Aerosol types

Carbonaceous aerosols

Reference: Park et al. [2003]. This paper describes the original formulation of carbonaceous aerosols in GEOS-Chem:

The simulation of carbonaceous aerosols in GEOS-Chem follows that of the Georgia Tech/Goddard Global Ozone Chemistry Aerosol Radiation and Transport (GOCART) model Chin et al. [2002], with a number of modifications described below. The model resolves EC and OC, with a hydrophobic and a hydrophilic fraction for each (i.e., four aerosol types). Combustion sources emit hydrophobic aerosols that then become hydrophilic with an e-folding time of 1.2 days following Cooke et al. [1999] and Chin et al. [2002]. We assume that 80% of EC and 50% of OC emitted from all primary sources are hydrophobic [Cooke et al. [1999]; Chin et al. [2002], Chung and Seinfeld [2002]. All secondary OC is assumed to be hydrophilic. The four aerosol types in the model are further resolved into contributions from fossil fuel, biofuel, and biomass burning, plus an OC component of biogenic origin, resulting in a total of 13 tracers transported by the model.

Simulation of aerosol wet and dry deposition follows the schemes used by Liu et al. [2001] in previous GEOS-Chem simulations of \(^{210}Pb\) and \(^{7}Be\) aerosol tracers. Wet deposition includes contributions from scavenging in convective updrafts, rainout from convective anvils, and rainout and washout from large-scale precipitation. Wet deposition is applied only to the hydrophilic component of the aerosol. Dry deposition of aerosols uses a resistance-in-series model Walcek et al. [1986] dependent on local surface type and meteorological conditions; it is small compared to wet deposition. Liu et al. [2001] found no systematic biases in their simulations of 210Pb and 7Be with GEOS-Chem.

We use global anthropogenic emissions of EC (6.4 Tg year1) and OC (10.5 Tg year1) from the gridded Cooke et al. [1999] inventory for 1984…. Cooke et al. [1999] do not resolve the contributions to EC and OC emissions from heating fuel. We assume these contributions to represent 8% (EC) and 35% (OC) of total anthropogenic emissions, based on data for the Pittsburgh area from Cabada et al. [2002] and apply local seasonal variations of emissions using the heating degree days approach Energy Information Administration [2001]; Cabada et al. [2002]. In this manner we find that anthropogenic EC emission in the United States in winter is 15% higher than in summer. For OC the anthropogenic winter emission is twice that in summer.

Biomass burning emissions of EC and OC are calculated using the global biomass burning inventory of Duncan et al. [2003].

Secondary formation of OC from oxidation of large hydrocarbons is an important source but uncertainties are large Griffin et al. [1999]; Kanakidou et al. [2000]; Chung and Seinfeld [2002]. Chung and Seinfeld [2002] find that biogenic terpenes are the main source of secondary OC aerosols. We assume a 10% carbon yield of OC from terpenes Chin et al. [2002], and apply this yield to a global terpene emission inventory dependent on vegetation type, monthly adjusted leaf area index, and temperature Guenther et al. [1995].

Since Park et al. [2003], there have been several notable updates, such as:

1. EC and OC biomass burning emissions now are taken from GFED instead of Duncan et al. [2003].

2. GEOS-Chem now also has the option of adding several secondary organic aerosol species to the simulation.

3. Anthropogenic emissions of EC and OC now come from the CEDS inventory.

Dust aerosols

The DustL23M mobilization scheme

Reference: Zhang et al. [2025]. From the abstract:

Accurate representation of mineral dust remains a challenge for global air quality or climate models due to inadequate parametrization of the emission scheme, removal mechanisms, and size distribution. While various studies have constrained aspects of dust emission fluxes and/or dust optical depth, annual mean surface dust concentrations still vary by factors of 5–10 among models. In this study, we focus on improving the annual simulation of fine dust in the GEOS-Chem chemical transport model, leveraging recent mechanistic understanding of dust source and removal, and reconciling the size differences between models and ground-based measurements. Specifically, we conduct sensitivity simulations using GEOS-Chem in its high performance configuration (GCHP) version 14.4.1 to investigate the effects of mechanism or parameter updates on annual mean concentrations. The results are evaluated by comparisons versus Deep Blue satellite-based aerosol optical depth (AOD) and AErosol RObotic NETwork (AERONET) ground-based AOD for total column abundance, and versus the Surface Particulate Matter Network (SPARTAN) for novel measurements of surface PM2.5 dust concentrations. Reconciling modelled geometric diameter versus measured aerodynamic diameter is important for consistent comparison. The two-fold overestimation of surface fine dust in the standard model is alleviated by 39 % without degradation of total column abundance by implementing a new physics-based dust emission scheme with better spatial distribution. Further reduction by 20 % of the overestimation of surface PM2.5 dust is achieved through reducing the mass fraction of emitted fine dust based on the brittle fragmentation theory, and explicit tracking of three additional fine mineral dust size bins with updated parametrization for below-cloud scavenging. Overall, these developments reduce the normalized mean difference against surface fine dust measurements from SPARTAN from 94 % to 35 %, while retaining comparable skill of total column abundance against satellite and ground-based AOD.

Validation

Dandan Zhang wrote:

I have inspected the implementation of 7 dust bins and related changes for dust scheme, emitted dust size distribution and dry/wet depositions. They all make sense to me.

For the total dust emission changes, my annual results in the year of 2018 also show emission increase of 45.3% relative to the base case (see Table 3 of Zhang et al. [2025]. The emission increase is dominated by the updated size distribution (Kok [2011]) with larger coarse dust than the default size distribution (see Figure 8a of the same paper). This is consistent with prior studies (e.g., Meng et al. [2022], Kok et al. [2017]) showing the underestimation of coarse dust and overestimation of fine dust.

For the total dust mass, I checked my annual results in the year of 2018. It shows increase of 34% than the base case. The less increase of dust mass than dust emissions is probably due to the modification of updated larger dry deposition rate as shown in Figure 9 in the paper. I also noticed a smaller increase in July 2018 from my results, showing increase of 22% than the base case for total dust mass.

For OH reduction with the increased total dust mass, it seems plausible as the heterogenous uptake of oxidants from increased dust concentration could reduce the OH concentrations.

I also confirmed that the reconciliation of aerodynamic diameter of dust is implemented in the default diagnostics of PM25 and PM10.

For the twice increase of the largest dust size bin from the updates than the default is mostly due to the new particle size distribution used. The mass fraction of each size bin is shown below:

Mass Fraction of Dust Size Bins

Bin	Size range (\(\mu m\))	Default (%)	Kok 2011 (%)	Fennec (%)
DSTbin1	0.2 — 0.36	0.0534	0.0334	0.269
DSTbin2	0.36 — 0.6	0.254	0.159	0.526
DSTbin3	0.6 — 1.2	1.90	1.19	2.54
DSTbin4	1.2 — 2.0	5.44	3.43	2.72
DSTbin5	2.0 — 3.6	19.2	12.5	11.6
DSTbin6	3.6 — 6.0	34.9	25.7	16.0
DSTbin7	6.0 — 12.0	38.2	57.0	66.4

The Kok [2011] is the particle size distribution used in GEOS-Chem 14.7.0 and later versions. The fraction of DSTbin7 is 1.5 times larger than the default, which is also shown in Figure 8a with no curving down beyond ~7 micrometer in geometric diameter of the Kok [2011] than the default particle size distribution. Combining with the increased total dust mass, the doubling of DSTbin7 looks reasonable to me.

Mass Tuning Factors for the DustL23M Scheme

Dandan Zhang generated mass tuning factors for DustL23M, in order to ensure consistent dust emission totals regardless of which combination of meteorology and horizontal resolution is used. The relevant value from the Abs. scaling factor column below will be copied into the --> Mass tuning factor entry in the HEMCO_Config.rc file that ships with your run directory.

Mass Tuning Factors - Lat-lon Grids

Meteorology	Resolution	Emissions (Tg/yr)	Rel. Scaling Factor	Abs. Scaling Factor
GEOS-FP	4x5	968.793	3.118	8.830E-03
GEOS-FP	2x2.5	1759.238	1.717	4.862E-03
GEOS-FP	0.25x0.3125	3020.560	1.000	2.832E-03
GEOS-IT	4x5	1288.555	2.344	6.639E-03
GEOS-IT	2x2.5	2336.167	1.293	3.662E-03
GEOS-IT	0.5x0.625	3547.693	0.851	2.411E-03
MERRA-2	4x5	817.376	3.695	1.047E-02
MERRA-2	2x2.5	1487.798	2.030	5.750E-03
MERRA-2	0.5x0.625	2276.232	1.327	3.758E-03

Mass Tuning Factors - Cubed-Sphere Grids

Meteorology	Resolution	Emissions (Tg/yr)	Rel. Scaling Factor	Abs. Scaling Factor
GEOS-FP	C24	1259.054	2.399	6.794E-03
GEOS-FP	C48	1943.159	1.554	4.402E-03
GEOS-FP	C90	2421.573	1.247	3.533E-03
GEOS-FP	C180	2776.547	1.088	3.081E-03
GEOS-FP	C360	2925.491	1.032	2.924E-03
GEOS-IT	C30	2021.690	1.494	4.231E-03
GEOS-IT	C48	2585.650	1.168	3.308E-03
GEOS-IT	C90	3272.238	0.923	2.614E-03
GEOS-IT	C180	3697.048	0.817	2.314E-03
MERRA-2	C24	1076.762	2.805	7.944E-03
MERRA-2	C48	2011.220	1.502	4.253E-03
MERRA-2	C90	1970.463	1.533	4.341E-03
MERRA-2	C180	2175.513	1.388	3.932E-03

Anthropogenic PM2.5 Dust Source in GEOS-Chem

Reference: Philip et al. [2017]. From the abstract:

We have added a new PM2.5 dust emission inventory (in addition to the default mineral dust simulation) into GEOS-Chem, termed as Anthropogenic Fugitive, Combustion and Industrial Dust (AFCID). In this dataset of fine anthropogenic dust, combustion and industrial sources dominate in most regions. The more commonly used term of fugitive dust is primarily coarse and not well represented in this fine dust inventory. Inclusion of AFCID improved the comparison of simulated dust and total PM2.5 mass in comparison with in situ observations (Philip et al. [2017]). Users have the option to turn on/off this inventory within the HEMCO_Config.rc configuration file.

Sea salt aerosols

The treatment of sea salt aerosols in GEOS-Chem has had two major stages of development:

1. Original formulation: Alexander et al. [2005]

2. Revised formulation: Jaeglé et al. [2011]

There have also been several additional modifications as described in the sections below.

Updated molecular weight of sea salt tracers

Molecular weights of sea salt species are 31.4 g/mol, which is consistent with the actual average composition of sea salt, and international guidelines from the IAPWS.

SST dependent sea salt emissions

Sea salt emissions now include both a wind speed and sea surface temperature (SST) dependence. The sea salt source function is based on Gong [2003], which is based on Monahan et al. [1986]. The Gong [2003] formulation expresses the density function \(dF/dr_{80}\) (in units of particles \(m^{-2} s^{-1} \mu m^{-1}\) as follows:

A and B are parameters depending on \(r_{80}\), the particle radius at RH = 80% (with \(r_{80}\) being close to twice the dry radius of sea salt particles).

Based on a comparison of GEOS-Chem sea salt simulation with coarse mode sea salt mass concentration observations obtained on 6 PMEL cruises, a new SST dependent source function was derived (Jaeglé et al. [2011]):

where \(T\) is the SST expressed in degrees Celsius (valid temperature range: 0 - 30C).

This new empirical source function leads to improved agreement of GEOS-Chem relative to sea salt mass concentration observations from cruises and ground-based stations, as well as AOD observations from MODIS and AERONET.

Recommended size range for sea salt:

• Accumulation mode: \(0.01 - 0.5\,\mu m\)

• Coarse mode: \(0.5 - 8\,\mu m\)

Note that in Jaeglé et al. [2011] we used 1 accumulation bin (\(0.01 - 0.5\mu m\)) and 2 coarse mode bins (\(0.5 - 4\,\mu m\); \(4 - 10\,\mu m\)). Due to the non-linearity of dry deposition, using a single coarse bin \(0.5 - 10\,\mu m\) leads to an overestimate of the sea salt burden, hence we recommend using \(0.5 - 8 \mu\,m\).

Updates to sea salt dry deposition

Over land, sea salt dry deposition velocities are calculated using the Zhang et al. [2001] scheme, which is based on the Slinn [1982] model for vegetated canopies. Over the oceans, we have implemented the Slinn and Slinn [1980] deposition model for natural waters. Following the recommendation of Lewis and Schwartz [2004] we assume RH = 98% in the viscous sublayer (0.1-1mm thick layer above the ocean surface). We integrate the dry deposition velocity over each size bin using a bimodal size distribution for sea salt (see below), which includes growth as a function of local RH (see below). Overall these changes lead to a factor of 3 increase in dry deposition velocity for coarse mode sea salt and a factor of 2 decrease for accumulation mode sea salt.

Updates to hygroscopic growth

The hygroscopic growth of sea salt aerosols is based on Equation (5) in Lewis and Schwartz [2006], which yields more accurate results at RH > 98% Gerber [1985] formulation previously used in GEOS-Chem.

Updates to optical properties

The size distribution of accumulation mode sea salt aerosols assumes a dry geometric radius \(r_{dg} = 0.085 \mu m\) with a geometric standard deviation \(2.03 \mu m\). This is based on cruises in the remote Pacific Ocean (Quinn et al., 1996). For coarse mode sea salt aerosols we use \(r_{dg} = 0.4 \mu m\) with a geometric standard deviation of \(1.8 \mu m\) based on Reid and others [2006].

These size distributions are used in the Mie theory calculation of extinction efficiency. They are also used in calculating the size integrated dry deposition velocity of sea salt aerosols.

Overall impact on distribution of sea salt

Implementing these changes leads to small changes in the mean global burdens of accumulation mode (20% decrease) and coarse model (25% increase) sea salt aerosols. However, the spatial changes are much larger, with a 30-50% decrease at high latitudes and a factor of ~2 increase over tropical regions. See this presentation for more information.

Sulfur and nitrogen aerosols

Reference: Park et al. [2004]. From the paper:

The sulfur simulation in GEOS-Chem is based on the Georgia Tech/Goddard Global Ozone Chemistry Aerosol Radiation and Transport (GOCART) model (Chin et al., 2000a], with a number of modifications described below. Our fossil fuel and industrial emission inventory is for 1999-2000 and is obtained by scaling the gridded, seasonally resolved inventory from the Global Emissions Inventory Activity (GEIA) for 1985 (Benkovitz et al. [1996]) with updated national emission inventories and fuel use data (Bey et al. [2001]). The emissions for the United States and Canada are from U.S. EPA [2001], and the emissions for European countries are from European Monitoring and Evaluation Programme (EMEP)/United Nations Economic Commission for Europe (UNECE). Asian sulfur emission in the model is 20 Tg S/yr, which can be compared to year 2000 estimates of 17 Tg S/yr by Streets and others [2003] and 25 Tg S/yr by Intergovernmental Panel on Climate Change (hereinafter IPCC) [2001].Anthropogenic sulfur is emitted as SO₂ except for a small fraction as sulfate (5% in Europe and 3% elsewhere) (Chin et al., [2000a]).

Other anthropogenic sources of SO₂ in the model include gridded monthly aircraft emissions (0.07 Tg S/yr) taken from Chin et al. [2000a] and biofuel use. We use a global biofuel CO emission inventory with 1° × 1° spatial resolution from Yevich and Logan [2003] and apply an emission factor of 0.0015 mol SO₂ per mole CO (Andreae and Merlet [2001]). Seasonal variations in biofuel emissions are specified from the heating degree days approach (Park et al. [2003]).

Natural sources of sulfur in the model include DMS from oceanic phytoplankton and SO₂ from volcanoes and biomass burning. The oceanic emission of DMS is calculated as the product of local seawater DMS concentration and sea-to-air transfer velocity. The seawater DMS concentrations are gridded monthly averages from Kettle and others [1999], and the transfer velocity of DMS is computed using an empirical formula from Liss and Merlivat [1986] as a function of the surface (10 m) wind speed. The GEOS surface winds used here assimilate remote sensing data from the Special Sensor Microwave Imager instrument. Volcanic emissions of SO₂ from continuously active volcanoes are included from the database of Andres and Kasgnoc [1998]. Emissions from sporadically erupting volcanoes show large year-to-year variability and are not included in the model. No major volcanic eruptions occurred in 2001. Biomass burning emissions of SO₂ are calculated using a gridded monthly biomass burning inventory of CO constrained from satellite observations in 2001 by Duncan et al. [2003] with an emission factor of 0.0026 mol SO₂ per mole CO (Andreae and Merlet [2001]).

The gas-phase sulfur oxidation chemistry in the model includes DMS oxidation by OH to form SO₂ and MSA, DMS oxidation by nitrate radicals NO₃ to form SO₂, and SO₂ oxidation by OH to form sulfate. Reaction rates are from DeMore et al. [1997] and the yields of SO₂ and MSA from DMS oxidation are from Chatfield and Crutzen [1990]. Aqueous-phase oxidation of SO₂ by O₃ and H₂O₂ in clouds to form sulfate is included using kinetic data from Jacob [1986] and assuming a pH of 4.5 for the oxidation by O₃. Cloud liquid water content is not available in the GEOS data, and we specify it instead in each cloudy grid box by using a temperature-dependent parameterization (Somerville and Remer [1984]). The cloud volume fraction in a given grid box is specified as an empirical function of the relative humidity following Sundqvist et al. [1989].

Ammonia emissions in the model are based on annual data for 1990 from the 1° × 1° GEIA inventory of Bouwman et al. [1997]. Source categories in that inventory include domesticated animals, fertilizers, human bodies, industry, fossil fuels, oceans, crops, soils, and wild animals. We view the first five as anthropogenic and the last four as natural. Additional emissions from biomass burning and biofuel use are computed using the global inventories of Duncan et al. [2003] and Yevich and Logan [2003] with an emission factor of 1.3 g NH₃ per kilogram dry mass burned (Andreae and Merlet [2001]).

Production of total inorganic nitrate (gas-phase nitric acid and aerosol nitrate) in the model is computed from the ozone-NOx-hydrocarbon chemical mechanism.

Important updates to the original formulation

Notable additions since Park et al. [2004]:

1. Biomass emissions of SO₂ and NH₃ are now computed by the GFED inventory.

   • The most recent version (Oct 2015) is GFED4

   • You may still use the older GFED2 or GFED3 inventories for research purposes.

2. Incorporation of new volcanic SO₂ emissions from Aerocom.

3. Alkalinity computation for sea salt aerosols.

4. Updates to regional and global anthropogenic emissions inventories.

5. Get liquid water content and cloud fraction directly from meteorological fields for SO₂ chemistry.

6. All species are no longer carried as per molecules sulfur or per molecules nitrogen.

7. Emissions inventories have been updated.

8. Other minor changes as described in the sections below.

Immplemented cloud Water pH for sulfate formation

From Alexander et al. [2012]:

Bulk cloud pH is calculated iteratively using concentrations of sulfate, total nitrate HNO₃ + NO₃), total ammonia (NH₃ + NH₄), SO₂, and CO₂ = 390 ppmv based on their effective Henry’s law constants and the local cloud LWC.

Over the oceans, the influence of cloud droplet heterogeneity in pH on in-cloud sulfate production rates is accounted for using the Yuen et al. (1996) parameterization. Based on isotopic evidence, this parameterization seems to work well over the oceans using sea salt aerosol as the coarse mode aerosol component, but tends to overestimate in-cloud sulfate production over land.

Implemented the Lana DMS climatology

Monthly average DMS seawater concentrations at 1° × 1° resolution (Lana et al. [2011]) are now read from disk via HEMCO.

Added sulfur oxidation by reactive halogens

Reference: Chen et al. [2017]. From the abstract:

Sulfur and reactive bromine (Bry) play important roles in tropospheric chemistry and the global radiation budget. The oxidation of dissolved SO₂ (S(IV)) by HOBr increases sulfate aerosol abundance and may also impact the Bry budget, but is generally not included in global climate and chemistry models. In this study, we implement HOBr + S(IV) reactions into the GEOS-Chem global chemical transport model and evaluate the global impacts on both sulfur and Bry budgets. Modeled HOBr mixing ratios on the order of 0.1–1.0 parts per trillion (ppt) lead to HOBr + S(IV) contributing to 8% of global sulfate production and up to 45% over some tropical ocean regions with high HOBr mixing ratios (0.6–0.9 ppt). Inclusion of HOBr + S(IV) in the model leads to a global Bry decrease of 50%, initiated by the decrease in bromide recycling in cloud droplets. Observations of HOBr are necessary to better understand the role of HOBr + S(IV) in tropospheric sulfur and Bry cycles.

Text S2 in this supporting document describes the parameterization of HOBr + S(IV) reactions in GEOS-Chem.

Added metal-catalyzed SO2 oxidation

Reference: Alexander et al. [2009]. From the author:

SO₂ is oxidized in clouds by transition metals (Fe and Mn). Natural Fe and Mn atmospheric concentrations are scaled to dust, and anthropogenic are scaled to primary anthropogenic sulfate. It is assumed that 1% of natural Mn and Fe is soluble, for anthropogenic it is 10%. The oxidation state of Fe and Mn depends on sunlight al. [2009] for more details.

Viral Shah implemented the metal catalyzed in-cloud SO₂ oxidation pathway originally described in Alexander et al. [2009]. He wrote:

My method largely follows Becky’s implementation. The main difference is that instead of using a tracer for primary sulfate to calculate anthropogenic Fe and Mn concentrations, I have added a tracer for anthropogenic Fe (pFe). pFe is emitted along with primary sulfate with an emissions ratio that equals the scaling factor used by Becky to calculate Fe concentrations from primary sulfate. This emission ratio is added as a scaling factor in HEMCO_Config and can be adjusted in the future. For wet and dry deposition, pFe is treated as an aerosol species. Anthropogenic Mn concentrations are calculated by scaling pFe concentrations. Note that Fe and Mn are also present in natural dust, and the GC dust species are used to calculate the natural Fe and Mn concentrations.

Set the molecular weights of SO4s and NITs to that of SALC

The reason for using SALC sea salt’s molecular weight for SO₄s and NITs is that these tracers are essentially internally mixed with coarse sea salt aerosol (SALC). As coarse sea salt aerosol likely dominates the mass of these aerosols, it is appropriate to use sea salt’s MW.

Another explanation is that since SO₄s and NITs are internally mixed with sea salt, they should be treated identically to SALC in the model for all processes.

Computing PM2.5 concentrations from GEOS-Chem output

For information on how to compute particulate matter (PM2.5) from GEOS-Chem diagnostic outputs, please see our Particulate matter Guide.