General comment:

This manuscript presents an investigation of brown carbon (BrC) from wildfire emissions using aircraft plume measurements as part of the FIREX-AQ study. The manuscript has two main foci. The first involves quantifying BrC light-absorption properties and contribution to wildfire aerosol absorption using an online method (which relies on online measurements of absorption coefficients using a PAS and BC concentrations using an SP2) and an offline method (which relies on solvent extraction using both water and methanol). The second involves investigating the evolution of BrC in the atmosphere including (i) overall change in BrC absorption, (ii) comparison of overall change in BrC absorption to 4-Nitrocatechol, (iii) effect of evaporation, and (iv) the role of ozone.

The manuscript is well-written and presents comprehensive high-quality data and analyses, which constitute an important contribution to the understanding of wildfire BrC. I find the following to be particularly interesting: (i) the comparison between offline and online BrC measurements, (ii) the comparison between 4-NC and overall BrC evolution, and (iii) illustrating the complex dynamics that govern BrC evolution. Below is a list of comments that I believe should be addressed in the revised version of the manuscript.

Major specific comments:

1) The comparison between b_ap_PASBrC and b_ap_TSBrC (Figure 6): The difference in comparison at different wavelength is interesting and should be further discussed. The slope of the comparison increases with wavelength, which indicates that (i) there are insoluble BrC species and (ii) these species have a smaller AAE than the TSBrC (which leads to the wavelength-dependent comparison). As the authors point out (Page 14 Line 29), the absolute comparisons (i.e. the slopes in Figure 6) are uncertain. However, the trends are still informative. Starting with the theoretical baseline that b_ap_TSBrC cannot exceed b_ap_PASBrC (because PASBrC represents the total BrC), the slopes in Figure 6 are likely underestimates. Nevertheless, the results highlight the importance of methanol-insoluble BrC, which, according to Figure 6b and 6c, contributes an average of ~65% and ~85% of BrC absorption at 532 nm and 664 nm, respectively. This qualitatively agrees with the findings of (Atwi et al., 2022) that biomass-burning BrC can be split into a less-absorbing methanol-soluble fraction (smaller MAC and larger AAE) and a more-absorbing methanol-insoluble fraction (larger MAC and smaller AAE), with the methanol-insoluble fraction dominating mid-visible absorption.

2) Figure 3 and associated discussion: The paragraph on top of Page 13 points out the lack of correlation between MAC and AAE. This is expected because there is no substantial variability in BrC sources and combustion conditions in this study. Due of the usually encountered large spread in BrC data (due to both true variability in optical properties as well as measurement uncertainty), the inverse MAC vs AAE trend becomes apparent only when comparing different BrC categories. For instance, in Figure 1 in Saleh (2020), the inverse MAC vs AAE trend would not be apparent if only looking at one category (e.g. smoldering biomass combustion or SOA from aromatic VOCs).

With that being said, comparing the average MAC and AAE of PASBrC with those of TSBrC would be informative (see (Atwi et al., 2022)). I suggest combining the two panels of Figure 3 in one figure (using different symbols for TSBrC and PASBrC) and showing the average and standard deviation for each group. Doing so will illustrate the inverse MAC vs AAE relation. Specifically, TSBrC will be shown to have a smaller MAC but larger AAE than PASBrC, in agreement with the results of Atwi et al. (2022).

Also, because of the inverse MAC vs AAE relation, the MAC values of different BrC categories start to converge at shorter wavelengths. It is therefore more informative to present Figure 3 at 532 nm, which would better illustrate the difference between TSBrC and PASBrC.

Finally, what is the reason for TSBrC having less data points than PASBrC?

3) There are several studies that pointed out that MCE does not correlate well with aerosol light-absorption properties (e.g. (McClure et al., 2019; Pokhrel et al., 2016)). I suggest using BC/OA as a proxy for combustion conditions in Figures 3, 5, and 6.

4) The manuscript presents details of uncertainty associated with each measurement / analysis, but the uncertainty is not reflected in the figures. Uncertainties can be added as light-gray error bars, which should not have a substantial effect on the clarity of the figures.

5) The calculation of PAS BrC absorption coefficients (b_ap_PASBrC) and the corresponding MAC_PASBrC include multiple steps that should be presented in the SI. Please show:

2. a) rBC size distributions and how they were adjusted to account for rBC outside the detection window of the SP2. Typically, how much rBC was found to be outside of the detection window of the SP2?
3. b) Particle size distributions measured with LAS and the corresponding lognormal fits. How much of the particle number was outside the detection window of LAS? Also, Page 8 Line 48 states that the variability in smoke aerosol refractive index causes an uncertainty of 20%. How is this estimated? In addition to this uncertainty, does the fact that the instrument was calibrated using non-absorbing aerosol (ammonium sulfate) cause any systematic uncertainty because the smoke aerosol is light-absorbing?

6) Section 3.4.3 and Figure 11: The data collected over this short period of time (4 minutes) does not provide enough evidence to arrive at the conclusion that (i) 4-NC contribution to absorption dropped from 10% to 3% and (i) there must be BrC production (which contradicts the statement earlier in the section that “chemical aging should be negligible during this time period.”). For instance, the variability in 4-NC concentration over a period of 20 s around 1:44 is approximately twice the inferred change over the measurement period (dotted blue line). Also, I would assume that if the difference between NEMR_PASBrC and NEMR_4-NC is plotted on Figure 11, the trend would be very similar to NEMR_PASBrC (i.e. it would not show any increase in non-4-NC BrC absorption).

Minor specific comments:

1) Line 65: This statement necessitates specifying an imaginary part of the refractive index cutoff above which the OA is said to be light-absorbing. I would rephrase this sentence to reflect the fact that OA is made of components with variable light-absorption properties that vary from negligibly absorbing to strongly absorbing.

2) Line 83: By definition, DRE of a certain component is obtained as the difference in radiative balance with and without the component. The statement here that BrC contributed 7 to 48% of DRE is not consistent with this definition because these values were obtained as the difference in radiative balance with and without BrC absorption, and should more accurately be referred to as ‘BrC absorption DRE,’ not DRE of BrC (see (Saleh, 2020; Wang et al., 2018)).

3) Section 3.4.2: Please provide 4-NC optical properties used to calculate absorption coefficients.

References:

Atwi, K., Perrie, C., Cheng, Z., El Hajj, O., & Saleh, R. (2022). A dominant contribution to light absorption by methanol-insoluble brown carbon produced in the combustion of biomass fuels typically consumed in wildland fires in the United States. Environ. Sci.: Atmos. https://doi.org/10.1039/D1EA00065A

McClure, C. D., Lim, C. Y., Hagan, D. H., Kroll, J. H., & Cappa, C. D. (2019). Biomass-burning derived particles from a wide variety of fuels: Part 1: Properties of primary particles. Atmospheric Chemistry and Physics Discussions, 2019, 1–37. https://doi.org/10.5194/acp-2019-707

Pokhrel, R. P., Wagner, N. L., Langridge, J. M., Lack, D. a., Jayarathne, T., Stone, E. a., et al. (2016). Parameterization of single-scattering albedo (SSA) and absorption Ångström exponent (AAE) with EC/OC for aerosol emissions from biomass burning. Atmospheric Chemistry and Physics, 16(15), 9549–9561. https://doi.org/10.5194/acp-16-9549-2016

Saleh, R. (2020). From Measurements to Models: Toward Accurate Representation of Brown Carbon in Climate Calculations. Current Pollution Reports. https://doi.org/10.1007/s40726-020-00139-3

Wang, X., Heald, C. L., Liu, J., Weber, R. J., Campuzano-Jost, P., Jimenez, J. L., et al. (2018). Exploring the observational constraints on the simulation of brown carbon. Atmospheric Chemistry and Physics, 18(2), 635–653. https://doi.org/10.5194/acp-18-635-2018

This manuscript investigates characteristics and evolution of brown carbon (BrC) using online photoacoustic spectrometer (PAS) that measures dry aerosol absorption of fine particles and offline filter-based approach using liquid spectrophotometric measurements of extracts of particles collected on filters. They compared the measurements at different wavelengths and found that good agreement of BrC absorption at 400 nm. While doing the comparisons, there are several assumptions and limitations, but it still provides useful information and worth publishing. The study claims that investigated samples falls under moderately absorbing class. They also investigated a particular BrC chromophore, 4-nitrocatechol and its evolution with plume ages. Results indicate that 4-nitrocatechol depleted with plume ages, while other BrC was much stable even with increasing temperature in downwind. However, some previous study reported that particulate nitrophenol and nitrocatechol isomers can contribute significantly to BrC absorption at 405 nm in aged wildfire smoke.

This is an interesting study and will be useful for the community. Overall, the manuscript is clearly written, some suggested clarifications are listed below. However, prior to acceptance, the authors should address the following questions/ suggestions and modify the manuscript accordingly.

Specific comments:

The comparison between bap, PASBrc and bap,TSBrC at 405 nm looks good. It might be good to add some discussion why the PAS derived BrC absorption is higher than the TS Brc at higher wavelength. I see that the authors add some discussion about the insoluble chromophores, but it will be good add this discussion in the results section and will be easier for readers to follow.

One of the main concerns of this manuscript is that applied method rely on several assumptions and approximation which can create a large uncertainty in estimation. I appreciate that the authors stated most of the uncertainties for example in extrapolating the wavelength-dependent differences. However, I think the authors should state overall uncertainties in estimating all the absorption values. For example, I think there is a large uncertainty in estimation of the conversion factor itself. How that translate to uncertainties in total absorption?

Page 12: Some more discussion about the absorption Angstrom exponent (AAE) measured by PAS and WS BrC and MS BrC and in context to previous study would be useful, like lower AAE value reported by PAS. I’m bit confused with the AAE values from PAS and from TS reported in Figure 3. And how did the authors calculate the modified combustion efficiency?

Size distribution data and black carbon data from SP2 are missing in the manuscript but it is important to have this information in the SI.

Authors discussed several other factors that may influence the evolution of BrC. I appreciate this discussion. However, some of the supporting data on this are not shown in the manuscript.  

Summary section can be improved by proving general applicability of the closure exercise and overall applicability of this study. Another thing I find it difficult to draw some firm conclusions as this study investigated several fires with different scales, while chasing one large fire with sufficient amount time would help to decipher some of the key aspect of BrC evolution.  I think this is still an important study but something to discuss in the summary section so future work can be better designed.

Minor comments:

Page 21, line 67: please check the sentence, 405 nm wavelength was mentioned twice

Overall, there are several acronyms and subscripts, it will be difficult for readers to follow, it will be good to have a table with all the acronyms.

Is there a reason to show just show one specific fire events for the wavelength dependencies in the main manuscript? I see the value for each fire events but how about combining all the dataset to get a broader picture?

Figure 4 can be move to SI. Did you calculate the correction factor for each fire events? I suggest adding some sorts of histograms and combining all the events, so reader can get an idea about the spread of the data.