Advanced modeling of gas chemistry and aerosol dynamics with SSH-aerosol v2.0
Abstract. SSH-aerosol is developed to represent the evolution of primary and secondary pollutants in the atmosphere by processes linked to gas-phase chemistry, aerosol dynamics (coagulation, condensation/evaporation and nucleation) and intra-particle reactions. The representation of process complexity can be adjusted based on the user's choices.
The model uses a sectional size distribution, and offers the capability to discretize chemical composition to account for the mixing state of particles.
The algorithms are designed to represent the evolution of ultrafine particles: conservation of mass and number during numerical resolution, taking into account the Kelvin effect, the condensation dynamics of nonvolatile compounds, and nucleation. Different parameterizations are provided for nucleation: binary, ternary, heteromolecular and organic nucleation depending on the compounds involved.
For gas-phase chemistry, schemes of different complexities can be handled: from simple schemes to model ozone, oxidants and inorganic chemistry (e.g. CB05, RACM2, Melchior2), to more complex schemes, e.g. from the Master Chemical Mechanism (MCM). The complexity of the schemes used for SOA formation may also be adjusted: from schemes built from chamber data to near-explicit schemes from MCM. SOA schemes reduced using the GENOA algorithm are also provided for several precursors (toluene, a sesquiterpene and three monoterpenes), together with their evaluation against chamber or flow-tube experiments. A wall-loss module has also been added for easier comparisons to chamber experiments.
Specific developments were made in version 2.0 to automatically link the chosen gas-phase mechanism to SOA formation by using the SMILES structure of organic compounds, allowing for the determination of their hydrophilic and hydrophobic properties and for the partitioning in both organic and aqueous phases. The gas/particle partitioning may also be represented with different complexities. For the organic phase, viscosity may be modelled, adapting the aerosol viscosity to its composition, and coupling organic and inorganic thermodynamics. The dynamic evolution of the partitioning may be computed explicitly or thermodynamic equilibrium may be assumed. Different options are also provided to simulate the chemistry of organic compounds inside the particles with different types of reactions: irreversible first order reactions, bulk oligomerization, hydratation of aldehydes and reactions of organic compounds with inorganic ions.
The SSH-aerosol model may be installed with a docker for standalone use. It has also been coupled to several 3D models to represent gas and aerosol concentrations: from the local scale with computational fluid dynamic and street network models to the regional scale with chemistry-transport models.
The paper by Sartelet and co-workers describes the new developments in the established aerosol model "SSH" (Sartelet et al., 2020), now termed SSH-2.0. The additions and changes to the model are comprehensive, which justifies publication of this update. Code is provided on github and Zenodo. One of the biggest advantages of this very complete aerosol model is its capability to be run both as box model to describe laboratory experiments, and as module in larger scale models. The authors added functionalities that seem helpful for both tasks. The authors have a difficult task at hand with this paper to describe the full capability of the model. This work details the development of advancements in gas-particle partitioning, viscosity and particle-phase reactions into the model, which I think is a good choice. However, I think the paper would benefit from some general clarifications about the SSH model and how it works (even though this has been established in previous work), possibly aided by some diagrams. After the questions and comments below are addressed, this will be a nice paper that fits well within the scope of GMD.
General Comments
Some of the new model capabilities advertised in Abstract and Conclusions are not sufficiently discussed in the paper, especially the treatment of new particle formation / nucleation and ultrafine particles. It would be good if these points would be strengthened.
Something that would be really instructive in this study would be to give the reader an idea how much more time consuming the model gets when the complexity is increased, e.g. by including more complex chemistry (e.g., Figs. 2, 3 etc.) or including particle-phase effects (e.g., Fig. 8). Could you time the model and report the runtime?
I think it would be generally worthwhile to add a short, general section on the numerical implementation of the model code (also to prepare Sect. 4.1, which details some of the improvements in SSH-2.0):
- I believe this is FORTRAN code?
- Which ODE solvers are being used?
- How is numerical convergence assured?
- Can the user change numerical integration tolerances?
- Are there problems with numerical stiffness, if yes, how can they be dealt with? For example, can evaporation lead to the full vanishing of layers?
For many of the examples, only the primary model output (e.g., SOA mass) is provided. I think this paper would be showing the capabilities of SSH-2.0 better if you would also show other model output, such as the concentrations of certain product or intermediates, particle size distributions, or diffusion gradients inside the particles.
It does not become clear enough how far the model can be customized by defining one's own particle-phase chemistry, or if only the four types of particle-phase chemistry that is described in section 4.3 and 4.4 can be used.
The English langugage could be improved at times (especially singular and plural nouns - for example, "dynamics" and "physics" are used as plural nouns in the English language), but language copy-editing will take care of that.
Overall, the methods in the paper are a bit difficult to understand at times or remain unexplained. I tried to outline the concepts which I find difficult in the specific comments below.
Specific Minor Comments
l. 47 - The sectional approach is not well described in this paper (despite being mentioned in the abstract). Is the approach to size distributions in this model using fixed or moving bins? This is important for discussung the implementation of nucleation in Sect. 5.1 later. In l. 342, I think size bins are mentioned for the first time without prior introduction. It would be good to clarify how size distributions work in the model.
l. 68 "oftheir" -> "of their"
l. 106 What is this "collision factor"?
l. 162 "isoprene SOAs prefer to condense onto aqueous phase" - I believe this is not clear. With "SOAs" you mean oxidation products here?
l. 211-224: This paragraph is hard to read for people not deeply familiar with these reaction schemes. What I find most confusing is the use of "oxidant" in quotation marks, describing the schemes (?), alongside oxidant without quotation marks. Why are they called "oxidant" schemes?
l. 254 "The influence of the RO2 pool option is low for BCARY." - Why is this the case?
Figure 2 - "SOA yield" is not defined in the manuscript.
Figure 3 - Please explain "Rdc" and "Expl." somewhere.
l. 282 "This is partly attributed to the impact of molecular rearrangement present in the near-explicit scheme, but which is missing in the H2O scheme." - This information does not help the reader understand what is going on. I would suggest to explain it briefly so it can be understood, or only refer to the original study.
l. 294 - "(upper left panel of Fig4)" - I believe this should read "(upper right panel of Fig. 4)". Because you explain first the upper right panel in the figure caption, I believe the order was reversed at some stage in the paper writing process. I think it would be good to discuss the figures from left to right.
Figure 6 - Since this chapter is about wall losses: what is the effect of the gas and particle wall loss parameterizations here?
l. 364 - What is "radical equilibrium"?
l. 381--385 - Figure 8 is hardly discussed in the paper. Figure 8 shows a better model-experiment agreement when using AIOMFAC-visc. Can this be understood? Is Naphthalene SOA expected to be more viscous than toluene SOA (Fig. 7)? What would the effect of using AIOMFAC-visc be on the calculations shown in Fig. 7, does it also affect this calculation?
l. 394 - "As in Couvidat and Sartelet (2015) and in SSH-aerosol v1.1, it does not represent the exchange of compounds between layers but assumes a characteristic time to reach the interface." - Can the authors comment why this implicit layer-to-interface approach has been chosen over a more explicit layer-to-layer diffusion approach? Do I understand correctly that the diffusion to the interface is only determined by the composition of the origin layer, not the composition of the layers between this layer and the interface? If yes, the authors should comment on the potential shortcomings of this method. A schematic figure might help the reader to visualize how diffusion is treated in the model.
Equation 10 - The use of long words in italic font as indices make the equations quite hard to read. Generally it is customary to use italic letters for variables representing numbers, but upright characters for words that do not ("diff"). It is not clear what bin, p and o stand for. What is the unit of J? Typically, flux is expressed as molecules per time and area (e.g., Pöschl, Rudich, Ammann 2007). Is this the case here (and e.g. in Eq. 23)?
l. 413 "By B^bin,layer_p,i(t+Δt) is the concentration estimated with Eq. 11" - Do you mean that the result of Eq. 11 (A^bin,layer_p,i(t+Δt)) is inserted here? The use of A and B for (same?) concentrations is highly confusing here.
l. 417 - What does "without diffusion" mean here?
Equation 20 and l. 451: R_diff should probably be R_diff,c (as in Eq. 19).
l. 458 - What does "default structure" mean here?
l. 465 - "However, at RH=70%, the aerosol is inviscid" - Is the viscosity truly zero here in the calculations, which is what inviscid means, or do just mean its low enough to not affect condensation?
Figure 9 - Why are the initial conditions not the same for these simulations? How is the model initialized? Is the POAmP concentration in the gas phase fixed (open system)?
l. 483 - It is not clear what is meant with "This type of reaction can account for catalysis by water and pH." - Can the rate coefficient of this reaction be expressed in dependence of humidity and pH?
l. 530 - "The influence of hydratation is assessed in the Platt isoprene test case with ammonia emission (see previous section)." - Where can this be seen?
l. 536 - "water of pH" should likely read "water or pH"
Equation 27 - "H_eff" is incorrectly subscripted
l. 573 - "the Henry constant of OH is adapted from Sander (2015) leading to OH aqueous concentration in agreement with typical mean value in the atmosphere from Herrmann et al. (2010)." - It's not clear what the importance of OH is in this section.
l. 588 - It is not clear what is meant with "inorganic aerosol are assumed to be metastable".
l. 593 - "as shown in the guide" - Here and in other places: I think this needs better referencing. What document is this, where can it be found (URL). The reader might benefit from an explicit page / section reference. In general, I don't think it makes much sense for this paper to describe a calculation, but then not show results.
Figure 13 - It is very hard to see and distinguish all different lines. The legend shows four different experiments, but only 3 are visible in the figure (dotted, dash-dotted, solid). From the legend, it is not clear which simulation the dash-dotted line refers to.
l. 615 - "An algorithm is used to redistribute the bulk concentrations between the size sections, taking into account the Kelvin effect." - This is not clear, please explain.
Section 5.1 - From this section, it does not become clear enough how nucleation works in the model.
- Does nucleation always occur into the smallest size bin?
- How is composition of the nucleation size bin treated when nucleation occurs?
- Can you show / visualize the evolution of the size distribution somehow?
- Is nucleation happening in the calculations for Fig. 15?
References
Pöschl, U., Rudich, Y., and Ammann, M.: Kinetic model framework for aerosol and cloud surface chemistry and gas-particle interactions – Part 1: General equations, parameters, and terminology, Atmos. Chem. Phys., 7, 5989–6023, https://doi.org/10.5194/acp-7-5989-2007, 2007.
Sartelet, K., Couvidat, F.,Wang, Z., Flageul, C., and Kim, Y.: SSH-Aerosol v1.1: A Modular Box Model to Simulate the Evolution of Primary and Secondary Aerosols., Atmosphere, 11, 525, https://doi.org/10.3390/atmos11050525, 2020.