GMDGeoscientific Model DevelopmentGMDGeosci. Model Dev.1991-9603Copernicus PublicationsGöttingen, Germany10.5194/gmd-10-721-2017The high-resolution version of TM5-MP for optimized satellite retrievals:
description and validationWilliamsJason E.williams@knmi.nlBoersmaK. Folkerthttps://orcid.org/0000-0002-4591-7635Le SagerPhillipeVerstraetenWillem W.https://orcid.org/0000-0002-7222-5212KNMI, De Bilt, the NetherlandsMeteorology and Air
Quality Group, Wageningen University, Wageningen, the NetherlandsKMI, Ukkel, Brussels, BelgiumJason E. Williams (williams@knmi.nl)15February201710272175018May201611July201613December20164January2017This 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://gmd.copernicus.org/articles/10/721/2017/gmd-10-721-2017.htmlThe full text article is available as a PDF file from https://gmd.copernicus.org/articles/10/721/2017/gmd-10-721-2017.pdf
We provide a comprehensive
description of the high-resolution version of the TM5-MP global chemistry
transport model, which is to be employed for deriving highly resolved
vertical profiles of nitrogen dioxide (NO2), formaldehyde (CH2O),
and sulfur dioxide (SO2) for use in satellite retrievals from platforms
such as the Ozone Monitoring Instrument (OMI) and the Sentinel-5 Precursor,
and the TROPOspheric Monitoring Instrument (tropOMI). Comparing simulations
conducted at horizontal resolutions of 3∘× 2∘ and
1∘× 1∘ reveals differences of ±20 % exist in
the global seasonal distribution of 222Rn, being larger near specific
coastal locations and tropical oceans. For tropospheric ozone (O3),
analysis of the chemical budget terms shows that the impact on globally
integrated photolysis rates is rather low, in spite of the higher spatial
variability of meteorological data fields from ERA-Interim at
1∘× 1∘. Surface concentrations of O3 in
high-NOx regions decrease between 5 and 10 % at
1∘× 1∘ due to a reduction in NOx recycling
terms and an increase in the associated titration term of O3 by NO. At
1∘× 1∘, the net global stratosphere–troposphere
exchange of O3 decreases by ∼ 7 %, with an associated shift in
the hemispheric gradient. By comparing NO, NO2, HNO3 and
peroxy-acetyl-nitrate (PAN) profiles against measurement composites, we show
that TM5-MP captures the vertical distribution of NOx and long-lived
NOx reservoirs at background locations, again with modest changes at
1∘× 1∘. Comparing monthly mean distributions in
lightning NOx and applying ERA-Interim convective mass fluxes, we show
that the vertical re-distribution of lightning NOx changes with enhanced
release of NOx in the upper troposphere. We show that surface mixing
ratios in both NO and NO2 are generally underestimated in both low- and
high-NOx scenarios. For Europe, a negative bias exists for [NO] at the
surface across the whole domain, with lower biases at
1∘× 1∘ at only ∼ 20 % of sites. For
NO2, biases are more variable, with lower (higher) biases at
1∘× 1∘ occurring at ∼ 35 % (∼ 20 %)
of sites, with the remainder showing little change. For CH2O, the impact
of higher resolution on the chemical budget terms is rather modest, with
changes of less than 5 %. The simulated vertical distribution of CH2O
agrees reasonably well with measurements in pristine locations, although
column-integrated values are generally underestimated relative to satellite
measurements in polluted regions. For SO2, the performance at
1∘× 1∘ is principally governed by the quality of
the emission inventory, with limited improvements in the site-specific
biases, with most showing no significant improvement. For the vertical
column, improvements near strong source regions occur which reduce the biases
in the integrated column. For remote regions missing biogenic source terms
are inferred.
Introduction
One application of chemistry transport models (CTMs) is to provide accurate
vertical and horizontal global distributions of trace gases such as ozone
(O3), nitrogen dioxide (NO2), sulfur dioxide
(SO2) and formaldehyde (CH2O) that are used as a priori best
guesses in the retrievals of tropospheric abundances from instruments mounted
on Earth-orbiting satellites such as the Tropospheric Emission Sounder (TES;
Worden et al., 2007) Global Ozone Monitoring Experiment (GOME), the SCanning
Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY; De
Smedt et al., 2008), the Ozone Monitoring Instrument (OMI; Boersma et al.,
2011), and GOME-2 (Valks et al., 2011). To date, although high-resolution
regional models have been employed for selected regions such as the US and
Europe (e.g. Russell et al., 2011; Zhou et al., 2012; Vinken et al., 2014),
at the global scale the CTM
resolutions employed are still rather coarse
(between 1.1 and 4.0∘ in latitude and between 1.1 and 6∘ in
longitude), resulting in “footprints” which aggregate hundreds of
kilometres in area. This has limitations as the resulting total columns are
sensitive to topography, surface albedo and the shape of the a priori
vertical profiles themselves. Using rather coarse resolution leads to
substantial errors in the retrievals (e.g. Boersma et al.,
2008;
Heckel et al., 2011; Russell et al., 2011) and imposes limitations on
capturing the regional-scale variability in short-lived trace gas abundances
observed from high-resolution satellite instruments such as the OMI.
This lack of spatial detail is particularly relevant for situations where
strong spatio-temporal variability in the vertical distribution of NO2,
SO2, and CH2O can be expected. Examples include shipping lanes in
the relatively unpolluted marine boundary layer (e.g. Vinken et al., 2014)
and coal-fired power plant SO2 pollution (e.g. Fioletov et al.,
2015). Moreover, during the
day the local lifetime and mixing ratios of trace gases such as nitric oxide
(NO) and NO2 are critically dependent on a host of variables, e.g.
temperature, surface albedo, cloud cover (via photolysis), chemical
conversion (i.e. the NO / NO2 ratio) and the extent of mixing by
convective upwelling (i.e. land type) and advective transport. Thus, the
information provided for the retrievals is affected by the coarsening of the
high-resolution meteorological data used to drive the CTM. Recently, Heckel
et al. (2011) demonstrated that there is an associated uncertainty of
∼ 2 using a priori data from a global CTM rather than a regional CTM,
principally due to loss of spatial information. Two other studies focusing on
the impact of horizontal resolution on the retrieval of vertical column
densities of NO2 suggested that errors of up to ∼ 50 % exist
(Yamaji et al., 2014; Lin et al., 2014). This problem becomes accentuated for
the next generation of Earth-orbiting satellites such as the Tropospheric
Monitoring Instrument (tropOMI), which has a smaller footprint compared to
its predecessors (Veefkind et al., 2012). Applications of TM5 include the
retrieval of NO2, CH2O, and SO2 column densities from the OMI
and tropOMI (e.g. van Geffen et al., 2016), where studies related to the
influence of horizontal resolution have been limited principally to NO2.
The dominant tropospheric loss terms for CH2O are photolysis and
scavenging into cloud droplets (wet deposition; Jacob, 2000). Thus the
atmospheric lifetime of CH2O is highly sensitive to the extent of cloud
cover and the vertical profiles of the photolysis rates. A dominant
application of CH2O retrievals is to provide constraints on tropical and
sub-tropical isoprene emission fluxes (e.g. Palmer et al., 2006; Stavrakou et
al., 2009; Marais et al., 2012). The resulting emission estimates are highly
sensitive to the stoichiometric yield of CH2O from isoprene oxidation,
the chemical lifetime of CH2O and spatial differences in land cover.
Other applications include estimating emissions released during biomass
burning (BB) episodes (Gonzi et al., 2011), whose spatial location is also
smeared via coarsening in TM5-MP. For SO2, which predominantly
originates from point sources, an adequate spatial distribution of such
sources is crucial for estimating accurate biases in existing emission
inventories.
In this paper we provide a comprehensive description of the global,
high-resolution 1∘× 1∘ version of the TM5 CTM
tailored for the application of satellite retrievals (hereafter referred to
as TM5-MP). In Sect. 2 we give details related to the modifications which
have been made to the TM5 model compared to previous versions, the emission
inventories employed, updates that have been made to the modified CB05
chemical mechanism, the stratospheric boundary conditions, the photolysis
scheme, the heterogeneous conversion and the overall model structure. In
Sect. 3 we analyse the impact on convective and advective transport of trace
species from the boundary layer (BL) of both increased horizontal resolution
and use of ERA-Interim convective mass fluxes, as derived using radon
(222Rn) distributions. In Sect. 4 we investigate the effects on regional
and global photolysis frequencies. In Sects. 5–9 we examine the differences
in the vertical and horizontal distributions of tropospheric O3,
NOx, lightning-induced NOx, N-containing species (i.e. nitric acid
(HNO3), PAN and lumped organic nitrates (ORGNTR)), CH2O and
SO2, where we make comparisons against both surface and aircraft
measurements to validate mixing ratios. Finally, in Sect. 10, we present our
conclusions.
Description of TM5-MP
Previous versions of TM5 (TM5-chem-v3.0, Huijnen et al., 2010) included a
two-way nested zooming option as described by Krol et al. (2005). This option
allowed high-resolution simulations to be performed over any pre-defined
regional domain, with boundary conditions being determined by the global
simulation at coarser resolution. Typically, global simulations at
3∘× 2∘ with zoom regions at
1∘× 1∘ were performed to alleviate the long runtime
of a global 1∘× 1∘ run. In the new version of TM5
(hereafter referred to as TM5-MP; the massively parallel version), the usage
of the Message Passing Interface (MPI) has been totally rewritten. Zoom
regions are no longer available, but datasets are distributed along
longitudes and latitudes, instead of model levels and tracers. The advantages
of that overhaul of domain decomposition are a smaller memory requirement and
the possibility of using more processors making global
1∘× 1∘ simulations feasible in terms of runtime and
affordable in terms of computing resources. A TM5-MP global
3∘× 2∘ (1∘× 1∘) run is
∼ 6 (∼ 20) times faster than the previous version of TM5 (Huijnen
et al., 2010) for similar resources. The following model description pertains
to both 3∘× 2∘ and 1∘× 1∘
simulations discussed in this paper.
Here we provide a comprehensive description of the modifications and updates
introduced into TM5-MP compared to TM5 v3.0 (Huijnen et al., 2010). The model
is driven using the ERA-Interim meteorological re-analysis (Dee et al., 2011)
and updated every 3 h, with interpolation of fields for the intermediate
time periods. Although TM5-MP can adopt all 60 vertical levels provided by
the ECMWF ERA-Interim reanalysis, we employ 34 vertical levels for this
study, with higher resolution in the troposphere and upper troposphere–lower
stratosphere (UTLS). Convective mass fluxes and detrainment rates are taken
from the ERA-Interim dataset to describe the updraft velocities from the BL
into the free troposphere, which replaces the parameterization of
Tiedtke (1989) used in previous versions. The vertical diffusion in the free
troposphere is calculated according to Louis (1979), and in the BL by the
approach of Holtslag and Boville (1993). Diurnal variability in the BL height
is determined using the parameterization of Vogelezang and Holtslag (1996).
We use the first-order moments scheme with an iterative time step to prevent
too much mass being transported out of any particular grid cell according to the preservation of the Courant–Friedrichs–Lewy
(CFL) criterion (Bregman et al., 2003), which is especially relevant when
reducing the size of grid cells as done here.
Details of the reaction rate data applied for NOx and nitrogen
reservoirs. The k0 terms are multiplied by the relevant air density to
calculate the correct forward and backward rate constants. The reaction data
and stoichiometery are taken from Atkinson et al. (2004) accommodating the
latest evaluation at http://iupac.pole-ether.fr.
The gas-phase chemistry in TM5-MP is described by an expanded version of the
modified CB05 chemical mechanism (hereafter mCB05; Williams et al., 2013). We
have placed emphasis on updating and expanding the fast NOx chemistry to
account for an accurate partitioning of nitrogen for higher NOx regimes
than those occurring at coarser horizontal resolutions. All reaction rate
data are now taken from the latest IUPAC recommendations (sited at
http://iupac.pole-ether.fr/; last access June 2016) using updated
formulations for third-body collisions, where the rate data for fast NOx
and CH2O chemistry are given in Table 1. This includes the recent update
to the formation rate of HNO3 determined by Möllner et al. (2010).
The most relevant modifications are that (i) the yield of CH2O, methanol
(CH3OH) and the hydro-peroxy radical (HO2) from the
self-termination of the methyl-peroxy radical (CH3O2) is increased
according to Yarwood et al. (2005), (ii) the direct formation of CH2O
from the reaction of CH3O2+ HO2 is added using the
temperature-dependent branching ratio defined in Atkinson et al. (2004),
(iii) the production of HNO3 during the oxidation of di-methyl sulfide
(DMS) by the NO3 radical is now included, (iv) explicit organic peroxy radicals
have been introduced as products from the oxidation of propene
(C3H6) and propane (C3H8) by OH, which are lost either by
the reaction with NO or HO2 allowing the in situ chemical
formation of acetone (CH3COCH3) and higher aldehydes (ALD2),
respectively, following the stoichiometry given in Emmons et al. (2010),
(v) a second product channel for N2O5 photolysis is added
producing NO, (vi) the formation and photo-dissociation of HONO have been
included, (vii) the formation and transport of
methyl-peroxy-nitrate (CH3O2NO2) are also included (Browne et
al., 2011), and (vii) modifications to the gas-phase chemistry involving
NH3 have been introduced following the stoichiometry given in
Hauglestaine et al. (2014). This version of the modified CB05 chemical
mechanism is hereafter referred to as mCB05v2.
The calculation of height-resolved photolysis rates (J values) is performed
using a tailored version of the Modified Band Approach (MBA). The
implementation and performance of this parameterization in TM5 have been
fully described in Williams et al. (2012). For the calculation of the
height-resolved actinic fluxes at the seven specific wavelengths used for
calculating the J values (these being 205.1, 287.9, 302.0, 311.0, 326.5,
385.0 and 610.0 nm), the two-stream radiative transfer solver of Zdunkowski
et al. (1980) is embedded in TM5-MP. For details regarding the
parameterizations used to account for the scattering and absorption
introduced by gaseous molecules, aerosols and clouds, the reader is referred
to Williams et al. (2012). For aerosols, the climatology of Shettle and Fenn
(1979) is included. The calculation of the effective radius (reff)
of cloud droplets is now performed using the approach of Martin et
al. (1994), where different parameter values are used for over the land and
ocean using cloud condensation nuclei concentrations of 40 and 900,
respectively. Due to potentially erroneous values at low horizontal
resolution, we weight the final reff value using the land fraction
in each grid cell. We apply limits between 4 and 16 µm to the
resulting reff values. This improves the representation of the
scattering component due to cloud droplets used for the calculation of the
actinic flux in the lower troposphere (LT; not shown). For the scattering
effects from cloud droplets, we subsequently downsize the physical
reff by ∼ 0.5–2 µm to account for the relationship
between the optical and physical reff values.
For aerosols, an aerosol scheme is available for use within TM5-MP (Aan den
Brugh et al., 2011), but we choose not to use it for the purpose of satellite
retrievals due to the extra computational expense needed when performing
high-resolution simulations that would potentially hinder operational use. We
acknowledge that the description of aerosols in this study is rather crude,
and increasing scattering could have an impact under instances of low cloud
coverage. For the application of TM5-MP to satellite retrieval, it is
preferable to use any advancements in computational performance on further
increases in the horizontal resolution employed. Therefore it is not
currently envisaged that a full description of aerosol processes will be
included during operational satellite retrievals.
However, heterogeneous conversion processes still need the description of the
total reactive surface area density
(SAD) from aerosols. In TM5-MP this is
assumed as the cumulative value of contributions from sulfate, nitrate,
ammonium and methane sulfonic acid as calculated by the EQuilibrium
Simplified Aerosol Model (EQSAM) approach (Metzger et al.,
2002); thus, the secondary
organic aerosol component is not included. The distribution of these aerosol
species is calculated online and coupled to the respective gaseous
precursors. The density of each aerosol type (1.7 g cm-3) and
reff (of between 0.18 and 0.2 µm) is prescribed as in
Huijnen et al. (2014). Swelling at higher relative humidities (> 70 %)
is crudely accounted for by increasing reff between 0.25 and
0.27 µm. The contributions due to sea salt, black carbon and
organic carbon towards heterogeneous loss are not accounted for.
Temperature-dependent gas-phase diffusion coefficients (Dg) are
used in the derivation of the pseudo first-order heterogeneous rate constants
based on the theory of Schwartz (1986).
For N2O5, the uptake coefficient (γ) is calculated using
the parameterization of Evans and Jacob (2005), and is therefore dependent on
both temperature and relative humidity. Once a surface reaction with H2O
occurs, 2 molecules of HNO3 are formed. No uptake on cirrus particles is
included for HNO3, which can lead to de-nitrification of the upper
troposphere (Lawrence and Crutzen, 1998; von Kuhlmann and Lawrence, 2006).
For HO2 we adopt a fixed γHO2= 0.06 across all aerosol
types as taken from Abbatt et al. (2012), and for NO3 we adopt a fixed
γNO3= 10-3 as recommended by Jacob (2000). For
HO2, heterogeneous conversion forms 0.5 molecules of hydrogen peroxide
(H2O2), whereas for NO3 it forms 1 molecule of HNO3
following Emmons et al. (2010). For the SAD associated with cloud droplets we
use the reff values that are calculated by Martin et al. (1994),
thus maintaining consistency between the size of the cloud droplets used for
the scattering component in the calculation of J values and heterogeneous
loss rates on the clouds. By using the ECMWF cloud fraction for each
respective grid cell, we assume that instantaneous mixing throughout the grid
cell does not occur, in order to avoid exaggerated conversion rates on cloud
surfaces.
As TM5-MP contains no explicit stratospheric chemistry, we apply constraints
above the tropopause to ensure realistic stratosphere–troposphere exchange
(STE) of O3 and for constraining the incoming radiation reaching the
troposphere needed for the MBA (Williams et al., 2012). For stratospheric
O3, we use total column values derived from the assimilation of
satellite observations as provided in the improved version of the
Multi-Sensor Re-analysis (MSR, van der A et al., 2010), which is vertically
distributed according to the climatology of Fortuin and Kelder (1998). Three
distinct zonal bands are used for nudging the stratospheric O3 fields,
these being 30∘ S–30∘ N, 30–66∘ S/N and
> 66∘ S/N, where nudging occurs at pressure levels < 45 hPa,
< 95 hPa and < 120 hPa, with relaxation times of 2.5, 3 and 4 days,
respectively.
For stratospheric CH4 we use the monthly 2-D climatological fields
provided by Grooß and Russell (2005), with the nudging heights and
relaxation times being identical to those used for stratospheric O3. For
stratospheric CO and HNO3 we constrain mixing ratios by using monthly
mean ratios of CO / O3 (Dupuy et al., 2004) and
HNO3/ O3 (Jégou et al., 2008; Urban et al., 2009) based on
the latitudinal climatologies derived from ODIN observations using data for
2003/2004 (CO) and 2001–2009 (HNO3). In order to avoid jumps in the
nudging constraints between months, we gradually change between ratios using
the total monthly difference/number of days in the month. These ratios are
applied using the monthly mean stratospheric O3 distribution in TM5-MP,
which is constrained by the MSR dataset (van der A et al., 2010). For both
species, model fields are nudged at 5.5, 10 and 28 hPa using relaxation
times of 5, 10 and 60 days, respectively. Previous versions of TM5 used a
HNO3 climatology from the UARS MLS instrument and applied nudging
constraints at 10 hPa only (Huijnen et al., 2010).
For our study on the impact of horizontal resolution on the performance of
TM5-MP, we present simulations for the year 2006, which has been used for
previous benchmarking studies (Huijnen et al., 2010; Williams et al., 2012).
We use a 1-year spin-up from the same initial conditions, where the initial
conditions are representative of the state-of-the-atmosphere for January 2005
taken from a previous simulation (see Zeng et al., 2015). The model is run
using 34 levels, as it will be used operationally for satellite retrievals,
with details of the pressure levels being given in Huijnen et al. (2010).
Emission inventories
All emission inventories applied in TM5-MP are year-specific, meaning that
the year-to-year variability in emission fluxes due to changes in
anthropogenic activity, biogenic activity and burning extent is taken into
account. For the anthropogenic emission of NOx, CO, SO2, NH3
and non-methane volatile organic compounds (NMVOC), we adopt the MACCity
emission estimates described in Granier et al. (2011). The lack of
sector-specific information complicates the use of daily cycles for e.g. the
road transport component, where a bi-sinusoidal distribution could be
applied, peaking in the morning and late afternoon to represent variability
in traffic volume. Aircraft emissions are included only for NO, using a
homogenous hourly flux estimate not related to regional flight times. For
volcanic SO2 emissions, the estimated emission flux has been scaled up
to 10 Tg S yr-1 based on Halmer et al. (2002). For the biogenic
component, where available we use the CLM-MEGANv2.1 emission inventories
produced for the Southern Hemispheric Multi-model Intercomparison Project
(SHMIP) as described in Zeng et al. (2015), with the missing trace species
(e.g. ethane, propane, higher organics) coming from alternative MEGAN
simulations as outlined in Sindelarova et al. (2014). A diurnal cycle is
imposed on the isoprene emissions and introduced into the first ∼ 50 m
between 20∘ S and 20∘ N, whereas for other latitudes a
continuous daily flux is applied. The BB emissions are taken from the monthly
estimates provided by the GFEDv3 inventory (van der Werf et al., 2010) and
latitude-dependent injection heights and a tropical burning cycle are
implemented following Huijnen et al. (2010). All emission inventories are
provided on a 0.5∘× 0.5∘ resolution and
subsequently coarsened onto the horizontal resolution employed in any
simulation. In TM5-MP all NOx emissions are introduced as NO, rather
than specifying a fraction that is emitted directly as NO2 (Carslaw and
Beevers, 2005). Global NOx emissions for the year 2006 total
49 Tg N yr-1 (including lightning). Other notable species include CO
(1081 Tg CO yr-1), SO2 (117 Tg S yr-1), CH2O
(13.5 Tg C yr-1) and isoprene (510 Tg C yr-1). An overview of
the global and zonal emission terms used in the simulations analysed here is
given in Table 3.
Details of updates made to the reaction data and stoichiometry of
the modified CB05 chemical mechanism for other reactions. Data are taken from
the following: (1) Atkinson et al. (2004) accommodating the latest evaluation
at http://iupac.pole-ether.fr, (2) a branching ratio (R) equal to
1/(1+498.×exp(-1160./T), (3) Yarwood et al. (2005), (4) Sander et
al. (2011), (5) Atkinson et
al. (2006), (6) Emmons et al. (2010), (7) Hauglustaine et al. (2014), (8) a
rate assumed equal to the NH2 analogue, (9) assumed to be equal to
HNO4 after Browne et al. (2011), and (10) E is an estimated value.
The zonally segregated emission totals introduced into TM5-MP for
the year 2006. All organic hydrocarbons are given in Tg C yr-1, except
for CO (Tg CO yr-1), CH2O (Tg CH2O yr-1) and
CH3OH (Tg CH3OH yr-1). All NOx emissions are introduced
as NO (Tg N yr-1). For SO2 emission totals are given as
Tg SO2 yr-1 and for NH3 as Tg NH3 yr-1. No
direct emissions occur for HNO3, PAN, ORGNTR, HONO, N2O5,
NO2, CH3O2NO2 or O3.
The tropospheric chemical budget terms and burden for O3 during
2006 for the 1∘× 1∘ simulation, with all quantities
being given in Tg O3 yr-1. The associated percentage changes are
given when comparing against the 3∘× 2∘ simulation
(equal to (1∘× 1∘) –
(3∘× 2∘)/3∘× 2∘).
The definition of the chemical
tropopause and the calculation of the STE are calculated using the
methodology outlined in Stevenson et al. (2006). The stratospheric nudging
term refers to total change in the mass of O3 in the stratospheric
column when constraining zonal distributions to observational values from the
MSR (Huijnen et al., 2010). The contributions to each term from the SH
extra-tropics/tropics/NH extra-tropics regions (defined as
90–30∘ S/30∘ S–30∘ N/30–90∘ N) are
provided.
For lightning NOx we use the parameterization which uses convective
precipitation fields (Meijer et al., 2001) and constrain the annual global
emission term at ∼ 6 Tg N yr-1. This uses the convective flux
values, meaning that re-scaling of the nudging term was necessary in order to
achieve similar total lightning NOx emissions across simulations. An
example of the resulting horizontal distributions in lightning NOx at
∼ 400 hPa for the tropics for both horizontal resolutions is shown in
the top panel of Fig. S1a in the Supplement. Although the spatial variability
increases at 1∘× 1∘, the global distribution
remains essentially the same, where the constraints on annual lightning
NOx emissions homogenize the total emission flux between resolutions.
One other factor affecting the vertical distribution of lightning NOx
emissions is the convective parameterization which is used. The lower panels
of Fig. S1a show that, at this altitude, the Tiedke (1989) approach increases
the NOx emissions at this level by ∼ 14 %, accompanied by a
significant re-distribution between regions (cf. the Southern Hemisphere (SH)
below 30∘ S and an significant increase in the tropical component in
the Tiedke simulation). However, this is altitude-dependent, where the
absolute differences in the vertical distribution in the monthly NOx
emissions for a selection of latitudes are shown in Fig. S1b. Here
comparisons are shown both over the continents (e.g. 50.5∘ N) and
the oceans (e.g. 59.5∘ S). Although the differences in the
integrated monthly global emission NOx flux are only around
∼ 1–2 %, the temporal and vertical distribution can be quite
different between convective schemes. Profiles show that in the upper
troposphere ERA-Interim consistently results in higher NOx emissions
around 300 hPa, especially for July.
Latitudinal constraints on CH4 global distributions are applied using
the methodology given in Bândǎ et al. (2015) with a 3-day relaxation
time. We also introduce similar constraints based on the appropriate surface
measurements for H2 in order to account for the latitudinal gradient and
variability across seasons, which replaces the fixed global value of 550 ppb
used in previous versions. Finally, for radon (Rn222) emissions we apply
the estimates of Schery (2004), whose global distribution is given in Zhang
et al. (2011).
Observations
Although the performance of mCB05 in TM5 v3.0 has been validated for selected
NMVOC, O3, CH2O, CO and NOy in both hemispheres (Williams et
al., 2013, 2014; Fisher et al., 2015; Zeng et al., 2015), the significant
changes made to both the chemical scheme and the rate parameters in mCB05v2
necessitate independent validation at both 3∘× 2∘
and 1∘× 1∘. We choose a range of ground-based and
airborne measurements taken at diverse locations during the year 2006,
representing different chemical regimes. Here we briefly describe the
observations utilized for this purpose.
For validation of simulated surface concentrations we use measurements of
gaseous O3, NO, NO2, HNO3 and SO2 available from the
European Monitoring and Evaluation program (EMEP, www.emep.int), where
we exploit measurements taken at various background sites in Norway, Finland,
the Netherlands, Belgium, Poland, the Czech Republic, Germany, Great Britain,
Spain, Slovakia, Italy and Portugal. The number of sites used for comparisons
of trace species other than O3 is smaller due to data availability. For
the model composites we extract data from 3-hourly instantaneous output in
order to assemble both the weekly and monthly mean values from the
simulations. For the weekly comparisons of NO2 and SO2 we use
values extracted at 13:00 local time, close to the overpass time of the OMI
instrument (e.g. Boersma et al., 2008). The selected stations allow
validation of the seasonality for both rural regions (FI37) and urban regions
(NL09), where we include identical stations where possible for both species.
For HNO3 we assemble the weekly values from the daily averages.
Measured [O3] in the EMEP network is obtained using UV monitors (Aas et
al., 2001). For all species, spatial interpolation of model data is
performed, accounting for the height of the measurement station and by
weighting using the distance of the station from the surrounding grid cells.
The wide range of measurement sites chosen ensures that both background and
polluted cases are assessed.
For validating the vertical distribution of relevant trace species such as
O3, SO2 and CH2O, we use measurements by the DC-8 aircraft
during the Intercontinental Chemical Transport Experiment B (INTEX-B; Singh
et al., 2009) that took place between March and May 2006. Observations of a
host of co-located nitrogen-containing species are available (namely NO,
NO2, PAN and HNO3). These flights were conducted over a wide
region, and we use all 3 months of measurements. Each month sampled a
different region representing different meteorological conditions and local
emission sources, namely, the Gulf of Mexico (90–100∘ W,
15–30∘ N), the remote Pacific (176–140∘ W,
20–45∘ N) and to the south and west of Alaska over the ocean
(160–135∘ W, 20–60∘ N). Measurements cover altitudes up
to 10.5 km, and we bin the values with respect to pressure using 50 hPa
bins or less in the LT. We interpolated 3-hourly output against measurements
for each respective day, similar to the comparisons performed in previous
evaluations of TM5 (e.g. Huijnen et al., 2010), but we segregate our
comparisons into the three distinct regions. For details relating to the
location of each flight the reader is referred to the campaign overview of
Singh et al. (2009).
For tropospheric O3, we supplement the INTEX-B comparisons with
measurements taken over more polluted regions as part of the Measurement of
Ozone, water vapour, carbon monoxide and nitrogen oxides by Airbus In-service
aircraft initiative (MOZAIC; Thouret et al., 1998). We aggregate the
measurements as seasonal means for December–January–February (DJF) and
June–July–August (JJA) in order to provide a robust number of samples for
each location. Here we choose to use profiles representative of the northern
mid-latitudes, namely, London (0.2∘ W, 51.2∘ N), Vienna
(16.5∘ E, 48.1∘ N), Washington (77.5∘ W,
38.9∘ N), Portland (122.6∘ W, 45.6∘ N), Shanghai
(121.8∘ E, 31.2∘ N) and Tokyo (140.4∘ E,
35.8∘ N).
We also make comparisons of O3, NO, NO2, selected N-reservoir
species, SO2 and CH2O profiles using measurements made aboard the
NOAA WP-3D aircraft as part of the Second Texas Air Quality Study (TexAQS II;
Parrish et al., 2009), which was conducted over the Texas seaboard during
September and October 2006. This allows the assessment of TM5-MP over a
region with higher NMVOC emissions and industrial activity. These
measurements were typically sampled at altitudes below 500 hPa; therefore,
no measurements in the UTLS are available from this campaign.
The effect on atmospheric transport
Here we analyse the differences in convective transport out of the BL by
analysing the vertical and horizontal distribution of 222Rn, which is a
diagnostic typically used for assessing the differences in transport in CTMs
(e.g. Jacob et al., 1997). 222Rn is emitted at a steady rate and
exhibits a half-life of ∼ 3.8 days, which is long enough to be
transported from the BL into the FT due to chemical passivity, with loss via
wet scavenging and dry deposition being negligible. Therefore, it acts as an
ideal tracer to assess differences in convective transport from the surface
out of the BL. The representation of BL dynamics for TM5-MP has recently
been assessed at 1∘× 1∘ using 222Rn
distributions for both the Tiedtke (1989) scheme and when adopting
convective mass transport values from the ERA-Interim meteorological data
(Koffi et al., 2016).
Figure 1 shows seasonal mean horizontal global distributions of 222Rn
for DJF and JJA in the 1∘× 1∘ simulation averaged
between 800 and 900 hPa (i.e. sampling the LT). Also shown are the
associated percentage differences against the re-binned
3∘× 2∘222Rn distribution, allowing a direct
comparison. Resolution-dependent differences result from the cumulative
effects of the use of higher-resolution mass fluxes from the ERA-Interim
meteorological data for describing convective activity and the more accurate
temporal distribution of regional 222Rn emissions at
1∘× 1∘. In general it can be seen that seasonal
differences of ±20 % exist, typically with increases over continents
and decreases over oceans in the 1∘× 1∘
simulations. Maximum differences of > 60 % occur near selected coastal
regions (California, western Africa, Madagascar) or in outflow regions such
as off South America and Africa, where differences exhibit a strong seasonal
dependency. This is due to the large differences in convective strength due
to the variability in heating rates, and thus temperatures, between land and
ocean (e.g. Sutton et al., 2007).
A comparison of the ratio of the monthly mean 222Rn profiles
(1∘× 1∘/3∘× 2∘)
extracted above for selected European cities for January (black) and July
(blue) 2006 is shown in Fig. S3 in the Supplement. The typical tropospheric
profile of 222Rn exhibits an exponential decay from the LT to the FT
(not shown). In order to homogenize the emission flux in the comparison, we
coarsen the 1∘× 1∘ data onto the
3∘× 2∘ grid by averaging the six individual values
into a representative mean column. The extent of the changes in the vertical
distribution of 222Rn is somewhat site-specific, meaning an in-depth
analysis is beyond the scope of this paper. In summary, the
1∘× 1∘ simulation generally provides stronger
convective activity for January, with the main impact occurring below
700 hPa (e.g. London and Paris). The changes in 222Rn in the LT range
between 2 and 10 % (i.e. ratios of 0.9 to 1.1), implying both weaker and
stronger convective transport, depending on changes in location (e.g.
orography and land type). The impact at Berlin is larger than e.g. Barcelona,
which shows that, surprisingly, the inclusion of a large ocean fraction (with
weaker convective mixing) in the 3∘× 2∘ cell does
not seem to introduce dominating effects. Recently Koffi et al. (2016) have
shown that comparisons of 222Rn at coastal sites in Europe at
1∘× 1∘ exhibit significant discrepancies compared
to more continental stations. For July the changes in the vertical
distribution extend into the FT up to 500 hPa, although changes in the upper
FT have a significant component due to changes in long-range transport. The
magnitudes of the changes are similar to those exhibited during January,
although maybe of the opposite sign (e.g. Rome). Thus the influence on e.g.
NO2, CH2O and SO2 a priori vertical profiles will be
non-negligible and diverse.
The seasonal distributions of Rn222 averaged between 800 and
900 hPa for DJF (top) and JJA (bottom) for the
3∘× 2∘ (left) and 1∘× 1∘
(right) simulations, with the associated percentage differences when compared
against the 3∘× 2∘ simulation.
For the tropical cities located in regions where convective mixing is
stronger, the corresponding differences between resolutions can reach ±20 %, especially near the surface (e.g. Caracas and Karachi). There is a
site-specific seasonal dependency in the magnitude of the changes related to
the regional land characteristics (e.g. Lagos versus Kuala
Lumpur) and the extent of the ocean within any
particular grid cell. Thus, differences in a priori vertical profiles of
trace gases using a resolution of 1∘× 1∘ can be
considerable compared to those provided at a 3∘× 2∘
resolution.
We also show ratios of profiles from 1∘× 1∘
simulations using the convective scheme of Tiedtke (1989) against those using
the convective mass fluxes from the ERA-Interim meteorological dataset
(Fig. S4), defined as ERA(1∘× 1∘)/Tiedtke
(1∘× 1∘). For this comparison no daily averaging is
employed, with 222Rn profiles extracted from 3-hourly instantaneous
sampling, with the profiles shown being interpolated directly above urban
conurbations (with high trace gas emissions). The ratios show that the
significant differences exist, with the convective mass fluxes from
ERA-Interim being somewhat weaker than those calculated online using
Tiedtke (1989), i.e. the ratio is typically less than 1, especially during
July. In the recent study by Koffi et al. (2016) performed at
1∘× 1∘ for the European domain, there was no
appreciable improvement in the correlation coefficients when distributions of
222Rn were compared against measurements, resulting in no strong
conclusion about which of the parameterizations results in better atmospheric
transport.
The impact on tropospheric photolysis frequencies
The changes in the spatio-temporal distribution of tropospheric clouds and
surface albedo have the potential to alter the incident flux of photolysing
light reaching the surface, and thus photochemical production and destruction
terms. When present, clouds dominate the integrated optical density in the
tropospheric column. TM5-MP uses a random overlap method for determining the
impact of clouds on actinic flux, which is weighted by cloud cover (Williams
et al., 2012). Comparing seasonal mean cloud coverage for DJF and JJA
(Fig. S5), we show that there are significant increases in the fractional
cloud cover (fcc) at 1∘× 1∘, resulting in fcc
values ranging between 0.1 and 0.8 (cf. 0.1–0.5 for
3∘× 2∘). Moreover, the definition of tropical
equatorial cloud systems becomes much more defined and there are significant
differences in the cloud distributions around the western coast of South
America. For DJF, the largest changes occur at high latitudes over the tundra
and oceans, but correspond to low-intensity incident radiation due to the
polar winter. For the SH, the seasonal fcc increases significantly, which
will potentially impose effects on Antarctic oxidative capacity (see
Sect. 5). For JJA, most increases in fcc do not occur directly above
high-NOx sources, but rather over the oceans. This limits the impact on
the lifetime of chemical precursors (e.g. NO2) as discussed in Sect. 6.
Examining similar plots for surface albedo (not shown) reveals that the
maximum differences (increases at 1∘× 1∘) again
occur in the polar regions under low temperatures related to sea ice and snow
coverage, typically during polar winters. For mid-latitudes and tropics,
although differences in the absolute albedo value can be significant
(±50 %), values are typically below 0.1, which will contribute to the
perturbations in the final J value tropospheric profiles as discussed
below. The monthly mean comparisons in surface J values provided in Fig. S6
show that any differences in instantaneous cloud cover are moderated to the
order of a few percent when looking at longer periods.
The similarity in the monthly mean photolysis frequencies for O3 and
NO2 across resolutions (hereafter denoted JO3 and
JNO2, respectively) are shown in Fig. S6 of the Supplement.
Comparisons of the monthly mean JO3 and JNO2 values are
shown at five different locations identical to those shown in Williams et
al. (2012). For JO3 the impact of increasing resolution is limited
to a few percent in the monthly mean values, even for regions which have high
surface albedo. At the global scale this leads to a reduction of
∼ 2 % in the total mass of O3 photolysed (not shown). For
JNO2, the corresponding differences become more appreciable, with
1∘× 1∘ exhibiting ∼ 5–10 % higher values
at high northern latitudes (associated with locations with high-NOx
regimes). Focusing on JNO2 and comparing seasonal mean values near
the surface show that very similar large-scale spatial patterns occur for
both simulations at the global scale (cf. Fig. S7). The highest
JNO2 values occur over the tropical oceans and high-altitude
regions (e.g. Nepal), with a latitudinal shift related to seasonal changes in
daylight hours. Although a more regional fine structure can be seen at
1∘× 1∘ (e.g. the south-western US and south-western
China for DJF), these seasonal averages show that the small perturbations in
JNO2 shown in Fig. S6 extend to the global scale, leading to only
modest changes in the tropospheric lifetime of NO2 (see Sect. 6).
Comparisons of monthly mean vertical profiles of JO3 and
JNO2 as sampled over selected tropical cities are shown in
Fig. S8a and b, respectively, in the Supplement. Here no averaging is
performed towards an identical horizontal resolution; therefore, values are
representative of the J values directly above the selected urban centres.
The JO3 profiles are affected to a larger extent than the
JNO2 profiles, due to the characteristic absorption spectra of
each species which make JO3 more sensitive to the additional
scattering introduced due to clouds. Profiles over Dubai act as a proxy for
clear-sky conditions, where values of unity exist in the residual of
JO3 and JNO2 calculated through most of the column. The
small difference at the surface is due to changes in the surface albedo
between resolutions, with Dubai being situated on the coast, meaning that a
sharp horizontal gradient exists in surface albedo. For other cities, the
largest perturbations occur away from the surface (e.g. Jakarta, Nairobi and
Lagos) around the altitude where tropospheric clouds are most abundant. There
are typically changes of between ±5 and 10 % in the monthly mean
profiles. The changes in JNO2 reflect those simulated for
JO3, with somewhat smaller perturbations.
Implications for oxidative capacity and tropospheric
O3
The partitioning of reactive N between the short- and long-lived chemical
N-reservoirs included in TM5-MP depends on the oxidative capacity simulated
for the troposphere via competition between the various different radicals
(i.e. OH, CH3C(O)O2, NO3 and CH3O2). Therefore,
changes to the distribution and resident mixing ratios of tropospheric
O3 subsequently impose changes in the fractional composition of the
NOy budget (Olszyna et al., 1994) and also the efficiency of the
NOx recycling terms by altering the chain length (Lelieveld et al.,
2002). In this section we analyse the global and zonal chemical budget terms
for tropospheric O3 to highlight the inter-hemispheric differences which
occur (i.e. under low- and high-NOx environments). An overview of the
resulting near-surface global distribution of tropospheric O3 for
May 2006 is shown in Fig. 3, which also includes the location of the regional
comparisons presented below in a larger global context. In general, the
pattern of minimum and maximum mixing ratios in O3 occurs in similar
locations, with the long-range transport component being more clearly defined
in the 1∘× 1∘ simulation. There is a distinct
latitudinal gradient in O3 mixing ratios imposed by the global
distribution of NOx emissions.
Table 4 provides the zonally segregated chemical budget terms for
tropospheric O3, from which the global component due to STE can be
determined by closing the budget terms following the methodology given in
Stevenson et al. (2006). The chemical tropopause calculated for
3∘× 2∘ is applied for the analysis of
1∘× 1∘ budget terms to ensure that a valid
comparison is performed, (i.e. the same mass of air is accounted for). For
computational efficiency the budget terms are aggregated in 10∘
latitudinal bins for each vertical level and summed across all longitudes,
providing the cumulative latitudinal terms.
The most significant change with resolution concerns STE. By using a
dedicated tagged stratospheric O3 tracer (which only undergoes
photo-chemical destruction and deposition in the troposphere; hereafter
denoted as O3S), changes in the zonal mean STE can be determined. The
stratospheric burden of O3 (BO3 (strat)) exhibits a strong
hemispheric gradient with much more downwelling occurring in the Northern
Hemisphere (NH), peaking during boreal
springtime. At the global scale the STE exchange is
579 Tg O3 yr-1, which agrees well with the multi-model mean for
STE of 556 ± 154 Tg O3 yr-1 in Stevenson et al. (2006),
with observational estimates being
∼ 550 ± 140 Tg O3 yr-1 (Olsen et al., 2001). The
∼ 7 % reduction of STE at 1∘× 1∘ is
encouraging considering that previous studies using TM5 have concluded that
STE in TM5 at 3∘× 2∘ was biased high compared to
STE inferred from TES and MLS satellite observations (Verstraeten et al.,
2015). The increase in STE in the SH, with an associated decrease in the NH
(see below), implies that there is a shift in circulation patterns at
1∘× 1∘ even though BO3 (strat) remains
essentially unchanged. Previous studies have shown that in order to resolve
the correct spatial and temporal stratosphere–troposphere flux, high
resolution is required both in the horizontal and the vertical gridding (e.g.
Meloen et al., 2002). The NH STE diagnosed with TM5-MP is an order of
magnitude smaller than estimates derived in a CTM study also conducted at a
1∘× 1∘ resolution (Tang et al., 2011;
∼ 200 Tg O3 yr-1), which identified deep convection as
important for STE. Here we use a different vertical grid and meteorological
dataset to drive TM5-MP, both of which affect the ability to capture an
accurate STE flux (Meloen et al., 2002). For the
1∘× 1∘ simulation using the Tiedtke scheme, there
is a further reduction in the STE component of 21 Tg O3 yr-1,
resulting in an STE component almost identical to the multi-model mean in
Stevenson et al. (2006).
Zonal mean seasonal distribution of the TM5-MP
O3S / O3 ratio for the 3∘× 2∘ (left)
and 1∘× 1∘ (right) simulations.
The near-surface distribution in tropospheric O3 (top) and
NO2 (bottom) for May 2006 from the 3∘× 2∘
(left) and 1∘× 1∘ (right) TM5-MP simulations. The
blue points represent the location of the MOZAIC airports used for
comparisons. Also shown are the locations of the INTEXB and Texas-AQSII
measurement campaigns, and the extent of the EMEP network in the European
domain, used for the validation of the resulting O3 and NO2
distributions.
The zonal seasonal means of the fraction of O3S to O3
(O3S / O3) for both simulations are shown in Fig. 2 for DJF and
JJA. There is a clear seasonal zonal shift in the fractional contribution due
to the O3 transported downwards from the stratosphere exhibiting a
longer lifetime in the winter hemisphere, reflecting a lower photochemical
destruction rate. At 1∘× 1∘ the largest increase in
STE occurs in the SH during JJA. Here ∼ 20–25 % of tropospheric
O3 is transported down from the stratosphere. Comparing the 0.2 contour
for the NH mid-troposphere shows significant changes, extending further down
towards the surface during boreal wintertime, leading to the higher total
mass of O3S in the troposphere. The extent of nudging towards the MSR
climatology is essentially constant across simulations (cf. Table 4).
Interestingly, less O3S reaches the surface in the tropics at
1∘× 1∘ due to the enhanced chemical destruction
term in the free troposphere. Approximately 10 % of the global deposition
term for O3 is associated with O3 that originates from the
stratosphere at 1∘× 1∘ (cf. ∼ 5 % at
3∘× 2∘). For the NH, this contributes to the
simulated increase in deposition of ∼ 9 %.
For tropospheric O3 there are similarities that occur between the NH,
tropics and SH, i.e. high- and low-NOx scenarios, resulting in a
cumulative decrease in O3 production of ∼ 2–4 % across zones.
For the chemical loss terms there is a modest decrease of ∼ 3 %
(∼ 2 %) in the NH (SH) reflective of the changes discussed for
JO3, which acts as the primary destruction term. Therefore, in the
SH the significant differences shown for fcc do not significantly impose a
lower photochemical destruction term on the annual tropospheric O3
budget. There is a zonal gradient in the tropospheric burden of O3
(BO3) following the zonal gradient in NOx emissions (cf. Fig. 3).
Comparing terms shows that BO3 decreases at
1∘× 1∘ by a few percent at the global scale
(∼ 7 Tg O3), making a rather small impact on oxidative capacity.
This is of the same order of magnitude as that found in previous studies
concerned with horizontal resolution (e.g. Wild and Prather, 2006).
Interestingly, changes in the deposition flux of O3 are rather small,
even though there is a larger amount of variability in the land surfaces and
a better-resolved land–sea contrast at 1∘× 1∘,
although differences in regional deposition fluxes can be more significant.
Multi-model inter-comparisons of surface deposition terms across models have
shown previous versions of TM5 to be at the low end of the model spread in
terms of O3 (Hardacre et al., 2015), suggesting that the surface
deposition flux should be increased by ∼ 10 % in TM5-MP towards the
multi-model mean value. This can be partly attributed to the large
uncertainty which exists related to the loss of O3 to the ocean
(Hardacre et al., 2015).
Comparisons of the seasonal variability in TM5-MP mass mixing ratios
for surface O3 against composites of measurements taken across the EMEP
monitoring network for 2006. Both the co-located TM5-MP
3∘× 2∘ and 1∘× 1∘ monthly
mean values are shown, along with the 1σ variability for Finland, the
Netherlands, Belgium, Poland, Slovakia and Italy. Individual stations that
are aggregated are given in the panels.
Figure 4 shows comparisons of simulated and observed mass mixing ratios of
surface O3 at EMEP sites across Europe (www.emep.int; Aas et al.,
2001), with stations chosen to cover a range of latitudes. Previous
comparisons using mCB05 have revealed high biases in surface O3,
especially during boreal summertime (Williams et al., 2013). These high
biases originate from cumulative effects associated with the accuracy of the
emission inventories, the convective and turbulent mixing component, the
underestimation of the scattering and absorption of photolysing light due to
aerosols and the chemical mechanism that is employed. For the emission
component it should be noted that even at 1∘× 1∘
coarsening is performed, where emission inventories are typically supplied at
0.5∘× 0.5∘ resolution. The seasonal cycle in
surface O3 is captured to a large degree, and the high bias exhibited by
the model is generally reduced by ∼ 2–5 ppb (or ∼ 20 %) at
1∘× 1∘. This is associated with perturbations in
the NOx recycling terms, chemical titration by NO, changes to the
turbulent diffusion and convective mixing out of the BL. The improvement in
biases is largest during boreal summertime, associated with the shorter chain
length of the NOx recycling term during boreal wintertime. However,
there is still a significant monthly mean bias in both simulations when
compared against observations throughout the year, especially for locations
impacted by a large anthropogenic NOx source. This is partly due to the
low NO / NO2 ratio as discussed in Sect. 6 below.
Comparing vertical profiles from composites assembled from the MOZAIC
measurements for DJF and JJA (Figs. S9a and S8b, respectively), INTEX-B
(Singh et al., 2009; Fig. S10) and TexAQS II (Parrish et al., 2009; Fig. S11)
show consistently that differences are small between simulations across
regions, and typically mimic those which occur at the surface. There is a
general positive bias of ∼ 20–40 % in mixing ratios
exhibited across all comparisons, although the variability in the vertical
gradients across regions is captured rather well. Such positive biases have
consequences for both the NOx recycling terms and HNO3 formation
discussed in the sections below.
Implications for the distribution of NO and NO2
The annual NO to NO2 recycling terms involving peroxy radicals
given in Tg N yr-1 for 2006 at 1∘× 1∘
resolution. In mCB05v2 XO2 represents lumped alkyl-peroxy radicals
(Yarwood et al., 2005). The NO + RO2 term is an aggregate of numerous
specific peroxy-radical conversion terms in the modified CB05 mechanism
(Williams et al., 2013; Tables 1 and 2). Also provided are the approximate
percentage differences when comparing with 3∘× 2∘
(equal to (1∘× 1∘) –
(3∘× 2∘)/3∘× 2∘). The
chemical tropopause is defined using the methodology outlined in Stevenson et
al. (2006).
Table 5 provides the zonally segregated annual NOx recycling terms
involving the main peroxy radicals and the direct titration term involving NO
for the 1∘× 1∘ simulation. The conversion rate of
NO back into NO2 decreases by ∼ 2–3 % across zones as a
consequence of an associated increase in the titration term and
re-partitioning of N into long-lived reservoir species (see below). For the
titration term involving NO, although the globally integrated flux remains
relatively constant, there is contrasting behaviour for the two most
important zones (TR, NH), which exhibit a lower and higher titration term,
respectively. It has been shown that for regions such as Europe the increased
titration results in lower surface O3 mixing ratios (cf. Fig. 4),
improving the boreal summertime high bias at the surface.
Important model uncertainties include the quality of the MACCity NOx
emission inventory, the lifetime of NO2 simulated in TM5, BL mixing and
the NOx recycling term via the chemical titration of O3. Figure 3
provides an illustration of the global distribution in surface NO2
during May 2006 for both the 3∘× 2∘ and
1∘× 1∘ simulations, where the short lifetime means
that the maximum mixing ratios occur directly near the strong source regions.
Most NOx is anthropogenic in origin; therefore, there is a strong
latitudinal gradient between the NH and SH, with ship tracks also visible.
The regions where validation occurs are also superimposed in the figure,
including the extent of the EMEP domain over which NO2 weekly
comparisons are made.
Comparison of TM5-MP weekly [NO] sampled at 13:00 UT each day
during 2006 with observed [NO] (µg m-3). The selected sites
shown are in the Czech Republic (top left), Spain (top right), Great Britain
(bottom left) and the Netherlands (bottom right).
Comparison of weekly TM5-MP [NO2] sampled at 13:00 UT each day
during 2006 with observed [NO2] (µg m-3). The selected
sites shown are in the Netherlands (left) and Spain (right).
Figures 5 and 6 show comparisons of weekly [NO] and [NO2] surface
measurements with the corresponding composites from both simulations, sampled
at 13:00 local time, which is close to the local overpass time for both OMI
and tropOMI. Although the number of EMEP sites conducting NOx
measurements is smaller than those measuring O3, we choose stations
located throughout Europe in both high- and low-NOx regimes. To
supplement these comparisons we provide the seasonal mean biases for DJF and
JJA from both simulations in Tables 6 and 7, respectively, calculated using
weekly binned data from all EMEP sites that measure hourly [NO] and
[NO2]. Here we perform an analysis across sites rather than focusing on
the behaviour at selected individual locations.
The seasonal mean absolute biases as calculated using weekly [NO]
values (µg m-3). The weekly means are composed from daily
measurements taken at 13:00 for DJF and JJA (given as the difference in the
measurements–model). Values are shown for both the
3∘× 2∘ and 1∘× 1∘
simulations for all stations with available data. Those with differences
< 5 % are considered to exhibit no discernible change in the bias.
For the determination of [NO2], the reduction of NO on a Molybdenum
convertor takes place with subsequent detection by chemi-luminescence, with
an associated detection limit of ∼ 0.4 ppb. Previous studies have
shown that some bias can result due to the oxidation of nitrogen reservoirs
such as PAN (Dunlea et al., 2007; Steinbacher et al., 2007). In TM5-MP all
NOx emissions are introduced as NO, although a fraction for road
transport is known to be emitted directly as NO2 (e.g. Carslaw and
Beevers, 2005). Many studies have been performed comparing satellite NO2
columns with model values, implying that inadequacies in emission inventories
are somewhat region-specific (e.g. Zyrichidou et al., 2015; Pope et al.,
2015).
Table 6 shows a negative bias of a few µg m-3 in TM5-MP in
seasonal surface [NO] in Europe. This is a cumulative effect of the accuracy
of the MACC NOx emission estimates, an overestimate in daytime vertical
mixing (Koffi et al., 2016) (enhanced dilution) and, to a lesser extent,
overly high surface [O3] (increasing the oxidation rate of NO to
NO2). As anthropogenic emissions are the principle source of NO, there
is no significant seasonal cycle in the monthly emission estimates in the NH.
Seasonal differences in convective mixing (i.e. lower BL heights) do cause
somewhat higher surface [NO] during DJF for approximately equal emission
terms. This is captured by TM5-MP, although under night-time conditions
TM5-MP has been shown to overestimate nocturnal BL heights (Koffi et al.,
2016). For ∼ 80 % of the EMEP sites we do not observe any significant
change in the quality of the comparisons. For ∼ 20 % of sites,
simulations of [NO] at 1∘× 1∘ introduce significant
improvements over those at 3∘× 2∘, and there is an
improvement regarding the extent of seasonal variability (Fig. 4).
Table 7 shows that for [NO2] the biases are more variable, being
typically in the range of ±0–6 µg m-3, with both
positive and negative biases occurring across sites. The conversion
efficiency from NO, loss to reservoir compounds (e.g. HNO3),
photo-dissociation rate, convective mixing and emission estimates
all contribute to these biases. The seasonal biases show
improvements at 1∘× 1∘ for ∼ 35 % of the
EMEP sites, accompanied by degradations at ∼ 20 % of the sites. The
maximum biases in [NO2] at 1∘× 1∘ can be
approximately double those for [NO]. For the corresponding NO / NO2
ratio, there will generally be an underprediction in the model due to the
negative biases shown for the [NO] comparisons. Analysing the corresponding
seasonal correlation coefficients (not shown) shows that in ∼ 25 % of
the cases there is little seasonal correlation between the weekly [NO2]
in TM5-MP and the measurements regardless of resolution for both seasons
(Pearson's r in the range -0.3 to 0.3). In ∼ 30 % of cases there
is actually a degradation in r between resolutions, the changes somewhat
reflect those seen in the seasonal biases, i.e. simultaneous changes to both
the meteorology and local emission fluxes do not necessarily improve the
performance of the model. Comparing 1∘× 1∘ values
both with and without the Tiedtke convection scheme shows that for the most
convective regions (e.g. south of 45∘ N) increases in r generally
occur during JJA when employing the ERA-Interim mass fluxes. Conversely for
e.g. Finland the correlation becomes worse.
Monthly mean comparisons of NO (left), NO2 (middle) and the
resulting NO / NO2 ratio (right) from INTEX-B measurements and
TM5-MP simulations. The dotted line represents the 1σ deviation in
the mean of the measurements. For details of the locations for each month the
reader is referred to Singh et al. (2009).
Beyond Europe, we compared monthly mean TM5-MP vertical distributions of NO
and NO2 between March and May 2006 against measurements taken during the
INTEX-B campaign in Fig. 7. In general differences between
1∘× 1∘ and 3∘× 2∘
simulations are of the order of a few percent, with NO2 biased low in
the LT by ∼ 70–80 %. This is partially associated with the take-off
and landing of the aircraft from polluted airfields, where point sources of
high anthropogenic emissions cannot be resolved even at
1∘× 1∘. For March, there is a strong signature from
biomass burning plumes in the middle troposphere which is not captured using
the monthly burning estimates. For the FT, TM5-MP captures the observed
gradient to a reasonable degree. In the UT there is a consistent high bias
for NO and an associated low bias for NO2, suggesting that the
conversion term is too low and that the NOx cycle is out of synch at
these cold temperatures despite the addition of new reservoir species (i.e.
CH3O2NO2) and application of new rate data.
One important gauge as to whether the chemical mechanism can capture the
correct recycling efficiency of NO into NO2 is to examine their ratio,
which is presented in the third column of Fig. 7. In the LT (< 900 hPa)
NO / NO2 ratios of 0.1–0.2 exist that are captured quite well by
TM5-MP, with negligible differences between 3∘× 2∘
and 1∘× 1∘ simulations. For the FT, TM5-MP
consistently overestimates the ratio in spite of a high bias in O3 (cf.
Fig. 4) which is imposed by the overestimates in NO2. This implies the
chemical conversion is too slow and, assuming representative JNO2
values, indicates a low bias in HO2 or an underestimation in the mixing
ratios of other long-lived and short-lived NOy compounds (see Sect. 7).
Monthly mean comparisons of NO (left), NO2 (middle) and the
resulting NO / NO2 ratio (right) from the TexAQSII campaign during
September and October 2006 and TM5-MP simulations. The dotted line represents
the 1σ deviation in the mean of the measurements. For details of the
locations for each month the reader is referred to Parrish et al. (2009).
Finally in Fig. 8 we show the corresponding comparisons against measurements
taken during the TexAQS II campaign (Parrish et al., 2009) for both September
and October 2006. As for the EMEP comparisons shown in Figs. 5 and 6, there
is a significant underestimation in NO and NO2 mixing ratios, with both
model profiles being outside the 1σ variability in the observational
mean. This is clearly related to the emission estimates for this region being
underestimated in the current emission inventories (e.g. Kim et al., 2011).
For the resulting NO / NO2 ratio, TM5-MP captures the correct ratio
in the lowest few hundred metres of the BL, but overestimates the ratio at
higher altitudes as for more pristine environments, although there is marked
improvement in the ratios simulated for October.
Changes in the NOy budgetLong-lived reservoirs
The resolution-dependent changes in the temporal distribution of [NO2]
and associated differences in NMVOC chemical precursor emissions have the
potential to alter the partitioning of reactive NOx between the three
main chemical reservoirs included in mCB05v2 (i.e. HNO3, PAN and
ORGNTR). The differences in both the deposition efficiency and tropospheric
lifetimes between trace species at 1∘× 1∘ suggests
that the fraction of NOx that can be transported out of source regions
could change significantly. Here we briefly examine the zonally integrated
nitrogen budget terms between simulations to quantify the effect of applying
a higher spatial resolution. The seasonal distribution of these three
dominant reservoir species at 1∘× 1∘ and their
individual contributions to total NOy are shown in Figs. S12–S15 for DJF
and JJA, respectively. Here we define NOy as the cumulative total of NO,
NO2, NO3, HNO3, PAN, CH3O2NO2, HONO,
2⋅N2O5, lumped organic nitrates (ORGNTR) and HNO4. It
should be noted that methyl-nitrate is not in this version of TM5-MP. These
figures are provided as a reference for the reader to aid understanding of
the discussion below.
Table S1 in the Supplement provides a zonal decomposition of the tropospheric
chemical budget terms for HNO3, PAN and ORGNTR. For HNO3, even
though the recent kinetic rate parameters increase (decrease) the chemical
production term at the surface (UTLS) compared to older rate data (e.g.
Seltzer et al., 2015), changes in the integrated column term are small. The
changes at 1∘× 1∘ are somewhat latitude-dependent
(corresponding to low- and high-NOx regimes), with only small increases
occurring in the NH and associated decreases in the tropics related to lower
[OH] (i.e. chemical production). Loss by cumulative deposition terms only
changes by a few percent, due to wet scavenging being so efficient for
HNO3 for any associated change in the SAD of cloud droplets
(cm2/cm3) introduced by changes in the liquid water product.
For PAN, both the production and destruction terms decrease marginally by
∼ 1–3 % across all zones, meaning the transport of NOx out of
the main source regions remains relatively robust. The total mass of N cycled
through PAN is ∼ 4 times that sequestrated as HNO3, although the
lifetime of PAN is shorter due to the efficient thermal decomposition. The
changes in the production term due to temporal increases in NO2 near
high-NOx source regions (cf. Fig. 3) are partially offset by a reduction
in the mixing ratios of the acetyl-peroxy radical (C2O3 in Table 1)
due to e.g. increased dry deposition of organic precursors at
1∘× 1∘. Although the chemical budget terms only
exhibit small changes, it can be expected that the global distribution of PAN
will be somewhat different due to the changes in the convective and advective
mixing due to the application of higher-resolution meteorological data (cf.
Sect. 3).
For ORGNTR, there is a 5 % reduction in the annual production term at
1∘× 1∘, with an associated decrease in the loss by
deposition. Both the largest production and, thus, destruction terms occur in
the tropics related to the strongest source of ORGNTR being biogenic
precursors in mCB05v2. Thus, at 1∘× 1∘, this
intermediate trace species becomes less important as a NOx reservoir.
Finally, the one additional intermediate not shown is
CH3O2NO2, which is primarily a stable vehicle for transporting
NOx from the surface up to the UTLS, where at cold temperatures it
accounts for a significant fraction of NO2 speciation along with
HNO4 (Browne et al., 2011). At the global scale 3 times as much nitrogen
cycles through CH3O2NO2 compared to PAN, although the thermal
stability is low at temperatures > 255 K; thus, resident mixing ratios
are typically small. This results in maximum mixing ratios occurring in the
cold upper troposphere (up to ∼ 0.2 ppb) and subsequently dissociating
primarily by thermal decomposition (photolytic destruction accounting for
> 0.1 % of all destruction). At 1∘× 1∘ there
is a few percent decrease in the chemical production term as a result of
lower CH3O2 mixing ratios and more variability in the temporal
temperature distribution.
Comparison of weekly [HNO3] (µg m-3) from both
3∘× 2∘ and 1∘× 1∘
simulations at 4 selected EMEP sites for 2006. The 1σ deviation in
the weekly observations are shown as error bars. The selected sites shown are
in Norway (top left), Germany (top right), Austria (bottom left) and Slovakia
(bottom right).
Monthly mean comparisons of HNO3 (left) and PAN (right) from
the INTEX-B measurements and TM5-MP simulations. The dotted line represents
the 1σ deviation in the mean of the measurements. For details of the
locations for each month the reader is referred to Singh et al. (2009).
Comparisons of weekly [HNO3] at the surface in Europe are shown in
Fig. 9 against measurements from the EMEP network. It has recently been
determined that HNO3 measurements are also sensitive to ambient
night-time [N2O5], which could result in a positive bias in the
observations (Phillips et al., 2013). In general, the modelled seasonal cycle
is not evident in the measurements, which exhibit a rather homogeneous
variation in mixing ratios throughout the year typically, thus being somewhat
decoupled from variability in photochemical activity. Comparisons show an
underestimation in TM5-MP during March and an overestimation during JJA. No
such seasonal pattern is observed for [NO2] (cf. Fig. 6); thus, seasonal
[OH] variability due to variations in photo-chemical activity and
[H2O(g)] and/or an incorrect wash-out term could both act as
likely causes. The impact of resolution on [HNO3] is rather muted for
most weeks, resulting in no significant changes to the seasonal biases (not
given), as constrained by the improvements in surface [NO2] (cf.
Fig. 6). The heterogeneous scavenging of HNO3 into ammonium nitrate can
act as a moderator of gaseous HNO3. Although this heterogeneous
conversion process is included in TM5-MP as described by the EQSAM approach,
low concentrations of e.g. ammonium nitrate (not shown) typically result.
Thus, gaseous [HNO3] remains too high due to too little conversion into
particles and subsequent deposition.
For other regions outside Europe, we make comparisons of vertical profiles of
HNO3 and PAN between March and September 2006 with those measured during
INTEX-B (Fig. 10) and TexAQSII (Fig. S16). PAN is a good marker for transport
in the free troposphere due to the relatively long lifetime at colder
temperatures. For all regions the vertical gradients for both species are
captured quite well, although some fine structure is lost due to the vertical
resolution of TM5-MP and insufficient precursor emissions. This implies that
the underestimation in NO2 simulated in the UTLS (Fig. 7) is not due to
insufficient transport of NO2 away from source regions, but rather
should be attributed to either missing chemistry or insufficient transport
down from the stratosphere. Finally, for more polluted regions, the vertical
gradient of HNO3 is rather less steep than that observed, with
significant low biases in the lower troposphere related to the low bias in
NO2 shown in Fig. 8. The impact of the 1∘× 1∘
resolution only results in a marginal improvement in the LT for HNO3
(again similar to NO2). For PAN the vertical profile in TM5-MP agrees
remarkably well, but is somewhat anti-correlated around 900 hPa in both
simulations, and the rapid decrease in the lowest kilometre is captured
sufficiently. This overestimation would likely be larger if the emission
estimates were increased as required to consolidate the NO2 comparisons
for the same campaign.
Short-lived reservoirs
Here we briefly discuss the perturbations introduced for the short-lived
N-reservoirs, namely HONO, HNO4 and N2O5, where the chemical
budget terms for all three species are provided in Table S2 in the
Supplement. For HONO it should be noted that many tropospheric CTMs have
difficulty in simulating observed mixing ratios (e.g. Gonçalves et al.,
2012), suggesting missing (heterogeneous) source terms. The global production
for HONO is an order of magnitude less than that for the other short-lived
N-reservoirs. At 1∘× 1∘ there is ∼ 10 %
more chemical production of HONO in high-NOx regions and no appreciable
effect in low-NOx regions. Thus the impact of increased resolution on
HONO production is rather small, which is surprising considering the higher
NO mixing ratios that occur in high-NOx regions (cf. Fig. 5). The muted
response is due to competing oxidative processes, which effectively lower the
OH available. For HNO4, approximately the same mass of N cycles through
this species as for PAN, although the shorter lifetime means that it is more
important at regional scale. Again, the impact of resolution on this species
is small, where decreases in [HO2] result in no significant net change
in production for the NH. The most significant changes occur for the global
production and heterogeneous conversion of N2O5, with enhanced
chemical production of ∼ 12 % at the global scale, increasing the
heterogeneous sink term by ∼ 6 %, although the changes in the total
mass of N converted are small. In general, this is due to an increase in the
production of the NO3 radical by ∼ 10 % at
1∘× 1∘ (not shown) resulting in enhanced
N2O5 mixing ratios.
Implications for tropospheric CH2O retrieval
The implications of applying a higher-resolution CTM for the global
distribution of CH2O are rather modest. Figure S17 shows the
near-surface global distributions of CH2O for May 2006, where maximum
mixing ratios occur near forested regions due to the link with isoprene
oxidation (e.g. Palmer et al., 2006). The tropospheric lifetime of CH2O
is of the order of a few days, meaning that transport has little impact
between simulations, apart from in low-emission areas. Also shown are the
locations where regional validation occurs. In Table 8, we show zonally
integrated chemical production and destruction terms for CH2O, which
suggests changes of the order of a few percent at the global scale. The most
notable difference is the increase in the cumulative deposition term of
∼ 4 % at 1∘× 1∘, thus reducing the
atmospheric lifetime of CH2O in TM5-MP. Again, this low impact shows
that the increase in the temporal variability of the meteorological data at
1∘× 1∘, and thus the local variability of cloud
SAD, only changes the net deposition term by a few percent. Even though the
temporal distribution of the surface mixing ratios shows more variability at
1∘× 1∘ due to the better representation of regional
precursor source terms (e.g.) isoprene and terpene, only moderate
improvements occur to the simulated profiles and total columns due to changes
in transport. For instance, when analysing individual production terms (not
given) for the tropics, decreases are related to small changes in the
dominating chemical source terms (e.g. oxidation of CH3OOH; a reduction
of ∼ 3–5 Tg CH2O yr-1). For the chemical destruction
term, the relative insensitivity of the photolysis of CH2O to resolution
(similar to JO3; cf. Fig. S4) results in small net decreases in
line with changes in the chemical production term.
The tropospheric chemical budget for the CH2O given in
Tg CH2O yr-1 during 2006 for the 1∘× 1∘
simulation. Percentage differences are shown against the corresponding
3∘× 2∘ simulation.
Comparisons of the vertical distribution of CH2O from both
3∘× 2∘ and 1∘× 1∘
simulations against measurements made as part of INTEX-B during 2006. The
dotted line represents the 1σ deviation in the mean of the
measurements. For details on the exact location of the flights the reader is
referred to Parrish et al. (2009).
Figure 11 compares monthly mean tropospheric profiles of CH2O measured
during INTEX-B (Singh et al., 2009) with those from both TM5-MP simulations
for March to May 2006. In general, there is a fair representation of the
vertical gradient of CH2O by TM5-MP for all months shown, although
surface mixing ratios are typically too high, suggesting that loss by
deposition to the ocean is underestimated (potentially related to
underestimations in surface area due to lack of 3-D wave structure) or that
the chemical production term is too efficient. Moreover, there appears to be
a missing (chemical) source term in the UTLS in TM5-MP, leading to a
∼ 30–50 % (∼ 0.05 ppb) low bias above 600 hPa, i.e. there is
no significant improvement to the underestimation in the SH CH2O column
in TM5-MP when compared to mCB05 (Zeng et al., 2015). Comparing profiles
shows that the changes in the vertical distribution of CH2O at
1∘× 1∘ are minimal in the chemical background
compared to 3∘× 2∘, with the main differences
originating from more efficient transport out of source regions (cf. March).
These findings are further confirmed by the comparisons of TM5-MP against
TexAQS II measurements for September and October 2006 (Fig. S18).
Implications for tropospheric SO2
retrieval
The global distributions of near-surface SO2 mixing ratios for both the
3∘× 2∘ and 1∘× 1∘
simulations are shown in the bottom panels of Fig. S17, where the
distribution shows the land-based point sources as applied from the MACCity
emission inventory. The high mixing ratios of SO2 correlate with the
location of strong anthropogenic emission sources due to the relatively short
atmospheric lifetime of SO2 (varying between ∼ 2 days during
winter and ∼ 19 h during summer; Lee et al., 2011), being rapidly
oxidized to sulfate (SO4=). Although the regional distributions are
similar, the 1∘× 1∘ simulation is able to
differentiate point sources to a much better degree, which enhances the
ability to derive more accurate emission fluxes. The in situ chemical
production term of SO2 from the oxidation of oceanic di-methyl sulfide
(DMS) is low; thus, there are very low SO2 mixing ratios in the chemical
background. Also shown are the regions used for validating the SO2
surface concentrations and vertical profiles (insets in Fig. S17; see below).
In Fig. 12, we compare weekly [SO2] for 2006 at a number of EMEP sites
in Austria (AT02, forested), the Netherlands (NL09, rural), Great Britain
(GB43, rural) and Spain (ES10, rural), with most sites being positioned away
from strong point sources. For SO2 in Europe, the main emission source
is anthropogenic (e.g. from the energy sector). High [SO2] has been
observed throughout the EMEP network in e.g. the Netherlands and Spain, which
is significantly higher than that measured in central
Europe (Tørseth et al., 2012). Although the measurement uncertainty is
somewhat site-specific due to the different methodologies employed, it is
typically around ∼ 1.3 µg m-3 (e.g. Hamad et al.,
2010). Comparing weekly averages shows that for most sites shown there is a
significant low bias at 3∘× 2∘, indicating
inaccuracies in the MACC emission inventory and the effect of coarsening on
the model resolution. At 1∘× 1∘ significant
improvements occur in the correlations as a result of the better temporal
distribution of anthropogenic emission sources.
The seasonal mean biases of daily [SO2] (µg m-3)
at 13:00 for DJF and JJA, when taking the difference between
measurements–model values. Values are shown for both the
3∘× 2∘ and 1∘× 1∘
simulations. Those with differences < 5 % are considered to exhibit no
discernible change in the bias.
Table 9 provides an overview of the seasonal biases for all of the EMEP sites
that measure hourly [SO2], with the biases calculated for the overpass
time of tropOMI aggregated on a weekly basis. Improvements occur at
1∘× 1∘ for ∼ 20 % of the sites during both
seasons, with the majority (∼ 50 %) of sites showing no significant
improvement (< 5 %). In such instances the local [SO2] is
determined more by long-range transport (and is thus sensitive to wash-out)
than a local emission source, where strong mitigation practises have been
implemented in Europe over the last few decades reducing resident [SO2]
significantly (Tørseth et al., 2012). For some sites there is a notable
increase in biases at 1∘× 1∘ (20 % DJF, 25 %
JJA), indicating that overly strong local emission sources occur in the MACC
inventories (e.g. ES13 and GR01). For others (e.g. ES08 and NL07)
significantly low biases occur, suggesting the opposite problem.
Comparison of weekly [SO2] (µg m-3) at 13:00
from both the 3∘× 2∘ and
1∘× 1∘ simulations at four selected EMEP sites for
2006. The selected sites shown are in Austria (top left), the Netherlands
(top right), Great Britain (bottom left) and Spain (bottom right).
Comparisons of the monthly tropospheric SO2 profile assembled
from data taken during September and October 2006 as part of TexAQS II. The
1σ deviation of the mean derived from the measurements is shown as
the dotted line. For details of the flight paths the reader is referred to
Parrish et al. (2009).
Finally, for the vertical profiles, we make comparisons against monthly mean
composites assembled from measurements taken during INTEX-B (Fig. S19) and
TexAQS II (Fig. 13) as for the other trace gas species. For the more pristine
locations there are typically low biases at 3∘× 2∘
for all months, especially at the surface during March, indicating a
significant underestimation of the emission fluxes of SO2. Increasing to
1∘× 1∘ only provides an improved correlation for
March, due to the transport in the FT being described better, as shown for
NO2 in Fig. 7. For April, the comparison shows a significant
underestimation in the column for both simulations, where corresponding
comparisons of the vertical profiles of DMS, which acts as a key source of
SO2 in the equatorial Pacific (Alonza Gray et al., 2011), agree quite
well (not shown). This points to an appreciable biogenic source term that is
currently missing from the inventories as proposed for organic nitrates
(Williams et al., 2014). For May, again no significant improvement occurs at
1∘× 1∘, although both simulations capture the peak
in SO2 mixing ratios at the top of the BL. More relevant for
satellite-based retrievals is the observed column near strong anthropogenic
source regions as shown in Fig. 13 over Texas during September and
October 2006. Here a clear improvement occurs at
1∘× 1∘, with the low bias in the BL being reduced
significantly, although the integrated column is still too low. Again this is
due to the underestimation in the source emission fluxes in the anthropogenic
emission inventory employed.
Conclusions
In this paper we have provided a comprehensive description of the
high-resolution 1∘× 1∘ version of TM5, which is to
be used for the purpose of providing a priori columns for the satellite
retrieval of trace gas columns of NO2, CH2O and SO2. By
performing identical simulations at a horizontal resolution of
3∘× 2∘ and 1∘× 1∘, and
comparing the resulting global distributions of trace gas species, photolysis
frequencies and chemical budget terms, we quantify and validate both the
near-surface and vertical distributions for the evaluation year of 2006.
Comparing the seasonal distribution in 222Rn between resolutions, we
show differences in the vertical distribution of up to ±20 % at the
global scale, with significantly larger impacts for specific coastal regions
and tropical oceans. In order to assess the changes in convective activity
above strong NOx sources, we show that differences of between ∼ 2
and 10 % (∼ 10–20 %) exist for the northern
mid-latitudes (tropics) at higher resolution, with both weaker and stronger
upwelling occurring depending on the region and the season. The magnitudes of
the changes are site-specific, being affected by local orography. We have
also made comparisons using a 1∘× 1∘ simulation
applying the Tiedtke (1989) convection scheme, showing that ERA-Interim mass
fluxes result in less transport of 222Rn out of the boundary layer,
where it has been shown that the use of ERA-Interim mass fluxes introduces
inconclusive improvements in surface 222Rn distributions (Koffi et al.,
2016). For lightning NOx
there is a vertical re-distribution in the emission flux, with enhanced NOx injected into the upper troposphere when using the ERA-Interim mass fluxes.
Although the impact of resolution on daily photolysis rates may be
appreciable, analysing global monthly mean JO3 and JNO2
surface values over a range of conditions shows that effects are limited to
∼ 2 % and ∼ 5–10 %, respectively. One contributing factor to
this rather muted impact is that the changes in surface albedo that occur at
1∘× 1∘ are largest at the poles during winter,
which has no impact on photo-chemistry due to the absence of photolysing
light (not shown). For cloud cover, a dominant term for determining total
optical depth, there are significant increases at
1∘× 1∘ over the oceans, although this is generally
related to instances of low photochemical activity. Examining the resulting
changes in JO3 and JNO2 which occur throughout the
tropospheric column reveals that significant differences of > 10 % can
occur at the top of the BL at tropical locations. Such modest changes
associated with this dominant loss term result in the change in the
integrated chemical budget terms for tropospheric O3 and NO2 being
rather low.
Analysing the chemical budget terms for tropospheric O3 shows (i) a
reduction in the stratosphere–troposphere exchange flux of ∼ 7 % to
597 Tg O3 yr-1, (ii) a repartitioning of the contribution from
stratospheric downwelling in both the Northern and Southern hemispheres,
(iii) no significant change in the tropospheric burden of O3 and
(iv) modest changes in the integrated chemical production and destruction
terms. Comparing simulated mixing ratios against surface measurements in
Europe shows that the positive bias present in TM5 decreases by
∼ 20 % at 1∘× 1∘ between 2 and
5 ppb month-1. This positive bias persists throughout the vertical
column across diverse global regions regardless of the local NOx
emissions, although the vertical gradient in tropospheric O3 through the
tropospheric column is captured quite well.
For NO and NO2 increasing horizontal resolution results in only modest
differences in the zonal mean recycling terms and the loss of O3 by
chemical titration. Comparisons against surface measurements in Europe show
that there is a consistent negative bias in weekly [NO] of a few
µg m-3 associated with both overly high surface O3
(enhanced NO titration) and the inaccuracy of the NOx emission
inventories. For NO2, the biases in the weekly concentrations are larger
and can be both positive and negative. Increasing horizontal resolution has
little effect on reducing the NO biases, but results in improvements for
NO2 at ∼ 35 % of the available sites, with ∼ 45 % of
sites showing limited changes. Examining correlation coefficients shows that
although there is typically a higher correlation at
1∘× 1∘, many sites still exhibit very low
correlation or anti-correlation for some seasons. For the tropospheric column
the improvement in the comparisons is only by a few percent, with a
significant underestimation in both NO and NO2 throughout the
tropospheric column. Analysing the NO / NO2 ratio and comparing
against observations show that although partitioning is captured in the BL,
there is a significant overestimation in the upper troposphere.
Finally, for CH2O and SO2, which can also be retrieved from
satellite measurements, the effect of increased resolution is rather modest
due to compensating changes to the chemical budget terms. When compared
against observations there is a persistent low bias for tropospheric
CH2O due to missing production terms, especially in the free
troposphere. SO2 comparison with surface observations in Europe shows
lower biases at 20 % of sites due to more accurate local emission fluxes,
whereas for the majority of cases (∼ 50 %) there is no significant
change. Comparing vertical profiles shows a significant underestimation in
the tropospheric column likely associated with either missing precursors or
an underestimation in the direct emission terms.
Future updates to TM5-MP will most likely focus on developing an online
secondary organic aerosol scheme, tropospheric halogen chemistry and
incorporating an updated isoprene oxidation scheme involving more
intermediate species. It will also be applied in the context of an Earth
system model (EC-EARTH) for allowing future studies concerning
chemistry–climate feedbacks. When computing resources allow, more expensive
simulations can be performed using the 60 vertical levels as defined in the
ERA-Interim meteorological dataset, approximately doubling the resolution of
the simulations presented here. An additional update to improve the STE would
be to apply the second-order moments scheme (Prather, 1986), whose
application has been shown to capture the seasonality and magnitude of STE
exchange to a better degree (Bönisch et al., 2008). In terms of oxidative
capacity, one means of reducing the tropospheric near-surface O3 mixing
ratios would be to improve loss to land surfaces (Hardacre et al., 2010),
although mixing ratios have been shown to be insensitive to the additional
loss term to oceans, which is currently missing from many CTMs (Ganzeveld et
al., 2009). Our comparisons of CH2O and SO2 show that there is a
significant uncertainty of chemical processes that affect distributions in
the pristine marine environment. For instance, the physical process of
deposition seems to be under-represented, possibly due to an overly low
surface area of the surface, i.e. the lack of a flat surface. The significant
underestimates in SO2 suggest missing biogenic source terms; therefore,
more understanding of biogenic emission terms is necessary.
Code availability
The TM5-MP code can be downloaded from the SVN server hosted at KNMI, the
Netherlands. A request to generate a new user account for access can be made
by e-mailing sager@knmi.nl. Any new user groups need to agree to the
protocol set out for use, where it is expected that any developments are
accessible to all users after publication of results. Attendance at 9-monthly
TM5 international meetings is encouraged to avoid duplicity and conflict of
interests.
The Supplement related to this article is available online at doi:10.5194/gmd-10-721-2017-supplement.
The authors declare that they have no conflict of interest.
Acknowledgements
This research has been supported by FP7 project Quality Assurance for
Essential Climate Variables (QA4ECV), no. 607405. We thank M. van Weele for
processing the MSR2 stratospheric ozone data record used for constraining the
overhead O3 field and T. P. C. van Noije for updating the SOx
emission estimates. We thank V. Huijnen for providing estimates on the
heterogeneous uptake coefficients.Edited by:
F. O'Connor Reviewed by: two anonymous referees
ReferencesAan de Brugh, J. M. J., Schaap, M., Vignati, E., Dentener, F., Kahnert, M.,
Sofiev, M., Huijnen, V., and Krol, M. C.: The European aerosol budget in
2006, Atmos. Chem. Phys., 11, 1117–1139, 10.5194/acp-11-1117-2011, 2011.Aas, W., Hjellbrekke, A.-G., Schaug, J., and Solberg, S.: Data quality 1999,
quality assurance and field comparisons, Kjeller, Norwegian Institute for
Air Research, EMEP/CCC Report 6/2001, 2001.Abbatt, J. P. D., Lee, A. K. Y., and Thornton, J. A.: Quantifying trace gas
uptake to tropospheric aerosol: recent advances and remaining challenges,
Chem. Soc. Rev., 41, 6555–6581, 10.1039/c2cs35052a, 2012.Alonza Gray, B., Wang, Y., Gu, D., Bandy, A., Mauldin, L., Clarke, A.,
Alexander, B., and Davis, D. D.: Sources, transport, and sinks of SO2
over the equatorial Pacific during the Pacific Atmospheric Sulfur Experiment,
J. Atmos. Chem., 68, 27–53, 10.1007/s10874-010-9177-7, 2011.Atkinson, R., Baulch, D. L., Cox, R. A., Crowley, J. N., Hampson, R. F.,
Hynes, R. G., Jenkin, M. E., Rossi, M. J., and Troe, J.: Evaluated kinetic
and photochemical data for atmospheric chemistry: Volume I – gas phase
reactions of Ox, HOx, NOx and SOx species, Atmos. Chem. Phys., 4,
1461–1738, 10.5194/acp-4-1461-2004, 2004.Atkinson, R., Baulch, D. L., Cox, R. A., Crowley, J. N., Hampson, R. F., Hynes, R. G., Jenkin, M. E., Rossi, M. J., Troe, J.,
and IUPAC Subcommittee: Evaluated kinetic and photochemical data for atmospheric chemistry: Volume II –
gas phase reactions of organic species, Atmos. Chem. Phys., 6, 3625–4055, 10.5194/acp-6-3625-2006, 2006.Bândă, N., Krol, M., van Noije, T., van Weele, M., Williams, J. E.,
Le Sager, P., Niemeier, U., Thomason, L., and Röckmann, T.: The effect of
stratospheric sulfur from Mount Pinatubo on tropospheric oxidizing capacity
and methane, J. Geophys. Res.-Atmos., 120, 1202–1220,
10.1002/2014JDO22137, 2015.Boersma, K. F., Jacob, D. J., Eskes, H. J., Pinder, R. W., Wang, J., and van
der A, R. J.: Intercomparison of SCIAMACHY and OMI tropospheric NO2
columns: observing the diurnal evolution of chemistry and emissions from
space, J. Geophys. Res., 2, 1–14, 10.1029/2007JD008816, 2008.Boersma, K. F., Eskes, H. J., Dirksen, R. J., van der A, R. J., Veefkind, J.
P., Stammes, P., Huijnen, V., Kleipool, Q. L., Sneep, M., Claas, J.,
Leitão, J., Richter, A., Zhou, Y., and Brunner, D.: An improved
tropospheric NO2 column retrieval algorithm for the Ozone Monitoring
Instrument, Atmos. Meas. Tech., 4, 1905–1928, 10.5194/amt-4-1905-2011,
2011.Bönisch, H., Hoor, P., Gurk, Ch., Feng, W., Chipperfield, M., Engel, A.,
and Bregman, B.: Model evaluation of CO2 and SF6 in the
extratropical UT/LS region, J. Geophys. Res., 113, D06101,
10.1029/2007JD008829, 2008.Bregman, B., Segers, A., Krol, M., Meijer, E., and van Velthoven, P.: On the
use of mass-conserving wind fields in chemistry-transport models, Atmos.
Chem. Phys., 3, 447–457, 10.5194/acp-3-447-2003, 2003.Browne, E. C., Perring, A. E., Wooldridge, P. J., Apel, E., Hall, S. R.,
Huey, L. G., Mao, J., Spencer, K. M., Clair, J. M. St., Weinheimer, A. J.,
Wisthaler, A., and Cohen, R. C.: Global and regional effects of the
photochemistry of CH3O2NO2: evidence from ARCTAS, Atmos. Chem.
Phys., 11, 4209–4219, 10.5194/acp-11-4209-2011, 2011.Carslaw, D. C. and Beevers, S. D.: Estimations of road vehicle primary
NO2 exhaust emission frcations using monitoring data in London, Atmos.
Environ., 39, 167–177, 2005.Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P.,
Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P.,
Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N.,
Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S.
B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P., Köhler,
M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J.-J.,
Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N., and
Vitart, F.: The ERA-Interim reanalysis: configuration and performance of the
data assimilation system, Q. J. Roy. Meteor. Soc., 137, 553–597, 2011.De Smedt, I., Müller, J.-F., Stavrakou, T., van der A, R., Eskes, H., and
Van Roozendael, M.: Twelve years of global observations of formaldehyde in
the troposphere using GOME and SCIAMACHY sensors, Atmos. Chem. Phys., 8,
4947–4963, 10.5194/acp-8-4947-2008, 2008.Dunlea, E. J., Herndon, S. C., Nelson, D. D., Volkamer, R. M., San Martini,
F., Sheehy, P. M., Zahniser, M. S., Shorter, J. H., Wormhoudt, J. C., Lamb,
B. K., Allwine, E. J., Gaffney, J. S., Marley, N. A., Grutter, M., Marquez,
C., Blanco, S., Cardenas, B., Retama, A., Ramos Villegas, C. R., Kolb, C. E.,
Molina, L. T., and Molina, M. J.: Evaluation of nitrogen dioxide
chemiluminescence monitors in a polluted urban environment, Atmos. Chem.
Phys., 7, 2691–2704, 10.5194/acp-7-2691-2007, 2007.Dupuy, E., Urban, J., Ricaud, P., Le Flochmoën, E., Lautié, N.,
Murtagh, D., De La Noë, J., El Amraoui, L., Eriksson, P., Forkman, P.,
Frisk, U., Jégou, F., Jiménez, C., and Olberg, M.: Strato-mesospheric
measurements of carbon monoxide with the Odin Sub-Millimetre Radiometer:
Retrieval and first results, Geophys. Res. Letts., 31, L20101,
10.1029/2004GL020558, 2004.Emmons, L. K., Walters, S., Hess, P. G., Lamarque, J.-F., Pfister, G. G.,
Fillmore, D., Granier, C., Guenther, A., Kinnison, D., Laepple, T., Orlando,
J., Tie, X., Tyndall, G., Wiedinmyer, C., Baughcum, S. L., and Kloster, S.:
Description and evaluation of the Model for Ozone and Related chemical
Tracers, version 4 (MOZART-4), Geosci. Model Dev., 3, 43–67,
10.5194/gmd-3-43-2010, 2010.Evans, M. J. and Jacob, D. J.: Impact of new laboratory studies of
N2O5 hydrolysis on global model budgets of tropospheric nitrogen
oxides, ozone and OH, Geophys. Res. Lett., 32, L09813,
10.1029/2005GL022469, 2005.Fioletov, V. E., McLinden, C. A., Krotkov, N., and Li, C.: Lifetimes and emissions of SO2 from point
sources estimated from OMI, Geophys. Res. Letts., 42, 1969–1976,
10.1002/2015GL063148, 2015.Fisher, J. A., Wilson, S. R., Zeng, G., Williams, J. E., Emmons, L. K.,
Langenfelds, R. L., Krummel, P. B., and Steele, L. P.: Seasonal changes in
the tropospheric carbon monoxide profile over the remote Southern Hemisphere
evaluated using multi-model simulations and aircraft observations, Atmos.
Chem. Phys., 15, 3217–3239, 10.5194/acp-15-3217-2015, 2015.Fortuin, J. P. F. and Kelder, H.: An ozone climatology based on ozonesonde
and satellite measurements, J. Geophys. Res., 103, 31709–31734, 1998.Ganzeveld, L., Helmig, D., Fairall, C. W., Hare, J., and Pozzer, A.:
Atmosphere-ocean ozone exchange: A global modeling study of biogeochemical,
atmospheric, and waterside turbulence dependencies, Global Biogeochem. Cy.,
23, GB4021, 10.1029/2008GB003301, 2009.Gonçalves, M., Dabdub, D., Chang, W. L., Jorba, O., and Baldasanoa, J.
M.: Impact of HONO sources on the performance of mesoscale air quality
models, Atmos. Environ., 54, 168–176, 10.1016/j.atmosenv.2012.02.079,
2012.Gonzi, S., Palmer, P. I., Barkley, M., De Smedt, I., and Roozendael, M. V.:
Biomass burning emission estimates inferred from satellite column
measurements of HCHO: Sensitivity to co-emitted aerosol and injection height,
Geophys. Res. Letts., 38, L14807, 10.1029/2011GL047890V, 2011.Granier, C., Bessagnet, B., Bond, T., D'Angiola, A., Denier van der Gon, H.,
Frost, G. J., Heil, A., Kaiser, J. W., Kinne, S., Klimont, Z., Kloster,
S.-F., Lamarque, J., Liousse, C., Masui, T., Meleux, F., Mieville, A., Ohara,
T., Raut, J.-C., Riahi, K., Schultz, M. G., Smith, S. J., Thompson, A., van
Aardenne, J., van der Werf, G. R., and van Vuuren, D. P.: Evolution of
anthropogenic and biomass burning emissions of air pollutants at global and
regional scales during the 1980–2010 period, Climate Change, 109, 163–190,
10.1007/s10584-011-0154-1, 2011.Grooß, J.-U. and Russell III, J. M.: Technical note: A stratospheric
climatology for O3, H2O, CH4, NOx, HCl and HF derived from HALOE
measurements, Atmos. Chem. Phys., 5, 2797–2807, 10.5194/acp-5-2797-2005,
2005.Halmer, M. M., Schmincke, H.-U., and Graf, H.-F.: The annual volcanic gas input
into the atmosphere, in particular into the stratosphere: a global data set
for the past 100 years, J. Volcanol. Geoth. Res., 115, 511–528, 2002.Hamed, A., Birmili, W., Joutsensaari, J., Mikkonen, S., Asmi, A., Wehner, B.,
Spindler, G., Jaatinen, A., Wiedensohler, A., Korhonen, H., Lehtinen, K. E.
J., and Laaksonen, A.: Changes in the production rate of secondary aerosol
particles in Central Europe in view of decreasing SO2 emissions between
1996 and 2006, Atmos. Chem. Phys., 10, 1071–1091,
10.5194/acp-10-1071-2010, 2010.Hardacre, C., Wild, O., and Emberson, L.: An evaluation of ozone dry
deposition in global scale chemistry climate models, Atmos. Chem. Phys., 15,
6419–6436, 10.5194/acp-15-6419-2015, 2015.Hauglustaine, D. A., Balkanski, Y., and Schulz, M.: A global model simulation
of present and future nitrate aerosols and their direct radiative forcing of
climate, Atmos. Chem. Phys., 14, 11031–11063, 10.5194/acp-14-11031-2014,
2014.Heckel, A., Kim, S.-W., Frost, G. J., Richter, A., Trainer, M., and Burrows,
J. P.: Influence of low spatial resolution a priori data on tropospheric
NO2 satellite retrievals, Atmos. Meas. Tech., 4, 1805–1820,
10.5194/amt-4-1805-2011, 2011.Huijnen, V., Williams, J., van Weele, M., van Noije, T., Krol, M., Dentener,
F., Segers, A., Houweling, S., Peters, W., de Laat, J., Boersma, F.,
Bergamaschi, P., van Velthoven, P., Le Sager, P., Eskes, H., Alkemade, F.,
Scheele, R., Nédélec, P., and Pätz, H.-W.: The global chemistry
transport model TM5: description and evaluation of the tropospheric chemistry
version 3.0, Geosci. Model Dev., 3, 445–473, 10.5194/gmd-3-445-2010,
2010.Holtslag, A. A. and Boville, B. A.: Local versus non-local boundary layer
diffusion in a global climate model, J. Climate, 10, 1825–1842, 1993.Huijnen, V., Williams, J. E., and Flemming, J.: Modeling global impacts of
heterogeneous loss of HO2 on cloud droplets, ice particles and aerosols,
Atmos. Chem. Phys. Discuss., 14, 8575–8632, 10.5194/acpd-14-8575-2014,
2014.Jacob, D. J., Prather, M. J., Rasch, P. J., Shia, R. L., Balkanski, Y. J.,
Beagley, S. R., Bergmann, D. J., Blackshear, W. T., Brown, M., Chiba, M.,
Chipperfield, M. P., de Grandpré, J., Dignon, J. E., Feichter, J.,
Genthon, C., W. L. Grose, W. L., Kasibhatla, P. S., Köhler, I., Kritz, M.
A., Law, K., Penner, J. E., Ramonet, M., Reeves, C. E., Rotman, D. A.,
Stockwell, D. Z., Van Velthoven, P. F. J., Verver, G., Wild, O., Yang, H.,
and Zimmermann, P.: Evaluation and intercomparison of global atmospheric
transport models using 222Rn and other short-lived tracers, J. Geophys.
Res., 102, 5953–5970, 10.1029/96JD02955, 1997.Jacob, J. D.: Heterogeneous chemistry and tropospheric ozone, Atmos.
Environ., 34, 2131–2159, 2000.Jégou, F., Urban, J., de La Noë, J., Ricaud, P., Le Flochmoën, E.,
Murtagh, D. P., Eriksson, P., Jones, A., Petelina, S., Llewellyn, E. J.,
Lloyd, N. D., Haley, C., Lumpe, J., Randall, C., Bevilacqua, R. M., Catoire,
V., Huret, N., Berthet, G., Renard, J. B., Strong, K., Davies, J., Mc Elroy,
C. T., Goutail, F., and Pommereau, J. P.: Technical Note: Validation of
Odin/SMR limb observations of ozone, comparisons with OSIRIS, POAM III,
ground-based and balloon-borne instruments, Atmos. Chem. Phys., 8,
3385–3409, 10.5194/acp-8-3385-2008, 2008.Kim, S.-W., McKeen, S. A., Frost, G. J., Lee, S.-H., Trainer, M., Richter,
A., Angevine, W. M., Atlas, E., Bianco, L., Boersma, K. F., Brioude, J.,
Burrows, J. P., de Gouw, J., Fried, A., Gleason, J., Hilboll, A., Mellqvist,
J., Peischl, J., Richter, D., Rivera, C., Ryerson, T., te Lintel Hekkert, S.,
Walega, J., Warneke, C., Weibring, P., and Williams, E.: Evaluations of
NOx and highly reactive VOC emission inventories in Texas and their
implications for ozone plume simulations during the Texas Air Quality Study
2006, Atmos. Chem. Phys., 11, 11361–11386, 10.5194/acp-11-11361-2011,
2011.Koffi, E. N., Bergamaschi, P., Karstens, U., Krol, M., Segers, A., Schmidt,
M., Levin, I., Vermeulen, A. T., Fisher, R. E., Kazan, V., Klein Baltink, H.,
Lowry, D., Manca, G., Meijer, H. A. J., Moncrieff, J., Pal, S., Ramonet, M.,
Scheeren, H. A., and Williams, A. G.: Evaluation of the boundary layer
dynamics of the TM5 model over Europe, Geosci. Model Dev., 9, 3137–3160,
10.5194/gmd-9-3137-2016, 2016.Krol, M., Houweling, S., Bregman, B., van den Broek, M., Segers, A., van
Velthoven, P., Peters, W., Dentener, F., and Bergamaschi, P.: The two-way
nested global chemistry-transport zoom model TM5: algorithm and applications,
Atmos. Chem. Phys., 5, 417–432, 10.5194/acp-5-417-2005, 2005.Lawrence, M. G. and Crutzen, P. J.: The impact of cloud particle
gravitational settling on soluble trace gas distributions, Tellus, 50B,
263–289, 1998.Lee, C., Martin, R. V., van Donkelaar, A., Lee, H., Dickerson, R. R., Hains,
J. C., Krotkov, N., Richter, A., Vinnikov, K. and Schwab, J. J.: SO2 emissions and lifetimes: Estimates from
inverse modeling using in situ and global, space-based (SCIAMACHY and OMI)
observations, J. Geophys. Res., 116, D06304, 10.1029/2010JD014758,
2011.Lelieveld, J., Peters, W., Dentener, F. J., and Krol, M. C.: Stability of
tropospheric hydroxyl chemistry, J. Geophys. Res., 107, 4715,
10.1029/2002JD002272, 2002.Lin, J.-T., Martin, R. V., Boersma, K. F., Sneep, M., Stammes, P., Spurr, R.,
Wang, P., Van Roozendael, M., Clémer, K., and Irie, H.: Retrieving
tropospheric nitrogen dioxide from the Ozone Monitoring Instrument: effects
of aerosols, surface reflectance anisotropy, and vertical profile of nitrogen
dioxide, Atmos. Chem. Phys., 14, 1441–1461, 10.5194/acp-14-1441-2014,
2014.Louis, J. F.: A parametric model of vertical eddy fluxes in the atmosphere,
Bound.-Lay. Meteorol., 17, 187–202, 1979.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.Martin, G. M., Johnson, D. W., and Spice, A.: The measurement and
parameterization of effective radius of droplets in warm stratocumulus
clouds, J. Atmos. Sci., 51, 1823–1842, 1994.Meijer, E. W., van Velthoven, P. F. J., Brunner, D. W., Huntrieser, H., and
Kelder, H.: Improvement and evaluation of the parameterisation of nitrogen
oxide production by lightning, Phys. Chem. Earth, 26, 557–583, 2001.Meloen, J., Siegmund, P., van Velthoven, P., Kelder, H., Sprenger, M.,
Wernli, H., Kentarchos, A., Roelofs, G., Feichter, J., Land, C., Forster, C.,
James, P., Stohl, A., Collins, W., and Cristofanelli, P.:
Stratosphere-troposphere exchange: A model and method intercomparison, J.
Geophys. Res., 108, 8256, 10.1029/2002JD002274, 2002.Metzger, S., Dentener, F., Pandis, S., and Lelieveld, J.: Gas/aerosol partitioning: 1. A
computationally efficient model, J. Geophys. Res., 107, 10.1029/2001JD001102, 2002.Möllner, A. K., Valluvadasan, S., Feng, L. Sprague, M. K., Okumura, M.,
Milligan, D. B., Bloss, W. J., Sander, S. P., Martien, P. T., Harley, R. A.,
McCoy, A. B., and Carter, W. P. L.: Rate of Gas Phase Association of Hydroxyl
Radical and Nitrogen Dioxide, Science, 330, 646–649,
10.1126/science.1193030, 2010.Olsen, S. C., McLinden, C. A., and Prather, M. J.: Stratospheric
N2O–NOy system: Testing uncertainties in a three-dimensional
framework, J. Geophys. Res., 106, 28771–28784, 2001.Olszyna, K. J., Bailey, E. M., Simonaitis, R., and Meagher, J. F.: O3
and NOy relationships at a rural site, 99, 14557–14563, J. Geophys.
Res., 10.1029/94JD00739, 1994.Palmer, P. I., Abbot, D. S., Fu, T.-M., Jacob, D. J., Chance, K., Kurosu, P.,
Guenther, A., Wiedinmyer, C., Stanton, J. C., Pilling, M. J., Pressley, S.
N., Lamb, B., and Sumner, A. L.: Quantifying the seasonal and interannual
variability of North American isoprene emissions using satellite observations
of the formaldehyde column, J. Geophys. Res., 111, D12315,
10.1029/2005JD006689, 2006.Parrish, D. D., Allen, D. T., Bates, T. S., Estes, M., Fehsenfeld, F. C.,
Feingold, G., Ferrare, R., Hardesty, R. M., Meagher, J. F., Nielsen-Gammon,
J. W., Pierce, R. B., Ryerson, T. B., Seinfeld, J. H., and Williams, E. J.:
Overview of the Second Texas Air Quality Study (TexAQS II) and the Gulf of
Mexico Atmospheric Composition and Climate study (GoMACCS), J. Geophys. Res.,
114, D00F13, 10.1029/2009JD011842, 2009.Phillips, G. J., Makkonen, U., Schuster, G., Sobanski, N., Hakola, H., and
Crowley, J. N.: The detection of nocturnal N2O5 as HNO3 by alkali-
and aqueous-denuder techniques, Atmos. Meas. Tech., 6, 231–237,
10.5194/amt-6-231-2013, 2013.Pope, R. J., Chipperfield, M. P., Savage, N. H., Ordóñez, C., Neal, L.
S., Lee, L. A., Dhomse, S. S., Richards, N. A. D., and Keslake, T. D.:
Evaluation of a regional air quality model using satellite column NO2:
treatment of observation errors and model boundary conditions and emissions,
Atmos. Chem. Phys., 15, 5611–5626, 10.5194/acp-15-5611-2015, 2015.Prather, M. J.: Numerical advection by conservation of second-order moments,
J. Geophys. Res., 91, 6671–6681, 1986.Russell, A. R., Perring, A. E., Valin, L. C., Bucsela, E. J., Browne, E. C.,
Wooldridge, P. J., and Cohen, R. C.: A high spatial resolution retrieval of
NO2 column densities from OMI: method and evaluation, Atmos. Chem. Phys.,
11, 8543–8554, 10.5194/acp-11-8543-2011, 2011.Sander, S. P., Abbatt, J. R., Burkholder, J. B., Friedl, R. R., Golden, D. M., Huie, R. E., Kolb, C. E.,
Kurylo, G., Moortgat, K., Orkin, V. L., and Wine, P. H.: Chemical kinetics and Photochemical Data
for Use in Atmospheric studies, Evaluation No.17, JPL Publication 10-6, Jet Propulsion
Laboratory, Pasadena, available at:http://jpldataeval.jpl.nasa.gov (last access: 26 February 2016), 2011.Schery, S. D.: Progress on Global Rn222 Flux Maps and Recommendations
for Future Research, in: 1st International Expert Meeting on Sources and
Measurements of Natural Radionuclides Applied to Climate and Air Quality
Studies (Gif-sur-Yvette, France, June 2003), edited by: Barrie, L. A. and
Lee, H. N., WMO TD 1201, Gif-sur-Yvette, France, 43–47, 2004.Schwartz, S. E.: Mass-transport considerations pertinent to aqueous-phase
reactions of gases in liquid-water clouds, in: Chemistry of Multiphase
Atmospheric Systems, edited by: Jaechske, W., Springer, Heidelberg,
415–471, 1986.Seltzer, K. M., Vizuete, W., and Henderson, B. H.: Evaluation of updated
nitric acid chemistry on ozone precursors and radiative effects, Atmos. Chem.
Phys., 15, 5973–5986, 10.5194/acp-15-5973-2015, 2015.Shettle, E. P. and Fenn, R. W.: Models for the aerosols of the lower
atmosphere and the effects of the humidity variations on their optical
properties, Environ. Res. Paper, 676, AFGL-TR-79-0114, 91 pp., 1979.Sindelarova, K., Granier, C., Bouarar, I., Guenther, A., Tilmes, S.,
Stavrakou, T., Müller, J.-F., Kuhn, U., Stefani, P., and Knorr, W.: Global
data set of biogenic VOC emissions calculated by the MEGAN model over the
last 30 years, Atmos. Chem. Phys., 14, 9317–9341,
10.5194/acp-14-9317-2014, 2014.Singh, H. B., Brune, W. H., Crawford, J. H., Flocke, F., and Jacob, D. J.:
Chemistry and transport of pollution over the Gulf of Mexico and the Pacific:
spring 2006 INTEX-B campaign overview and first results, Atmos. Chem. Phys.,
9, 2301–2318, 10.5194/acp-9-2301-2009, 2009.Stavrakou, T., Müller, J.-F., De Smedt, I., Van Roozendael, M., van der
Werf, G. R., Giglio, L., and Guenther, A.: Global emissions of non-methane
hydrocarbons deduced from SCIAMACHY formaldehyde columns through 2003–2006,
Atmos. Chem. Phys., 9, 3663–3679, 10.5194/acp-9-3663-2009, 2009.Steinbacher, M., Zellweger, C., Schwarzenbach, B., Bugmann, S., Buchmann,
B., Ordónez, C., Prevot, A. S. H., and Hueglin, C.: Nitrogen oxides
measurements at rural sites in Switzerland: bias of conventional measurement
techniques, J. Geophys. Res., 112, D11307, 10.1029/2006JD007971, 2007.Stevenson, D. S., Dentener, F. J., Schultz, M. G., Ellingsen, K., van Noije,
T. P. C., Wild, O., Zeng, G., Amann, M., Atherton, M., Bell, N., Bergmann, D.
J., Bey, I., Bulter, T., Cofala, J., Collins, W. J., Derwent, R. G., Doherty,
R. M., Drevet, J., Eskes, H. J., Fiore, A. M., Gauss, M., Hauglustaine, D.
A., Horowitz, L. W., Isaksen, I. S. A., Krol, M. C., Lamarque, J.-F.,
Lawrence, M. G., Montanaro, V., Müller, J. F., Pitari, G., Prather, M. J.,
Pyle, J. A., Rast, S., Rodriguez, J. M., Sanderson, M. G., Savage, N. H.,
Shindell, D. T., Strahan, S. E., Sudo, K., and Szopa, S.: Multimodel ensemble
simulations of present-day and near future tropospheric ozone, J. Geophys.
Res., 111, D08301, 10.1029/2005JD006338, 2006.Sutton, R. T., Dong, B., and Gregory, J. M.: Land/sea warming ratio in
response to climate change: IPCC AR4 model results and comparison with
observations, Geophys. Res. Letts., 34, L02701, 10.1029/2006GL028164,
2007.Tang, Q., Prather, M. J., and Hsu, J.: Stratosphere-troposphere exchange
ozone flux related to deep convection, Geophys. Res. Letts., 38, L03806,
10.1029/2010GL046039, 2011.Thouret, V., Marenco, A., Logan, J. A., Nédélec, P., and Grouhel,
C.: Comparisons of ozone measurements from the MOZAIC airborne program and
the ozone sounding network at eight locations, J. Geophys. Res., 103,
25695–25720, 1998.Tiedtke, M.: A comprehensive mass flux scheme for cumulus parameterization
in large-scale models, Mon. Weather. Rev., 117, 1779–1800, 1989.Tørseth, K., Aas, W., Breivik, K., Fjæraa, A. M., Fiebig, M.,
Hjellbrekke, A. G., Lund Myhre, C., Solberg, S., and Yttri, K. E.:
Introduction to the European Monitoring and Evaluation Programme (EMEP) and
observed atmospheric composition change during 1972–2009, Atmos. Chem.
Phys., 12, 5447–5481, 10.5194/acp-12-5447-2012, 2012.Urban, J., Pommier, M., Murtagh, D. P., Santee, M. L. and Orsolini, Y. J.:
Nitric acid in the stratosphere based on Odin observations from 2001 to 2009
– Part 1: A global climatology, Atmos. Chem. Phys., 9, 7031–7044,
10.5194/acp-9-7031-2009, 2009.Valks, P., Pinardi, G., Richter, A., Lambert, J.-C., Hao, N., Loyola, D., Van
Roozendael, M., and Emmadi, S.: Operational total and tropospheric NO2
column retrieval for GOME-2, Atmos. Meas. Tech., 4, 1491–1514,
10.5194/amt-4-1491-2011, 2011.van der A, R. J., Allaart, M. A. F., and Eskes, H. J.: Multi sensor
reanalysis of total ozone, Atmos. Chem. Phys., 10, 11277–11294,
10.5194/acp-10-11277-2010, 2010.van der Werf, G. R., Randerson, J. T., Giglio, L., Collatz, G. J., Mu, M.,
Kasibhatla, P. S., Morton, D. C., DeFries, R. S., Jin, Y., and van Leeuwen,
T. T.: Global fire emissions and the contribution of deforestation, savanna,
forest, agricultural, and peat fires (1997–2009), Atmos. Chem. Phys., 10,
11707–11735, 10.5194/acp-10-11707-2010, 2010.van Geffen, J. H. G. M., Boersma, K. F., Eskes, H. J., Maasakkers, J. D., and
Veefkind, J. P.: TROPOMI Algorithm Theoretical Basis Document (ATBD)
tropospheric and total NO2, S5P-KNMI-L2-0005-RP, 56 pp., 2016.Veefkind, J. P., Aben, I., McMullan, K., Förster, H., de Vries, J.,
Otter, G., Claas, J., Eskes, H. J., de Haan, J. F., Kleipool, Q., van Weele,
M., Hasekamp, O., Hoogeveen, R., Landgraf, J., Snel, R., Tol, P., Ingmann,
P., Voors, R., Kruizinga, B., Vink, R., Visser, H., and Levelt, P. F.:
TROPOMI on the ESA Sentinel-5 Precursor: A GMES mission for global
observations of the atmospheric composition for climate, air quality and
ozone layer applications, Remote. Sens. Environ., 120, 70–83,
10.1016/j.rse.2011.09.027, 2012.Verstraeten, W. V., Neu, J. L., Williams, J. E., Bowman, K. W., and Worden,
J. R.: Rapid increases in tropospheric ozone production and export from
China, Nat. Geosci., 8, 690–695, 10.1038/ngeo2493, 2015.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.Vinken, G. C. M., Boersma, K. F., Maasakkers, J. D., Adon, M., and Martin, R.
V.: Worldwide biogenic soil NOx emissions inferred from OMI NO2
observations, Atmos. Chem. Phys., 14, 10363–10381,
10.5194/acp-14-10363-2014, 2014.Vogelezang, D. H. P. and Holtslag, A. A. M.: Evaluation and model impacts of
alternative boundary-layer height formulations, Bound.-Lay. Meteorol., 81,
245–269, 1996.von Kuhlmann, R. and Lawrence, M. G.: The impact of ice uptake of nitric acid
on atmospheric chemistry, Atmos. Chem. Phys., 6, 225–235,
10.5194/acp-6-225-2006, 2006.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.Williams, J. E., Strunk, A., Huijnen, V., and van Weele, M.: The application
of the Modified Band Approach for the calculation of on-line
photodissociation rate constants in TM5: implications for oxidative capacity,
Geosci. Model Dev., 5, 15–35, 10.5194/gmd-5-15-2012, 2012.Williams, J. E., van Velthoven, P. F. J., and Brenninkmeijer, C. A. M.:
Quantifying the uncertainty in simulating global tropospheric composition due
to the variability in global emission estimates of Biogenic Volatile Organic
Compounds, Atmos. Chem. Phys., 13, 2857–2891, 10.5194/acp-13-2857-2013,
2013.Williams, J. E., Le Bras, G., Kukui, A., Ziereis, H., and Brenninkmeijer, C.
A. M.: The impact of the chemical production of methyl nitrate from the NO
+ CH3O2 reaction on the global distributions of alkyl nitrates,
nitrogen oxides and tropospheric ozone: a global modelling study, Atmos.
Chem. Phys., 14, 2363–2382, 10.5194/acp-14-2363-2014, 2014.Worden, J., Liu, X., Bowman, K., Chance, K., Beer, R., Eldering, A., Gunson,
M., and Worden, H.: Improved tropospheric ozone profile retrievals using OMI
and TES radiances, Geophys. Res. Lett., 34, L01809,
10.1029/2006GL027806, 2007.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, 2014.Yarwood, G., Rao, S., Yocke, M., and Whitten, G.: Updates to the carbon bond
chemical mechanism: CB05, Final report to the US EPA, EPA Report Number:
RT-0400675, available at: www.camx.com (last access: 15 January 2014),
2005.Zdunkowski, W. G., Welsch, R. M., and Kord, G. J.: An investigation of the
structure of typical 2-stream methods for the calculation of solar fluxes and
heating rates in clouds, Contributions to Atmospheric Physics, 53, 215–238,
1980.
Zeng, G., Williams, J. E., Fisher, J. A., Emmons, L. K., Jones, N. B., Morgenstern, O., Robinson, J., Smale, D., Paton-Walsh, C., and Griffith, D. W. T.: Multi-model simulation of CO and HCHO
in the Southern Hemisphere: comparison with observations and impact of biogenic emissions, Atmos. Chem. Phys., 15, 7217–7245, 10.5194/acp-15-7217-2015, 2015.Zhang, K., Feichter, J., Kazil, J., Wan, H., Zhuo, W., Griffiths, A. D.,
Sartorius, H., Zahorowski, W., Ramonet, M., Schmidt, M., Yver, C., Neubert,
R. E. M., and Brunke, E.-G.: Radon activity in the lower troposphere and its
impact on ionization rate: a global estimate using different radon emissions,
Atmos. Chem. Phys., 11, 7817–7838, 10.5194/acp-11-7817-2011, 2011.Zhou, Y., Brunner, D., Hueglin, C., Henne, S., and Staehelin, J.: Changes in
OMI tropospheric NO2 columns over Europe from 2004 to 2009 and the
influence of meteorological variability, Atmos. Environ., 46, 482–495, 2012.Zyrichidou, I., Soukouli, M. E., Balis, D., Markakis, K., Poupkou,
A., Katragkou, E., Kioutsioukis, I., Melas, D., Boersma, K. F., and van
Roozendael, M.: Identification of surface NOx emission sources on a
regional scale using OMI NO2, Atmso. Environ., 101, 82–93, 2015.