the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Aug 2026
Measurement report: Long-term 2D MAX-DOAS observations of HONO vertical distributions in a rice–wheat rotation agricultural region in Shouxian, China
Abstract. Nitrous acid (HONO) photolysis is an important source of hydroxyl radicals (OH) in the atmospheric boundary layer, but long-term observations of HONO vertical distributions and their responses to agricultural activities remain limited. We conducted nearly 2 years (2022–2023) of two-dimensional multi-axis differential optical absorption spectroscopy (2D MAX-DOAS) observations in a rice–wheat rotation region in Shouxian, China. Retrieval reliability and stability were assessed through an intercomparison between two synchronized systems, spectral fit quality, degrees of freedom for signal (DOFS), and averaging-kernel analysis. Meteorological data, farming records, backward trajectories, potential source contribution function (PSCF) analysis, and Tropospheric Ultraviolet and Visible (TUV) calculations were integrated to investigate HONO variability, fertilization effects, and potential OH production. HONO was concentrated below 0.5 km, with morning and late-afternoon peaks and a noon minimum. Seasonal means followed the order winter > autumn > spring > summer. In spring and summer, near-surface HONO decreased toward midday, whereas HONO increased aloft, suggesting possible radiation-associated replenishment in elevated layers. During the 10 d following fertilization on 15 March 2023, HONO concentrations at 0.05, 0.20, and 0.50 km increased by factors of 5.3, 3.7, and 2.5, respectively, relative to the preceding 10 d. Near-site PSCF values suggested local and short-range agricultural influences. At 0.05 km, potential OH production from HONO photolysis increased after fertilization by 710 % under clear-sky conditions and by 86 % under cloudy conditions. These findings demonstrate the importance of resolving HONO vertical structure and agricultural-event responses when evaluating its photochemical relevance in agricultural regions.
- Preprint
(7387 KB) - Metadata XML
-
Supplement
(1177 KB) - BibTeX
- EndNote
Status: open (until 14 Nov 2026)
- RC1: 'RC1: Comment on egusphere-2026-4334', Anonymous Referee #1, 13 Oct 2026 reply
-
RC2: 'Comment on egusphere-2026-4334', Anonymous Referee #2, 08 Nov 2026
reply
General comment
This study addresses HONO sources in an agricultural region, an important topic for understanding the contribution of HONO to the oxidative capacity of the atmospheric boundary layer. MAX-DOAS is a well-established technique for such investigations, and the two-year dataset is valuable. However, several methodological issues need to be addressed before robust conclusions can be drawn. In particular, the profile retrievals require further evaluation and improvement to enhance their sensitivity to HONO and NO2 at higher altitudes. Because clouds can substantially affect profile retrievals, an appropriate cloud-screening procedure should be applied before statistical analysis. Furthermore, given the low HONO abundances, a rigorous uncertainty assessment is essential to establish the robustness of the statistical results and their interpretation. If the vertical sensitivity of the retrievals cannot be improved sufficiently, I recommend focusing the statistical analysis on vertical column densities (VCDs) and surface concentrations rather than on vertically resolved profiles.
Specific comments
- Please explain how the HONO DOAS fitting settings listed in Table 2 were selected. If they were adopted from previous studies, please provide the relevant references. If they were developed specifically for this study, please present sensitivity tests examining the effects of the fitting window, intensity offset correction, H2O and O4 cross-sections, Ring-effect treatment, polynomial order, and other relevant fitting parameters. Please also describe the components labelled “Ring 1” and “Ring 2” and provide the corresponding references. The study published in Atmos. Meas. Tech., 10, 3719–3742, 2017 recommended including Taylor expansion terms for NO2 absorption to account for the wavelength dependence of the air mass factor (AMF). This effect is particularly important under high NO2 loadings, which appear relevant to the measurement site considered here. If these terms were included in the fitting procedure, please state this explicitly and cite the relevant study. Otherwise, please demonstrate that their omission has a negligible effect on the HONO retrievals.
- In line with Referee 1’s comment, please include an appropriate reference for the PriAM algorithm in Section 2.2.3.
- Section2.3 should explain how the a priori profiles, a priori covariance matrix (Sa), vertical grid, and measurement-error covariance matrix (Se) were specified. If these settings were adopted from previous studies, please provide the relevant references and sufficient detail to make the retrieval configuration clear. Sensitivity tests reported in previous work, for example, Atmos. Meas. Tech., 13, 5087–5116, 2020, demonstrate that these choices can affect the balance between retrieval responsiveness and stability.
- Section2.4 describes the quality-control procedure, but it is unclear whether cloud-contaminated measurements were systematically identified, flagged, and excluded from the statistical analysis. Although precipitation data and an AOD threshold were used to exclude thick-cloud conditions, broken clouds and non-precipitating clouds may remain in the dataset. Please apply an appropriate cloud-classification method and assess the influence of cloud contamination on the HONO and aerosol retrievals. A suitable cloud-screening procedure should then be implemented to ensure the reliability of the statistical analysis.
- Section 2.3 states that AOD retrieved from MAX-DOAS was used to calculate photolysis rates. However, MAX-DOAS aerosol retrievals do not intrinsically distinguish aerosol extinction from cloud effects. Consequently, under cloudy conditions, the retrieved AOD may differ substantially from the actual aerosol optical depth. This limitation makes rigorous cloud screening essential. Are independent aerosol measurements available at the site, for example, from a sun photometer? In addition to cloud screening, a comparison with independent aerosol observations is needed to establish the reliability of the MAX-DOAS AOD used in the photolysis-rate calculations.
- The example HONO DOAS fit presented in Figure 6 shows a very weak HONO absorption signal. Please consider selecting a morning measurement with stronger HONO absorption to illustrate the fitting performance more clearly. The RMS residual is comparable to that reported in Atmos. Meas. Tech., 10, 3719–3742, 2017, where a distinct HONO absorption structure was evident for a HONO dSCD of 6.9 × 10¹⁵.
- The low DOFS shown in Figure 5 indicate limited independent information on the vertical profiles and suggest a strong influence of the a priori constraints. This raises concerns about whether the retrieved profiles can reliably characterise the vertical distributions of the species. To demonstrate their robustness, please perform synthetic sensitivity tests examining the retrieval response to different true profiles and the influence of alternative a priori profiles. Previous studies (Atmos. Meas. Tech., 13, 5087–5116, 2020; Atmos. Chem. Phys., 19, 5417–5449, 2019) showed that PriAM configured with “constant Sa” could achieve substantially higher DOFS, typically around 3.8, and reproduce different true HONO profiles using different a priori profiles. I therefore recommend either improving the profile retrievals or focusing the analysis on surface concentrations and VCDs. As currently presented, the profiles in Figure 7 do not provide sufficient evidence to support a robust interpretation of HONO vertical distributions.
- In Section 3.3.1, the authors state: “Elevated HONO VCD frequently coincided with lower temperatures and higher relative humidity, particularly when relative humidity was approximately 70 %–90 %, suggesting meteorological conditions favourable for HONO accumulation.” However, Figure 10 primarily shows the seasonal variations in HONO VCD, temperature, and relative humidity. Shared seasonal patterns alone do not establish a relationship between these variables. Please provide a direct statistical assessment of these associations or qualify the interpretation accordingly.
- The statement that “Elevated HONO VCD occurred more frequently under low-wind-speed conditions” does not appear to be consistently supported by Figure 11. For example, higher HONO VCDs are evident at wind speeds of 4–5 m/s in spring and summer. Please reconcile this statement with the patterns shown in the figure. I also recommend presenting analogous plots for surface HONO concentrations to help distinguish the influence of transport on near-surface HONO from its influence on the total column.
- The discussion of Figure 13 states that “In autumn and winter, HONO and NO2 generally increased together from morning to midday in the elevated layers”. However, the averaging kernels in Figure 5 indicate reduced retrieval sensitivity at higher altitudes, where the retrieved concentrations are more strongly constrained by the a priori profiles. The apparent diurnal variations in these layers may therefore reflect retrieval artefacts rather than atmospheric variability. As noted above, please improve the profile retrievals and quantify the altitude-dependent uncertainties before interpreting these features.
- The manuscript does not quantify the uncertainties in HONO, NO2, or AOD retrieved from MAX-DOAS. These uncertainties should be assessed and considered when interpreting the statistical results, including the relative diurnal variations discussed in relation to Figure 13. Variations smaller than the relevant measurement or retrieval uncertainties should be identified explicitly and interpreted with caution rather than used as the basis for firm conclusions.
- The fertilisation event shown in Figure 15 is a particularly interesting case that merits a more focused discussion of the potential contribution of fertilisation-related HONO sources. To support this hypothesis, please provide additional information on meteorological conditions, wind patterns, and, if available, measurements of other tracers associated with fertilisation, such as ammonia.
- Please mark the location of the measurement site on the maps in Figure 17.
Citation: https://doi.org/10.5194/egusphere-2026-4334-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 225 | 146 | 82 | 453 | 82 | 88 | 80 |
- HTML: 225
- PDF: 146
- XML: 82
- Total: 453
- Supplement: 82
- BibTeX: 88
- EndNote: 80
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Hu et al. present a two-year study of HONO in a rice–wheat agricultural region in China, using two MAX-DOAS systems. The study primarily focuses on the vertical profiles of HONO retrieved using the PRiAM profiling algorithm to investigate its temporal and vertical variability and identify potential sources. To support this analysis, the authors examine HONO alongside NO2, AOD, and meteorological parameters. They further assess the implications of the observed HONO levels for atmospheric oxidation by analyzing OH radical production rates and HONO photolysis frequencies, providing additional insights into the potential sources and temporal evolution of HONO.
As HONO is typically present at low atmospheric concentrations, its reliable retrieval using DOAS can be challenging. Therefore, studies that improve our understanding of HONO measurements are valuable. In particular, investigating HONO emissions from rice–wheat agricultural regions via remote sensing is of considerable interest, as these environments remain relatively understudied despite their growing importance for understanding regional atmospheric chemistry.
General Comments
The overall concept of this study is convincing, and the methods employed are generally appropriate for the objectives of the manuscript. However, the main limitation is the extremely low HONO concentrations observed during the measurement period at the study site, which substantially challenges the reliability of the retrieved HONO profiles. The authors state that "the adopted fitting settings adequately reproduced the major absorption features" (L250) and "the averaging-kernel patterns, DOFS values, and differences between the retrieved and a priori profiles support the reliability of the main retrieved lower-atmospheric feature" (L273). Unfortunately, I do not believe that the results presented adequately support either of these claims. In particular, the spectral fit shown in Fig. 4 raises serious concerns about the reliability of the HONO retrieval at the observed concentrations. In my experience, the fit is unusually poor compared with what I would typically expect from a published HONO retrieval study. Furthermore, degrees of freedom (DOF) values below 2 generally indicate limited vertical information content, although the implications depend on the specific retrieval setup. In this case, the low DOFS values raise substantial concerns about the extent to which the retrieved vertical profiles are constrained by the measurements rather than by the a priori assumptions. Consequently, the interpretation of the retrieved vertical structure is not sufficiently substantiated.
For these reasons, I cannot recommend final publication of the manuscript in its current form. Nevertheless, I would encourage the authors to consider submitting a substantially restructured manuscript that focuses on the HONO vertical column densities (VCDs) rather than the retrieved vertical profiles.
I believe that the underlying dataset may still be of considerable scientific value, provided that the limitations of the spectral fitting and the information content of the measurements are clearly acknowledged and the conclusions are revised accordingly.
Specific Comments to preprint version
Abstract and throughout the manuscript: Please do not abbreviate day as d
L96: Please add information on the spectrometers and CCDs
L102: Please provide a reference for PRiAM
Table 2: Why was H2O only included in one fit? Why did you use a different approach for the Intensity Offset?
L147-148: Please add cities and countries for the institutes
2.2.3 Please add a priori information. Which profile shapes, aerosol parameterizations, etc., were used?
L175: What does this mean? "The altitude range constrained by the measurements was further evaluated by comparing the retrieved and a priori profiles"
L191: Please explain your interpolation and smoothing approach.
Fig. 4: There are almost no visible spectral features of HONO! Also, the optical depth of HONO is of the same order of magnitude as the RMS, indicating that you are close to or even below the detection limit.
Fig. 5: Please add information regarding the a priori profiles (see comment above). It would be better if the profiles matched the fitting example above. If not, please explain why you chose to show these profiles.
Fig. 6: Why is there a larger deviation for higher HONO values?
Was any cloud filtering applied? Since you discuss results with and without clouds below, it would be good to add details about your cloud treatment somewhere in the manuscript.
Fig. 7: HONO shows a clear diurnal variation. What is the point of averaging over a full month? The temporal evolution would be better discussed using a time series.
Fig. 8: This looks to me as if it were just a scaling of exponential profiles, probably due to the limited information content!
Furthermore, there seems to be a small background concentration close to the surface in all seasons. This again indicates that there is too much forcing towards the a priori profile.
L319–320: "Relatively elevated" sounds as if there were a HONO distribution at higher altitudes, detached from the surface. I cannot see such a layer. If there is one, perhaps changing the colour scale would improve this figure.
L339: "Elevated HONO": Please rephrase this so that it is clear that you mean higher concentrations and not a detached concentration layer.
Here, I was also wondering how many data points went into these plots. As there are typically fewer measurements in winter due to less sunlight, and the measurements tend to be more strongly affected by clouds and aerosols,
this might impact the data quality, statistics, and discussion.
Fig. 10: The impact of humidity is not as clear to me from this figure as it is discussed in this section. If there is a relevant correlation, please add supporting information.
Section 3.3.1: Since you discuss the PBL, it would be worth adding some information on the PBL height, either in numbers or by adding another (sub)figure.
Fig. 11: Please add information about the typical prevailing wind directions for each season, if there are any.
What is the minimum number of data points per wind direction? Why are higher HONO concentrations observed at higher wind speeds in the NE direction in spring?
Fig. 13: In my view, a time series of VCDs (or near-surface concentrations) would already explain most of the findings. Since I have little confidence in the full profiles, I would recommend simplifying this plot.
Fig. 14: Why is there a period between September and October during which the NO2 background values seem to be much lower than during the other periods?
Fig. 15: Please discuss your findings using VCDs only.
Fig. 16: I do not really see the point of discussing post-to-pre-fertilization ratios. The temporal development of HONO can already be seen in time series or even in the contour plots.
Either shorten this section or highlight the benefits of this analysis and how it supports your findings.
L503: "Plotting resolution": If the resolution is insufficient, please increase it.
Section 3.4: This is the first section in which you discuss findings with and without clouds. However, there is no information about your cloud treatment.
Furthermore, since clouds strongly affect radiative transfer, all profiling results that have not been cloud-filtered have extremely high uncertainties.
Please add a section discussing the impact of clouds, including information on how clouds are treated throughout the manuscript.