APM aerosol microphysics

From Geos-chem
Revision as of 02:01, 11 November 2011 by Fangqun Yu (Talk | contribs) (APM User Groups)

Jump to: navigation, search

This page describes the APM (Advanced Particle Microphysics) option in GEOS-Chem. APM is one of two microphysics packages being incorporated into GEOS-Chem, the other being TOMAS.

Overview

The Advanced Particle Microphysics (APM) package was developed for implementation into GEOS-Chem at State University of New York (SUNY) at Albany. In the present version of the APM module, size-resolved microphysics for secondary particles (i.e., those formed from gaseous species) and sea salt has been treated with 40 sectional bins to represent sulfate (or secondary) particles and 20 sectional bins to represent sea salt particles. The bin structure is chosen to have relatively high resolution for the size range important to the growth of nucleated particles (a few nanometers) to cloud condensation nuclei (CCN). The growth of nucleated particles through the condensation of sulfuric acid vapor and equilibrium uptake of nitrate, ammonium, and secondary organic aerosol is explicitly simulated, along with the scavenging of secondary particles by primary particles (dust, black carbon, organic carbon, and sea salt). The amounts of secondary species coated on primary particles (through condensation, coagulation, equilibrium uptake, and aqueous chemistry) are tracked.

Authors and collaborators

  • Fangqun Yu (SUNY Albany) -- Principal Investigator
  • Gan Luo (SUNY Albany)

APM User Groups

User Group Personnel Projects
SUNY Albany Fangqun Yu, Gan Luo, Xiaoyan Ma, Kevin Bartlett ...

Computational Information

The APM model contains a number of computationally efficient schemes:

  1. Usage of pre-calculated lookup tables for nucleation rates and coagulation kernels;
  2. Variable size ranges for particles of different types;
  3. Variable bin resolution;
  4. Variable and optimized time steps for the coagulation calculations;
  5. The coating of primary particles by sulfate is tracked using one tracer (sulfate mass) for each type of primary particles;
  6. Nitrate, ammonium, and SOAs asso-ciated with sulfate are calculated based on the equilibrium partition.

The above schemes enable the APM model to capture the main properties of atmospheric particles important for their direct and indirect radiative forcing while keeping the computational costs quite low.

In the study reported in Yu and Luo [2009], all simulations are running on 8-CPU Linux workstations with the 2.2 Ghz Dual Quad-Core AMD Opteron Processor 2354. The model system is compiled using OpenMP for running in parallel. The original GEOS-Chem code has 54 tracers, and it takes 24.23 hours for one year full-chemistry simulations at 4x5 horizontal resolutions and 47 layers (GEOS-5 data). The GEOS-Chem with APM model incorporated has 127 tracers (73 additional tracers: 40 for sulfate, 20 for sea salt, one for H2SO4 gas, 4 tracers for BC/OC from fossil fuel, 4 tracers for BC/OC from biomass/bio-fuel, and 4 for sulfate attached to dust, BC, primary OC, and sea salt particles). With full size-resolved microphysics (nucleation, condensation, coagulation, deposition, and scavenging) and chemistry, it takes the model (127 tracers) 52.35 hours for the same year simulations on the same machine. In other words, the efficient schemes allow the increase in the computing cost per 100% increase in number of tracers (associated with particle size information) to (52.35/24.23-1)/(127/54-1) = 86%. Such a relatively small increase in the computing cost associated with full size-resolved microphysics is desirable and makes the future coupling of APM model with global climate model feasible.

Model Details

Particle Types and Representation

 Mixing State -- Semi-externally mixed: 
    Secondary particles (SP): sulfate, plus nitrate/ammonium/SOA in equilibrium
    Primary particles:  black carbon (BC), primary organic carbon (POC), dust, and sea salt 
                        + coated SP species on each type of primary particles.

 Size structures: 
    SP: 40 bins (1.2 nm - 12 μm)
    BC: two log-normal modes for hydrophobic BC and two log-normal modes for hydrophilic BC
    POC: two log-normal modes for hydrophobic OC and two log-normal modes for hydrophilic OC
    Dust: 15 bins (30 nm - 50 μm)
    Sea salt: 20 bins (12 nm – 12 μm)

Microphysics

Nucleation

Nucleation (or new particle formation) is one of the key processes connecting gas-phase chemistry to aerosol microphysics and controlling number concentrations (and size distributions) of atmospheric particles. The APM model employs ion-mediated nucleation (IMN) (Yu, JGR [2010]) and binary homogeneous nucleation (BHN) (Yu, JGR [2008]), in term of look-up tables. Other nucleation schemes can also be included.

Based on IMN mechanism, sulfuric acid vapor concentration ([H2SO4]), temperature (T), relative humidity (RH), ionization rate (Q), and surface area of pre-existing particles (S) have profound and non-linear impacts on nucleation rates. The sensitivities of nucleation rates to the changes in these key parameters may imply important physical feedback mechanisms involving climate and emission changes, chemistry, solar variations, nucleation, aerosol number abundance, and aerosol indirect radiative forcing (Yu, JGR [2010]).

Growth

H2SO4 vapor concentration is a tracer and the condensation of H2SO4 on all particles is explicitly simulated. Many field measurements indicate significant contribution of secondary organic gases (SOGs)to the growth of secondary particles. Scheme to consider the explicit condensation of low volatile SOGs is under development (Yu, ACPD [2010]).

Coagulation

Coagulation is a process in which particles of various sizes and compositions collide with each other and coalesce to form larger particles. In the atmosphere, coagulation is an important process scavenging small particles and turning externally mixed particles into internally mixed particles. In the present model, the mass conserving semi-implicit numerical scheme is employed to solve the self coagulation of size-resolved sulfate and sea salt particles, as well as the scavenging of sulfate particles by sea salt, dust, BC, and POC particles.

Coagulation is the most time-consuming process among various size-resolved microphysical processes (nucleation, growth, coagulation, and deposition). The reason is that coagulation involves particles of different sizes and thus adds two additional dimensions (size of particle A and size of particle B) into 3-dimensional spatial grid system. For example, for 40 bins of sulfate, coagulation among sulfate particles is equivalent to solving 40x40 = 1600 reaction equations. To reduce the computing cost of 3-D sectional aerosol microphysics model, it is critical to optimize the number of bins and coagulation calculation.

Deposition and Scavenging

To be added.

Implementation notes

We have downloaded the recently released version of GEOS-Chem v8-03-01. We are in the process of incorporating the APM model into v8-03-01. We are trying to have the GEOS-Chem and APM integrated as much as possible while minimize the modifications to GeosCore codes. A single switcher (LAPM, specified in input.geos) has been designed to turn on/off APM processes. When LAPM = F, the model should reduce to the original GeosCore simulation. We are still optimizing and testing the code. We will get the GEOS-chem + APM to the GEOS-Chem support team as soon as possible.

Updates:

  • 2010/08/25: Finally we got the APM properly incorporated into GEOS-Chem v8-03-01. The switcher LAPM appears to work very well and the results look reasonable. We are in the process of doing final clean-up and adding comments.
  • 2010/09/15: A new SOA aging and condensation scheme has been developed based on v8-02-02 (Yu, ACPD [2010]). We have been trying to add new SOA scheme into version v8-03-01 which has additional anthropogenic organic species. More time is needed to make it work properly and evaluate the results. Considering that we have already been later in getting the code to GEOS-Chem support team and per Bob's suggestion, we decide to send the working code (v8-03-01 + APM without the SOA aging and condensation scheme) to the support team first and add SOA condensation scheme separately in the future.
  • 2010/09/23: Following Bob's suggestion, Gan is in the process of adding and testing "#if defined( APM )" in the code to block out APM-specific sections. Will need some additional time to finish the task. Upon completion, the APM code will be ignored during compilation if we want to run the original version. If we compile the code with APM, we will still have switcher available in input.geos to turn on/off APM.
  • 2010/10/18: Gan has finished adding "#if defined( APM )" to the integrated GEOS-Chem + APM. It took some additional time to re-arrange/re-organize the code in a way to facilitate the adding of "#if defined (APM)-- #endif" blocks. In addition, we spent sometime to move the APM code in the previous GEOS-Chem version to the most recent version(v8-03-02). Based on preliminary tests, the model works pretty well. The fully integrated APM model should be able to run under all the configurations that original GEOS-Chem v8-03-02 can run. Fangqun is doing the final cleanup and evaluation. We expect to deliver the integrated code to Bob shortly.
  • 2010/11/03: The APM code fully integrated into GEOS-Chem-v8-3-2 has been delivered to Bob.
  • 2011/2/17: Bob has finished merging APM into the most recent GEOS-Chem version under development (which will be released as v9-01-02). There were many changes between v8-03-02 and v9-01-01, so Bob decided to wait until v9-01-01 was stable before proceeding. We are in the process of evaluating the merged code.
  • 2011/3/10: There were some bugs in the merged code which took us quite some time to debug and fix. We have finished 1 yr 4x5 simulation (GEOS5 met, 47 layers). Preliminary analysis indicates that the aerosol results are reasonable. There exist some differences between results of this version and those of previous versions, but these differences are expected considering the substantial changes/updates in v9-01-02. We are doing some final checks and will get the code back to Bob shortly.
  • 2011/3/11: We have finished the evaluation of the GEOS-Chem v9-01-02 APM simulations and sent the final code back to Bob.
  • 2011/3/15: APM is integrated into GEOS-Chem v9-01-02.
  • 2011/8/23: APM is re-integrated into GEOS-Chem. 1-month benchmark simulation v9-01-02l ensured that the import of APM did not break the standard chemistry simulation.
  • 2011/9/7: Fixed some outstanding minor issues in APM (bug fixes from G. Luo).

--Bob Y. 15:30, 7 September 2011 (EDT)

Validation, Application, and Development

The first application of GEOS-Chem + APM focuses on predicting the number concentrations of particles in the troposphere. Significant amount of efforts have been devoted to validate the simulated global spatial distributions of particle number abundance, using a large amount of land-, ship-, and aircraft- based measurements. See following publications for details (full citations are given in the in the References section).

  1. Yu and Luo, ACP [2009]
  2. Yu et. al., JGR [2010]

The model has been applied in a number of other studies and results have been reported in the following papers:

  1. Luo and Yu, ACP [2011]
  2. Luo and Yu, ACP [2010]
  3. Yu and Luo, Atmosphere [2010]

Ongoing Developments:

  1. Extended SOA formation model that considers oxidation aging and kinetic condensation: Yu, ACP [2011].
  2. Scheme to calculate aerosol optical properties based on size, composition, and mixing state resolved particle properties predicted by APM.
  3. One-column radiative transfer module to calculate online direct radiative forcing of various aerosols predicted by APM.
  4. Parametrization of size-resolved primary sulfate emission as a function of OH concentration, temperature, condensation sink, and other parameters (Yu, EST [2010]).

Validation plots:

  1. This plot shows simulated annual mean [SO2], [H2SO4], nucleation rates, and total number concentration of particles larger than 10 nm (CN10), averaged in the boundary layer. This is from a 1-year, 4x5 simulation for 2005 done by the SUNY/Albany group.

--Bob Y. 11:42, 15 March 2011 (EDT)

AEROCOM intercomparison

The AEROCOM-project is an open international initiative of scientists interested in the advancement of the understanding of the global aerosol and its impact on climate. We have made a commitment to represent the GEOS-Chem community to submit aerosol microphysics simulations based on GEOS-Chem + APM.

We submitted our size-resolved microphysics simulation results for A2-CTRL-2006 (GEOS-5, 2x2.5) to AEROCOM in January, 2011. The submitted results include OC and SOA mass concentrations which have been used in AEROCOM OA intercomparison. We also managed to submit AOD, AAOD, aerosol direct radiative forcing results to AeroCom in September 2011.

References

  1. Luo, G., and F. Yu, Sensitivity of global cloud condensation nuclei concentrations to primary sulfate emissions parameterizations, Atmos. Chem. Phys., 11, 1949-1959, doi:10.5194/acp-11-1949-2011, 2011. PDF
  2. Luo, G., and F. Yu, A numerical evaluation of global oceanic emissions of alpha-pinene and isoprene, Atmos. Chem. Phys., 10, 2007-2015, 2010. PDF
  3. Yu, F., Updated H2SO4-H2O binary homogeneous nucleation rate look-up tables, J. Geophy. Res., 113, D24201, doi:10.1029/2008JD010527, 2008. PDF
  4. Yu, F., Ion-mediated nucleation in the atmosphere: Key controlling parameters, implications, and look-up table, J. Geophys. Res., 115, D03206, doi:10.1029/2009JD012630, 2010. PDF
  5. Yu, F., A secondary organic aerosol formation model considering successive oxidation aging and kinetic condensation of organic compounds: global scale implications, Atmos. Chem. Phys., 11, 1083-1099, doi:10.5194/acp-11-1083-2011, 2011. PDF
  6. Yu, F., Diurnal and seasonal variations of ultrafine particle formation in anthropogenic SO2 plumes, Environmental Science & Technology, 44 (6), 2011-2015, DOI: 10.1021/es903228a, 2010. PDF
  7. Yu, F., and G. Luo, Simulation of particle size distribution with a global aerosol model: Contribution of nucleation to aerosol and CCN number concentrations, Atmos. Chem. Phys., 9, 7691-7710, 2009. PDF
  8. Yu, F., and G. Luo, Oceanic dimethyl sulfide emission and new particle formation around the coast of Antarctica: A modeling study of seasonal variations and comparison with measurements, Atmosphere, 1(1):34-50, 2010. PDF
  9. Yu, F., G. Luo, T. Bates, B. Anderson, A. Clarke, V. Kapustin, R. Yantosca, Y. Wang, S. Wu, Spatial distributions of particle number concentrations in the global troposphere: Simulations, observations, and implications for nucleation mechanisms, J. Geophys. Res., 115, D17205, doi:10.1029/2009JD013473, 2010. PDF

Known issues

None at this time.