CMAQv5.4 Release Notes: Chemistry: Aerosol Dynamics - fisidi/CMAQ GitHub Wiki

Two-Moment Option for Particle Coagulation

Ben Murphy, U.S. Environmental Protection Agency
Type of update: Bug Fix/Science Update Release Version/Date: CMAQv5.4

Description:
The current three-moment approach for modeling coagulation for the Aitken and Accumulation modes (intermodal and intramodal coagulation) was found to be inconsistent with sectional and particle-resolved approaches via box modeling. A two-moment approach was documented by Whitby et al. (1991) and is implemented here as a default option.

Significance and Impact:
Minimal change to PM concentration predictions or code structure. Non-negligible impact on fine-mode particle sizes and number concentrations.

References:
Whitby, McMurry, Shankar, and Binkowski. Modal Aerosol Dynamics Modeling, Report to the Atmospheric Research and Exposure Assessment Laboratory, US EPA, 1991.

Internal PRs PR#879

Sulfuric Acid Conservation in VOLINORG

Ben Murphy, U.S. Environmental Protection Agency
Type of update: Bug Fix Release Version/Date: CMAQv5.4

Description:
The current implementation of sulfuric acid mass transfer from the gas to particle phase in the VOLINORG subroutine assumes that the sulfuric acid concentration leaving the subroutine equals the concentration calculated using a pseudo-steady-state assumption with the gas production rate divided by the condensation sink. This concentration could be higher or lower than the abundance of sulfuric acid predicted after execution of the chemistry model, so sulfur is not conserved. Since H2SO4 at the beginning of VOLINORG already has the gas-phase production in it from the chemistry module, a more conservative approach is now used, equivalent to: H2SO4_FINAL = H2SO4_INITIAL - [Change in ASO4]* MWSO4/MWH2SO4.

Significance and Impact: Differences in sulfuric acid concentrations at the surface up to a factor of 100, but these are always during cases where the absolute concentration is below 0.1 ug m-3. Generally dry and wet deposition decrease, but not by large amounts when run on the Southeast benchmark case.

References: Internal PRs PR#636

Update Aerosol Size Distribution Check for ICs and BCs

Ben Murphy, U.S. Environmental Protection Agency
Type of update: Bug Fix/Science Update Release Version/Date: CMAQv5.4

Description: The aerosol size distribution parameters (number, surface area, and speciated mass) from the initial and boundary conditions is read in and 'conditioned' by the aero_check_icbc routine in AERO_DATA.F. This routine checks to make sure that the total aerosol mass, number and surface area are above zero and that they combine to give appropriate modal diameter and standard deviation (i.e. within the bounds asserted in AERO_DATA.F). If the parameters do not meet the constraints imposed, then the number and/or surface area may be adjusted in order to fit reference values for diameter and standard deviation.

There are complexities to this issue that were not adequately accounted for in the current implementation. First, the current implementation only supports inputs where the surface area specifies the dry distribution, but inputs could be configured for the wet deposition and thus contradict the expectations of the conditioning module. An option is now provided in the CCTM runscript to specify the nature of the surface area input for initial and boundary conditions independently.

Secondly, it is possible that no surface area parameter is provided on the inputs. This use case is now supported with options on the CCTM runscript that control the initial and boundary conditions independently.

Third, the consideration of grid cells has been limited to just those actually on the boundaries of the grid domain. All boundary cells for interior subdomains and interior boundary cells for subdomains on the domain edge are ignored. This saves model computation time.

Finally, the logical flow has been made more thorough and robust to catch extreme errors in diameter and standard deviation, but pass minor errors (i.e. sigma = 1.04999 instead of 1.05) and correct them without warning.

Deprecated variables including AVISDIAG, ZERO_PCSOA and AOD are also removed since their functionality is no longer supported.

A related and potentially more impactful issue looms with respect to how CMAQ is treating SOA. Currently semivolatile organic species are considered 'wet' and their impact is negated when transporting aerosol surface area. It is difficult to disentangle that option from the assumptions of the boundary conditions. With this Pull Request, we will have the option to ignore the aerosol surface area from the BCs (BC_AERO_M2USE = F), which will provide an incremental step toward allowing the SOA to be treated as 'dry'.

Significance and Impact: Impact on surface concentrations and domain-wide budget were assessed for a July 2016 case. Average PM2.5 surface concentrations changed by less than 0.025 ug m-3. Average Accumulation and Aitken mode diameters changed by less than 2.5 nm.

References: N/A Internal PRs PR#874

Reclassify all SOA species as 'dry' aerosol

Ben Murphy, U.S. Environmental Protection Agency
Type of update: Science Update Release Version/Date: CMAQv5.4

Description: SOA is currently categorized as 'wet' aerosol for the purposes of calculating aerosol surface area for transport. CMAQ recalculates the surface area ignoring wet aerosol components in order to prevent unstable calculations during vertical updrafts. The approach of considering SOA as one of these wet components was made more than a decade ago and doesn't appear necessary for current CMAQ numerical stability.

This PR builds on top of PR #874 by using the improved IC/BC size distribution conditioning. It's important to note that if all organics are treated as dry aerosol, then there is an inconsistency with existing boundary conditions, which have aerosol surface area consistent with assuming the SOA is wet. Until new boundary conditions are generated with SOA as dry, it is strongly recommended to ignore aerosol surface area inputs from ICs and BCs. This is achieved with a runscript option introduced in PR #874 (IC_AERO_M2USE and BC_AERO_M2USE).

Significance and Impact:
Modest decreases to organic aerosol concentrations throughout the US domain for a July 1-31, 2016 case. Some reductions in bias, but changes are not very large.

References: N/A Internal PRs: PR#890

Simplify Aerosol Chemical Namelist

Ben Murphy, U.S. Environmental Protection Agency
Type of update: Interface Update
Release Version/Date: CMAQv5.4

Description:
Aerosol species are currently provided to the AE namelist with mode suffixes indicating the size of the particles the species is applicable for (e.g. aitken, accumulaiton, or coarse). This PR removes the suffix from all chemical species names on the AE namelist, removes duplicative species names so that only one entry for each chemical species persists, and then adds 3 columns with T/F values indicating for which modes this species should be considered. The CGIRD_SPC routine reads in the aerosol chemical species and distributes the species in modal space using a new routine added to the AERO_DATA file outside of the AERO_DATA module. This routine must be outside of AERO_DATA because AERO_DATA uses CGRID_SPCS for information as well.

Code has also been added to chemmech to process the new namelist format correctly. The dependence on the AERO_DATA table has been removed here so there are in effect fewer checks for chemmech than there are for the CCTM, but users will still encounter the checks when they eventually run the CCTM. The alternative is to have AERO_DATA tables in two places and this is not desired at this time.

Significance and Impact:
No impact on predictions should be observed.

References: N/A
Internal PRs: PR#885