Liang et al. measured gas-particle partitioning of 89 compounds through SV-TAG and cTAG either by measuring the quantity of each compound in the gas and particle phase, or the particle phase only and taking the difference to obtain the concentration of the compound in the gas phase. By measuring the gas phase and particle phase, the measured fraction in the particle can be compared to the saturation mass concentration to obtain information about gas-particle partitioning for individual compounds in the ambient atmosphere. The main body of the paper looks at gas-particle partitioning in an urban environment with influence from biomass burning. This paper demonstrates that polar compounds partition into biomass burning organic aerosol more readily than nonpolar compounds, which preferentially evaporate into the gas phase with increasing BBOA.

The measurement technique is well-characterized, and supported with additional measurement techniques to ensure unambiguous identification and quantification of the reported compounds. The findings are unique in that they report direct measurements of gas-particle partitioning in the ambient atmosphere with some mechanistic understanding of the change in activity coefficients in the presence of different organic aerosol. The paper is well organized with appropriate figures. This paper should be published after responding to the comments below.

General comments:

The analysis, interpretation, discussion and conclusions in this manuscript do not address any uncertainty within the vapor pressure group contribution methods used to estimate the saturation mass concentration of the identified compounds. The interpretation that the activity coefficient is solely responsible for differences between measured and expected fraction of compound in particle makes sense mathematically, but relies on the assumption that the uncertainty in the group contribution methods used is small relative to the total discrepancy observed. Evaluations of the uncertainty of SIMPOL have been conducted in e.g. Bilde et al., 2015, Chem. Rev., 4115-4156 and Valorso et al., 2011, Atmos. Chem. Phys., 6895-6910, and there is evidence that vapor pressures of multifunctional compounds tend to be inaccurate when by calculated SIMPOL and other group contribution methods in e.g. Barley and McFiggans, 2010, Atmos. Chem. Phys., 749-767, Dang et al., 2019, Aerosol Sci. Tech., 1040-1055, and O’Meara et al., 2014, Phys. Chem. Chem. Phys., 19453-19469. The error in the calculated saturation mass concentration could produce deviations from gamma = 1 that are not related to the activity coefficient. A discussion of the uncertainty of the group contribution methods is critical for the interpretation of the data presented in this manuscript, and must be included. The discussion should include whether the uncertainties in the group contribution methods are random or systematic, and how this can impact the conclusions stated in the manuscript.

Minor Comments:

Pg. 2 line 48: Finewax et al. did not assume that nitrophenol compounds were only in the particle-phase, but measured the saturation mass concentration of 4-nitrocatechol, C* = 13 ug m^-3, and concluded that it was almost entirely in the particle phase of the particle concentrations used in their laboratory study. This was supported by Fredrickson et al., 2022, ACS Earth & Space Chemistry, which reported measured vapor pressure of 2.4 – 12 ug m^-3. Please revise this, and include the Finewax citation in the discussion of 4-nitrocatechol vapor pressure on pg 7 lines 199-208.

Pg. 6-7 Lines 189-193: Please cite literature e.g. Dang et al., 2019, Aerosol Sci. Tech., 1040-1055, to demonstrate precedent of different isomers having significantly different vapor pressures.

Pg. 9 line 225: Are activity coefficients of 10^-3 expected? Please include literature values of gamma to support this finding.

Pg. 10 Figure 2: It’s not unambiguous that Figure 2B demonstrates that the gas-particle partitioning of compounds is better described by the model in the strong BB case, merely that compounds are closer to the gamma = 1 line. This could suggest that predicted vapor pressure based on group contribution methods fail to accurately determine the vapor pressure of multi-functional compounds, or it could suggest that gas-particle partitioning may not be instantaneous for BBOA at these atmospheric conditions (e.g. RH, temperature). It could demonstrate the uncertainty in the group contribution methods for calculating compound vapor pressure. I believe that the statement on pg 10 lines 252-253 does not necessarily follow without also assuming that the uncertainty in the vapor pressure calculations are very small.

Pg. 15 Lines 340-346: Viscosity discussion should include the age of the biomass burning organic aerosol. Based on viscosity measurements of BBOA in the literature, and the size distribution of the particles, can the gas-particle partitioning timescale be calculated? Based on the timescale calculated and the age of the smoke plume, the authors should be able to determine whether gas-particle partitioning has reached equilibrium as it approaches the sampling site.

Technical Comments:

Figure 2: Legend should be placed elsewhere to not block x-axis label.

Figure S13: These compounds are all non-polar. Revise the caption.

The authors determined the particle-phase fraction (Fp) of the individual semivolatile organic compounds measured by two TAG instruments and looked at the gas-particle partitioning behavior of SVOCs under the influences of biomass burning at two locations in two separate years. The manuscript provides insight into the phase partitioning of SVOCs and factors influencing the behavior in wildfire plumes. I recommend the manuscript for publication after addressing the following comments:

Specific comments:

1.       Please explain why out of the many SVOCs identified, the authors selectively discussed a few compounds in the text (bold in Figure 1) in Section 3.1. It is not clear what’s special about these compounds or their significance; why discuss these over the others. In addition, the authors claimed in Lines 175-176 and lines 200-201 that the higher observed Fp makes levoglucosan and nitrocatechols etc “very good BB marker compounds”. I would provide references that these compounds are exclusively emitted/formed from biomass burning with negligible other sources.  Please also explain why high Fp makes them good BB markers.

2.       The evaluation of the performance of the equilibrium absorptive partitioning model would benefit from additional explanation/description. It is unclear what criteria were used to assess the predictive capabilities of the model and why.

a.        Line 215-217, comparing C* vs C^o does not provide insight into the performance of the equilibrium absorptive partitioning model. I would clarify that calculated C* using both the measured Fp and the predicted Fp, and did the comparison between these two sets of results.

b.       Line 225-226. Why does the measured data falling between the lines of ү=1 and ү=10^-3 in Figure 2 suggest that the model underpredicts the F_p?

c.       Line 244-245. It is hard to follow why the data points being closer to the line of ү=1 suggests a better-described phase partitioning behavior of compounds.

d.       Line 252-253. The activity coefficients were included in the calculation of predicted C*. I don’t see any cases illustrating that excluding activity coefficients leads to a worse prediction.

3.       Did the authors look at the Fp of SVOCs under non-BB periods? How did that compare to different BB scenarios? The title highlighted “when wildfire smoke comes to Town”. However, the manuscript only presented/discussed results under BB influences. It will be useful to add the non-BB scenario to the figures and table (Figure 5, Figure S12 – 15, Table 1) and discussions in Section 3.3 to set the base case of “background” air in Town to reflect the title.

4.       The manuscript used data collected during two studies: the 2017 Berkeley wildfire and the 2018 McCall FIREX-AQ. However, the main text mainly presented the 2017 Berkely wildfire, and the discussions were mainly focused on the Berkely results too, except for the Radom forest algorithm results where the 2017 Berkely study had limited data. Can the authors discuss the 2018 results more, and comment on the similarities and differences between the 2017 vs 2018 results and any insights gained from comparing the two studies?

Minor comments:

1.        How many compounds in total were measured/identified from these two data sets? I would state it at the beginning of the results.

2.        Lines 159 – 161, I would be more specific with the range of ү that indicates different phasing.

3.        Line 203, please clarify what “half of these compounds would stay in the gas phase” mean. Half number, half mass?

4.        Line 322-323.  I disagree with the statement. Figure S15 shows that f44 and O/C values under lower BB influences had much broader distributions than those under stronger BB influences. This on the contrary suggests that the background aerosols when BB influence was low had different compositions from the BBOAs at McCall. Did you check the tracers for cooking and traffic aerosols?

Technical comments:

1.        Figure 2, I would clarify in the caption that the markers are the C* derived from the measured Fp (C* was not directly measured as stated in line 255), and that the lines are the C* derived from the predicted Fp.

2.        Figure 3, the upper right half of the matrix is the repetition of the lower left half of the matrix. This is redundant and confusing. I would show only half of the matrix. The same applies to Figure S8

3.        Figure S3, the left axis label should be Fp. The legend of the figure should be modified to include “Measured” for the markers of Alkane and Acid, and “Predicted” for the lines with different gamma values.