then decreases as resolution increases
further to 0.25∘× 0.3125∘. This shows that finer
modes of variability are more difficult to capture in models of commensurate
resolution, consistent with the results of previous studies such as
and .
Increasing model resolution leads to faster conversion of NOx to nitric
acid, because OH correlates positively with NOx, while slowing down
conversion to organic nitrates. However, the effect on the mean budget of
reactive nitrogen oxides (NOy) in the mixed layer is small. Furthermore,
comparisons of NOx and ozone concentrations in the lower free troposphere
indicate no significant sensitivity to model resolution. Previous model
studies have pointed out significant chemical non-linearities at the scale of
urban and industrial plumes, smaller than the
0.25∘× 0.3125∘ resolution used here. However, the
ability of GEOS-Chem to simulate the variability of concentrations observed
in SEAC4RS implies that such small-scale effects are not important on the
regional scale.
The overall relative insensitivity of oxidant chemistry to model resolution
in the challenging environment of the Southeast US suggests that mean
statistics of regional boundary layer chemistry can be simulated adequately
at coarse resolution (such as 2∘× 2.5∘) for global
modeling purposes.
Acknowledgements
We are grateful to the entire NASA SEAC4RS team for their help in the
field. This work was funded by the NASA Atmospheric Composition Modeling and
Analysis Program and by the NASA Tropospheric Chemistry Program.
Jenny A. Fisher acknowledges financial support from a University of
Wollongong Vice Chancellor's Postdoctoral Fellowship. Isoprene measurements
during SEAC4RS were supported by the Austrian Federal Ministry for
Transport, Innovation and Technology (bmvit) through the Austrian Space
Applications Programme (ASAP) of the Austrian Research Promotion Agency
(FFG). Armin Wisthaler and Tomas Mikoviny received support from the Visiting
Scientist Program at the National Institute of Aerospace
(NIA). Edited by: M. C. Facchini
ReferencesArunachalam, S., Holland, A., Do, B., and Abraczinskas, M.: A quantitative
assessment of the influence of grid resolution on predictions of future-year
air quality in North Carolina, USA, Atmos. Environ., 40, 5010–5026,
10.1016/j.atmosenv.2006.01.024, 2006.Atkinson, R. and Arey, J.: Atmospheric degradation of volatile organic
compounds, Chem. Rev., 103, 4605–4638, 10.1021/cr0206420, 2003.Chen, D., Wang, Y., McElroy, M. B., He, K., Yantosca, R. M., and Le Sager,
P.: Regional CO pollution and export in China simulated by the
high-resolution nested-grid GEOS-Chem model, Atmos. Chem. Phys., 9,
3825–3839, 10.5194/acp-9-3825-2009, 2009.Chock, D. P., Winkler, S. L., and Sun, P.: Effect of grid resolution and
subgrid assumptions on the model prediction of a reactive bouyant plume under
convective conditions, Atmos. Environ., 36, 4649–4662,
10.1016/S1352-2310(02)00422-3, 2002.Crounse, J. D., McKinney, K. A., Kwan, A. J., and Wennberg, P. O.:
Measurement
of Gas-Phase Hydroperoxides by Chemical Ionization Mass Spectrometry, Anal.
Chem., 78, 6726–6732, 10.1021/ac0604235, 2006.Davis, D. D., Grodzinsky, G., Kasibhatla, P., Crawford, J., Chen, G., Liu,
S., Bandy, A., Thornton, D., Guan, H., and Sandholm, S.: Impact of ship
emissions on marine boundary layer NOx and SO2 Distributions over
the Pacific Basin, Geophys. Res. Lett., 28, 235–238,
10.1029/2000gl012013,
2001.Esler, J. G., Roelofs, G. J., Köhler, M. O., and O'Connor, F. M.: A
quantitative analysis of grid-related systematic errors in oxidising capacity
and ozone production rates in chemistry transport models, Atmos. Chem. Phys.,
4, 1781–1795, 10.5194/acp-4-1781-2004, 2004.Fiore, A. M., Jacob, D. J., Mathur, R., and Martin, R. V.: Application of
empirical orthogonal functions to evaluate ozone simulations with regional
and global models, J. Geophys. Res., 108, 4431, 10.1029/2002jd003151,
2003.Fisher, J. A., Jacob, D. J., Travis, K. R., Kim, P. S., Marais, E. A., Chan
Miller, C., Yu, K., Zhu, L., Yantosca, R. M., Sulprizio, M. P., Mao, J.,
Wennberg, P. O., Crounse, J. D., Teng, A. P., Nguyen, T. B., St. Clair, J.
M., Cohen, R. C., Romer, P., Nault, B. A., Wooldridge, P. J., Jimenez, J. L.,
Campuzano-Jost, P., Day, D. A., Shepson, P. B., Xiong, F., Blake, D. R.,
Goldstein, A. H., Misztal, P. K., Hanisco, T. F., Wolfe, G. M., Ryerson, T.
B., Wisthaler, A., and Mikoviny, T.: Organic nitrate chemistry and its
implications for nitrogen budgets in an isoprene- and monoterpene-rich
atmosphere: constraints from aircraft (SEAC4RS) and ground-based (SOAS)
observations in the Southeast US, Atmos. Chem. Phys. Discuss.,
10.5194/acp-2016-52, in review, 2016.Gillani, N. and Pleim, J.: Sub-grid scale features of anthropogenic emissions
of NOx and VOC in the context of regional Eulerian models, Atmos.
Environ., 30, 2043–2059, 10.1016/1352-2310(95)00201-4, 1996.Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T.,
Emmons, L. K., and Wang, X.: The Model of Emissions of Gases and Aerosols
from Nature version 2.1 (MEGAN2.1): an extended and updated framework for
modeling biogenic emissions, Geosci. Model Dev., 5, 1471–1492,
10.5194/gmd-5-1471-2012, 2012.Hair, J. W., Hostetler, C. A., Cook, A. L., Harper, D. B., Ferrare, R. A.,
Mack, T. L., Welch, W., Izquierdo, L. R., and Hovis, F. E.: Airborne High
Spectral Resolution Lidar for profiling aerosol optical properties, Appl.
Optics, 47, 6734–6752, 10.1364/AO.47.006734, 2008.Hansel, A., Jordan, A., Warneke, C., Holzinger, R., Wisthaler, A., and
Lindinger, W.: Proton-transfer-reaction mass spectrometry (PTR-MS): on-line
monitoring of volatile organic compounds at volume mixing ratios of a few
pptv, Plasma Sources Sci. T., 8, 332, 10.1088/0963-0252/8/2/314, 1999.Henderson, B. H., Jeffries, H. E., Kim, B.-U., and Vizuete, W. G.: The
influence of model resolution on ozone in industrial volatile organic
compound plumes, J. Air Waste Manage., 60, 1105–1117,
10.3155/1047-3289.60.9.1105, 2010.Hudman, R. C., Moore, N. E., Mebust, A. K., Martin, R. V., Russell, A. R.,
Valin, L. C., and Cohen, R. C.: Steps towards a mechanistic model of global
soil nitric oxide emissions: implementation and space based-constraints,
Atmos. Chem. Phys., 12, 7779–7795, 10.5194/acp-12-7779-2012, 2012.Ito, A., Sillman, S., and Penner, J. E.: Global chemical transport model
study
of ozone response to changes in chemical kinetics and biogenic volatile
organic compounds emissions due to increasing temperatures: Sensitivities to
isoprene nitrate chemistry and grid resolution, J. Geophys. Res.-Atmos., 114,
D09301, 10.1029/2008jd011254, 2009.Jacob, D. J. and Wofsy, S. C.: Budgets of reactive nitrogen, hydrocarbons,
and
ozone over the Amazon forest during the wet season, J. Geophys. Res.-Atmos.,
95, 16737–16754, 10.1029/JD095iD10p16737, 1990.Jang, J.-C., Jeffries, H., Byun, D., and Pleim, J.: Sensitivity of zzone to
model grid resolution – I. Application of high resolution regional acid
deposition model, Atmos. Environ., 29, 3085–3100,
10.1016/1352-2310(95)00118-I, 1995.Kiley, C. M., Fuelberg, H. E., Palmer, P. I., Allen, D. J., Carmichael,
G. R.,
Jacob, D. J., Mari, C., Pierce, B., Pickering, K. E., Tang, Y., Wild, O.,
Fairlie, T. D., Logan, J. A., Sachse, G. W., Shaack, T. K., and Streets,
D. G.: An intercomparison and evaluation of aircraft-derived and simulated
CO from seven chemical transport models during the TRACE-P experiment, J.
Geophys. Res., 108, 8819, 10.1029/2002jd003089, 2003.Kim, P. S., Jacob, D. J., Fisher, J. A., Travis, K., Yu, K., Zhu, L.,
Yantosca, R. M., Sulprizio, M. P., Jimenez, J. L., Campuzano-Jost, P., Froyd,
K. D., Liao, J., Hair, J. W., Fenn, M. A., Butler, C. F., Wagner, N. L.,
Gordon, T. D., Welti, A., Wennberg, P. O., Crounse, J. D., St. Clair, J. M.,
Teng, A. P., Millet, D. B., Schwarz, J. P., Markovic, M. Z., and Perring, A.
E.: Sources, seasonality, and trends of southeast US aerosol: an integrated
analysis of surface, aircraft, and satellite observations with the GEOS-Chem
chemical transport model, Atmos. Chem. Phys., 15, 10411–10433,
10.5194/acp-15-10411-2015, 2015.Kumar, N. and Russell, A. G.: Multiscale air quality modeling of the
northeastern United States, Atmos. Environ., 30, 1099–1116,
10.1016/1352-2310(95)00317-7, 1996.Liang, J. and Jacobson, M. Z.: Effects of subgrid segregation on ozone
production efficiency in a chemical model, Atmos. Environ., 34, 2975–2982,
10.1016/S1352-2310(99)00520-8, 2000.Liu, S. C., Trainer, M., Fehsenfeld, F. C., Parrish, D. D., Williams, E. J.,
Fahey, D. W., Hübler, G., and Murphy, P. C.: Ozone production in the rural
troposphere and the implications for regional and global ozone distributions,
J. Geophys. Res.-Atmos., 92, 4191–4207,
10.1029/JD092iD04p04191, 1987.Mao, J., Paulot, F., Jacob, D. J., Cohen, R. C., Crounse, J. D., Wennberg,
P. O., Keller, C. A., Hudman, R. C., Barkley, M. P., and Horowitz, L. W.:
Ozone and organic nitrates over the eastern United States: Sensitivity to
isoprene chemistry, J. Geophys. Res.-Atmos., 118, 11256–11268,
10.1002/jgrd.50817, 2013.Marais, E. A., Jacob, D. J., Kurosu, T. P., Chance, K., Murphy, J. G.,
Reeves, C., Mills, G., Casadio, S., Millet, D. B., Barkley, M. P., Paulot,
F., and Mao, J.: Isoprene emissions in Africa inferred from OMI observations
of formaldehyde columns, Atmos. Chem. Phys., 12, 6219–6235,
10.5194/acp-12-6219-2012, 2012.Marais, E. A., Jacob, D. J., Jimenez, J. L., Campuzano-Jost, P., Day, D. A.,
Hu, W., Krechmer, J., Zhu, L., Kim, P. S., Miller, C. C., Fisher, J. A.,
Travis, K., Yu, K., Hanisco, T. F., Wolfe, G. M., Arkinson, H. L., Pye, H. O.
T., Froyd, K. D., Liao, J., and McNeill, V. F.: Aqueous-phase mechanism for
secondary organic aerosol formation from isoprene: application to the
southeast United States and co-benefit of SO2 emission controls, Atmos.
Chem. Phys., 16, 1603–1618, 10.5194/acp-16-1603-2016, 2016.Mathur, R., Shankar, U., Hanna, A. F., Odman, M. T., McHenry, J. N., Coats,
C. J., Alapaty, K., Xiu, A., Arunachalam, S., Olerud, D. T., Byun, D. W.,
Schere, K. L., Binkowski, F. S., Ching, J. K. S., Dennis, R. L., Pierce,
T. E., Pleim, J. E., Roselle, S. J., and Young, J. O.: Multiscale Air
Quality Simulation Platform (MAQSIP): Initial applications and performance
for tropospheric ozone and particulate matter, J. Geophys. Res.-Atmos., 110,
D13308, 10.1029/2004jd004918, 2005.Murray, L. T., Jacob, D. J., Logan, J. A., Hudman, R. C., and Koshak, W. J.:
Optimized regional and interannual variability of lightning in a global
chemical transport model constrained by LIS/OTD satellite data, J. Geophys.
Res.-Atmos., 117, D20307, 10.1029/2012jd017934, 2012.
National Research Council: A National Strategy for Advancing Climate
Modeling, Report, National Research Council Board on Atmospheric Science and
Climate, Washington, D.C., 2012.Palmer, P. I., Jacob, D. J., Fiore, A. M., Martin, R. V., Chance, K., and
Kurosu, T. P.: Mapping isoprene emissions over North America using
formaldehyde column observations from space, J. Geophys. Res.-Atmos., 108,
4180, 10.1029/2002jd002153, 2003.Parrella, J. P., Jacob, D. J., Liang, Q., Zhang, Y., Mickley, L. J., Miller,
B., Evans, M. J., Yang, X., Pyle, J. A., Theys, N., and Van Roozendael, M.:
Tropospheric bromine chemistry: implications for present and pre-industrial
ozone and mercury, Atmos. Chem. Phys., 12, 6723–6740,
10.5194/acp-12-6723-2012, 2012.Richter, D., Weibring, P., Walega, J., Fried, A., Spuler, S., and Taubman,
M.:
Compact highly sensitive multi-species airborne mid-IR spectrometer, Appl.
Phys. B, 119, 119–131, 10.1007/s00340-015-6038-8, 2015.Ryerson, T. B., Williams, E. J., and Fehsenfeld, F. C.: An efficient
photolysis
system for fast-response NO2 measurements, J. Geophys. Res.-Atmos., 105,
26447–26461, 10.1029/2000JD900389, 2000.SEAC4RS Science Team: Studies of Emissions and Atmospheric Composition,
Clouds and Climate Coupling by Regional Surveys (SEAC4RS),
10.5067/Aircraft/SEAC4RS/Aerosol-TraceGas-Cloud, 2016.Shrestha, K. L., Kondo, A., Kaga, A., and Inoue, Y.: High-resolution modeling
and evaluation of ozone air quality of Osaka using MM5-CMAQ system, J.
Environ. Sci., 21, 782–789, 10.1016/s1001-0742(08)62341-4, 2009.Sillman, S., Logan, J. A., and Wofsy, S. C.: A regional scale model for ozone
in the United States with subgrid representation of urban and power plant
plumes, J. Geophys. Res., 95, 5731–5748, 10.1029/JD095iD05p05731,
1990.Taylor, K. E.: Summarizing multiple aspects of model performance in a single
diagram, J. Geophys. Res., 106, 7183, 10.1029/2000jd900719, 2001.Tie, X., Brasseur, G., and Ying, Z.: Impact of model resolution on chemical
ozone formation in Mexico City: application of the WRF-Chem model, Atmos.
Chem. Phys., 10, 8983–8995, 10.5194/acp-10-8983-2010, 2010.Toon, O. B., Maring, H., Dibb, J., Ferrare, R., Jacob, D. J., Jensen, E. J.,
Luo, Z. J., Mace, G. G., Pan, L. L., Pfister, L., Rosenlof, K. H., Redemann,
J., Reid, J. S., Singh, H. B., Yokelson, R., Chen, G., Jucks, K. W., and
Pszenny, A.: Planning, implementation, and scientific goals of the Studies
of Emissions and Atmospheric Composition, Clouds, and Climate Coupling by
Regional Surveys (SEAC4RS) field mission, J. Geophys. Res.-Atmos.,
in review, 2016.Travis, K. R., Jacob, D. J., Fisher, J. A., Kim, P. S., Marais, E. A., Zhu,
L., Yu, K., Miller, C. C., Yantosca, R. M., Sulprizio, M. P., Thompson, A.
M., Wennberg, P. O., Crounse, J. D., St. Clair, J. M., Cohen, R. C., Laugher,
J. L., Dibb, J. E., Hall, S. R., Ullmann, K., Wolfe, G. M., Pollack, I. B.,
Peischl, J., Neuman, J. A., and Zhou, X.: NOx emissions, isoprene
oxidation pathways, vertical mixing, and implications for surface ozone in
the Southeast United States, Atmos. Chem. Phys. Discuss.,
10.5194/acp-2016-110, in review, 2016.Valari, M. and Menut, L.: Does an increase in air quality models' resolution
bring surface ozone concentrations closer to reality?, J. Atmos. Ocean Tech.,
25, 1955–1968, 10.1175/2008jtecha1123.1, 2008.Valin, L. C., Russell, A. R., Hudman, R. C., and Cohen, R. C.: Effects of
model resolution on the interpretation of satellite NO2 observations,
Atmos. Chem. Phys., 11, 11647–11655, 10.5194/acp-11-11647-2011, 2011.Vinken, G. C. M., Boersma, K. F., Jacob, D. J., and Meijer, E. W.: Accounting
for non-linear chemistry of ship plumes in the GEOS-Chem global chemistry
transport model, Atmos. Chem. Phys., 11, 11707–11722,
10.5194/acp-11-11707-2011, 2011.Wild, O. and Prather, M. J.: Global tropospheric ozone modeling: Quantifying
errors due to grid resolution, J. Geophys. Res., 111, D11305,
10.1029/2005jd006605, 2006.
Yamaji, K., Ikeda, K., Irie, H., Kurokawa, J., and Ohara, T.: Influence of
model grid resolution on NO2 vertical column densities over East Asia,
J. Air Waste Manage., 64, 436–444, 10.1080/10962247.2013.827603, 2014.Zhang, L., Jacob, D. J., Downey, N. V., Wood, D. A., Blewitt, D., Carouge,
C. C., van Donkelaar, A., Jones, D. B. A., Murray, L. T., and Wang, Y.:
Improved estimate of the policy-relevant background ozone in the United
States using the GEOS-Chem global model with 1/2∘× 2/3∘ horizontal resolution over North America, Atmos. Environ., 45,
6769–6776, 10.1016/j.atmosenv.2011.07.054, 2011.Zhu, L., Jacob, D. J., Kim, P. S., Fisher, J. A., Yu, K., Travis, K. R.,
Mickley, L. J., Yantosca, R. M., Sulprizio, M. P., De Smedt, I., Gonzalez
Abad, G., Chance, K., Li, C., Ferrare, R., Fried, A., Hair, J. W., Hanisco,
T. F., Richter, D., Scarino, A. J., Walega, J., Weibring, P., and Wolfe, G.
M.: Observing atmospheric formaldehyde (HCHO) from space: validation and
intercomparison of six retrievals from four satellites (OMI, GOME2A, GOME2B,
OMPS) with SEAC4RS aircraft observations over the Southeast US, Atmos.
Chem. Phys. Discuss., 10.5194/acp-2016-162, in review, 2016.