Development of Ozone Reactivity Scales for Volatile Organic Compounds in a Chinese Megacity

ACP-2021-44

Zhang, Xue, Carter, Pei, Chen, Mu, Wang, Zhang and Wang

Zhang et al. describe a comprehensive analysis of reactivity scales for a Chinese megacity, Guangzhou. They use two methods to characterise base conditions for their Master Chemical Mechanism (MCM) box model. In one, model inputs are based on observations and in the other they are based on emissions. Four reactivity scales were determined using base conditions, maximum incremental reactivity (MIR), maximum ozone reactivity (MOR) and equal benefit incremental reactivity (EBIR) conditions. MOR and MIR reactivity scales estimated for Guangzhou were compared with the corresponding scales for USA conditions using the same chemical mechanism and with the SAPRC-07 mechanism. Sensitivity tests were performed to investigate the influence of environmental conditions on the estimated reactivity scales.

This is a comprehensive study which is well worth publishing in ACP. It should provide the essential data required in the formulation of strategies for tackling elevated ozone levels across China. I particularly liked Figure 2 and how the study links the MIR scale to VOC-limited conditions, MOR to mixed and EBIR to NO_x-limited conditions.

The study fulfils an important service in providing in Table 1 the IRs for all 116 VOC species in the MCM under the four sets of conditions. Elsewhere in the study more use should be made of relative incremental reactivities, by expressing them relative to a specific VOC such as ethene. Zhang et al. introduce the concept as MIR/Ethene though much more use could be made than the brief mention in section 3.2 (lines 227 onwards). The advantage of relative reactivities is that they clearly reveal small differences between reactivity scales. When discussing the influence of background conditions (lines 270 onwards), these would be much clearer if they were presented as ratios to ethene: IRs/IR_ethene. Table 2 should be replaced with Tables of IR/IR_ethene values then we could see if the background conditions really changed the reactivities for particular VOCs. Also, in Figures 5 and 6, we see the differences in reactivity scales between Guangzhou and USA. But these would be much more illuminating if they were presented as IR/IR_ethene values rather than as ranks. Ranks disguise the magnitudes of the differences.

In reading through the preprint, some trivial issues were noted.

1. line 30: Agathokleous et al. is not the best reference to give here. There is an excellent reference available from the Tropospheric Ozone Assessment Report in Elementa.
2. line 47: What is meant here? A mechanism is either explicit or not. If it contains non-stoichiometric chemical equations then it is not explicit.
3. line 72: ‘results obtained’.
4. lines 86-87: It is widely recognised that the diurnal cycle in ozone is caused by changes in the stability of the boundary layer and not by intense in situ photochemical ozone production. This is explained in the Tropospheric Ozone Assessment Report.
5. line 93: ppbv ppbv^-1.
6. line 99: Reference to the MCM website at the University of York would be more up-to-date.
7. line 112: Where does these initial concentrations come from if not from observations? There are many sets of OVOC data for China.
8. line 128: Replace ‘folds’ with scaling factors.
9. line 134 onwards: please explain what the ‘base’ scenario is.
10. line 176: The chemspider website reference provides rate coefficients presumably and not reaction fluxes.
11. line 238-242: It would be exceptionally useful if a little more detail was given here about why the five outliers are difficult to represent in chemical mechanisms. Benzene and styrene, presumably like phenol, are strong inhibitors of ozone formation. Is there a simple explanation how this works mechanistically. Then we have n-octane through n-decane. Presumably the mechanism of inhibition is different here and it would be interesting to know why this is. Why does it begin with n-octane and not n-heptane?
12. line 241: This point would be self-evident if the presentation had been done with MIR/MIR_ethene ratios.
13. line 244: MIR
14. line 250: What was done with one quarter of the base ratios?
15. Section 3.2: This is a big section that would benefit considerably from being split into smaller sub-sections.
16. line 336 onwards. This is an important recommendation and should be in the Conclusions section with a little more explanation.

In this paper, Carter's method and local data of Guangzhou were applied to construct new MIR, MOR and EBIR scenarios under observation and emission methods through a box model equipped with MCMv3.3.1 mechanism. At the same time, the MIR-characteristics of VOCs are described and analyzed. The application of recent research results to achieve migration is meaningful and has universality in application. At the same time, the research needs to carry out data processing for each VOC, and the workload is heavy, which is worthy of recognition. After review, it is considered that the article still needs to address the following concerns:

1. For the box model, the observation based method adopted by the author to constrain the concentration of both NO₂ and NO at the same time, which made the O₃ concentration largely fixed by the ratio of NO₂/NO. In this case, could the impact of VOCs on O₃ be reasonably reflected?
2. The running step of the box model is 1h, is it too sparse for the total integral period of 10h? In the paper, it may be necessary to include the graphs or tables of O₃ concentration changes in the two observation-based and emission models within 10 hours.
3. The MIR table of VOCs species is suggested to refer to the article published by Carter in 2007, which is arranged in the order of commonly used alkanes, olefins and aromatic hydrocarbons.
4. Fig. 5a is one of the most important conclusions of the whole paper, which is used to compare the MIR-values of the article and the MIR-values in the literature. Compared with the logarithmic axis, the comparison results of the conventional axis are more convincing. At the same time, scatter plots similar to the size order in Fig. 5b should be reduced, because the deviation of MIR order in a considerable number of VOCs species is large, and R² is of little significance. This paper needs to further prove the validity of the calculated MIR in Guangzhou, so it can be used to replace the MIR from US (mainly Carter’s publication) for Guangzhou.
5. In Table2a, why change the simulation time to 3 days? 10h can describe the period of time during which VOCs receive light and undergo photolysis reaction cycle in real day, and the situation of 3 days lacks practical significance. However, for the MIR scenario and MOR scenario, there were differences in RMA slope changes in the 3-day simulation, and some R² were too small. It needs to be clarified.