the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Aug 2026
Decadal transition of PM2.5 source structure over North China: Weakening combustion and emerging nitrate-driven pollution
Abstract. Long-term clean-air actions have reshaped fine-particle pollution in North China, but it remains unclear how the source structure of PM2.5 has changed at the regional background scale and which chemical components are amplified. We combine long-term observations from Tuoji Island (2012–2023), a coastal background site, with chemical analysis, positive matrix factorization (PMF), ion balance, and thermodynamic diagnostics. Although PM2.5 mass did not decrease monotonically, its composition and source structure changed substantially. Elemental carbon decreased from 1.46 to 0.81 μg m-3, and the biomass/coal combustion contribution decreased from 43.37 % to 13.74 %, indicating weakened primary-combustion signals. SO42- decreased from 9.86 to 4.10 μg m-3, while the NO3-/SO42- ratio increased from 0.64 to 2.08, showing a shift from sulfate dominance toward greater nitrate importance. After mineral-dust outliers were removed, NO3- accounted for 19.19 % of PM2.5 in the highest 20 % of samples during 2021–2023, and the PMF secondary nitrate contribution reached 27.59 %, making it the dominant amplified source factor. Greater NH4+ excess and model-diagnosed aerosol liquid water content indicate increased NH4+ availability after sulfate neutralisation and wetter particles, favouring NH4NO3 formation and persistence. These findings show that emission reductions reorganised PM2.5 pollution toward weaker primary combustion and sulfate influence but stronger nitrate amplification during late-stage high-PM2.5 conditions.
- Preprint
(1165 KB) - Metadata XML
-
Supplement
(1363 KB) - BibTeX
- EndNote
Status: open (until 17 Sep 2026)
- RC1: 'Comment on egusphere-2026-3299', Anonymous Referee #1, 13 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-3299', Anonymous Referee #3, 13 Aug 2026
reply
The manuscript addresses a relevant ACP question: how the composition and source-related characteristics of PM2.5 evolved at a regional-background receptor during China's clean-air-control period. The study contributes a long-term dataset from a location where continental transport, coastal processing, sea salt, and shipping-related influence can overlap. The combination of measured composition and PMF provides a convincing description of weakening combustion-related and sulfate signals and increasing nitrate importance.
My main recommendation is to keep the interpretation proportional to the observational design. The manuscript provides strong evidence for a stage-dependent compositional transition at Tuoji Island. It provides supporting thermodynamic context, but it does not independently quantify the effects of emission policy, meteorology, transport, or gas-phase precursor availability. The revised manuscript should distinguish these evidence levels clearly and moderate the corresponding causal claims. I recommend the paper for publication in ACP after the following specific comments are addressed:
Specific comments
1. Define the principal novelty more directly. The sulfate-to-nitrate transition has been reported in Beijing, Tianjin, and other North China studies cited by the authors. Sect. 3.3 already identifies the added value of the long-term coastal-background record; please make the same point explicitly in the Introduction so that the new contribution is clear before the results are presented.
2. Moderate causal language concerning emission controls. The decline in EC, the combustion-related PMF factor, and sulfate is consistent with the timing and known direction of emission controls. However, the study does not perform a formal policy-attribution or meteorological-normalisation analysis. The current discussion is mostly careful, but the Abstract statement that emission reductions "reorganised" PM2.5 pollution is stronger than the direct evidence. Wording such as "during the emission-control period" or "consistent with the emission-reduction trajectory" would be more balanced.
3. Separate supporting thermodynamic diagnostics from independent lines of evidence. The increase in NO₃⁻ concentration and fraction and the nitrate-rich PMF factor are two related but useful forms of evidence. By contrast, NH₄⁺ excess, reverse-mode ALWC, and model-resolved NH₄NO₃ all depend on the observed particle-phase ion composition; they strengthen the thermodynamic consistency of the interpretation but are not fully independent confirmation. Please state this relationship once in Sect. 3.3. In the same revision, replace "greater NH₄⁺ availability" with a precise particle-phase stoichiometric description (positive NH₄⁺ excess does not directly measure gaseous NH₃ availability), and describe ALWC as a model diagnostic associated with the late-stage aerosol state rather than as evidence that wetter particles caused the nitrate increase.
4. Acknowledge the possible influence of meteorology and transport. Add a concise statement in Sect. 3.3 explaining that changes in meteorology and regional transport may also contribute to the stage differences; where readily available, a brief supporting discussion of observed wind conditions would strengthen this context.
5. Distinguish late-stage average conditions from high-PM2.5 amplification. In 2021–2023, the port-related factor is the largest stage-average PMF factor, whereas secondary nitrate becomes the largest amplified factor in the high-PM2.5 subset. The title, abstract, and conclusion should preserve this distinction. "Increasing nitrate importance" or "nitrate amplification during high-PM2.5 conditions" is more precise than a general statement that all late-stage pollution is nitrate-driven.
6. Improve the balance between results and limitations. The conclusion repeats many values already presented in Sect. 3.1–3.3. Condense these repeated results and add a concise paragraph stating the main interpretation limits: non-continuous stage coverage (including COVID-affected sampling in 2020 and differing seasonal coverage among stages), the smaller PMF subset, and the diagnostic nature of the reverse-mode thermodynamic results.Citation: https://doi.org/10.5194/egusphere-2026-3299-RC2 -
RC3: 'Comment on egusphere-2026-3299', Anonymous Referee #4, 13 Aug 2026
reply
This manuscript presents a 12-year (2012–2023) record of PM2.5 chemical composition and source apportionment at Tuoji Island, a coastal background site in the Bohai Strait. The integration of filter-based chemical characterization, positive matrix factorization (PMF), ion-balance diagnostics, and ISORROPIA-II thermodynamic modeling is well aligned with the scope of ACP. The long observation period at a regional-background setting, where continental transport, coastal processing, sea salt, and shipping influence can overlap, distinguishes the study from urban or episode-based source-apportionment work. However, the scientific contribution of the manuscript is constrained by a fundamental problem of novelty differentiation. The core conclusion that a transition from combustion- and sulfate-dominated pollution toward nitrate-driven secondary pollution under China's clean-air actions, has been extensively reported in urban North China (Zhai et al., 2021, Nat. Geosci.; Dai et al., 2023; Wang et al., 2022; Chu et al., 2023). The manuscript does not sufficiently articulate or demonstrate what the background-site perspective adds beyond confirming urban trends at a different location. In addition, several issues needed to be clarified. I recommend the paper for publication in ACP after substantial revision addressing the following specific comments.
Specific comments
- The sulfate-to-nitrate transition has been documented in Beijing (Wang et al., 2022), Tianjin (Dai et al., 2023), and the BTH "2+26" cities (Chu et al., 2023), and mechanistically explained by Zhai et al. (2021). Sect. 3.3 identifies the added value of the long-term coastal-background record, but the Introduction does not state explicitly what remains unknown after these studies and how the Tuoji Island record fills those gaps. Please articulate the specific quantitative insight that the background site provides, for example, the relative magnitude of regional transport versus local formation, or the background-level compositional response to emission controls, and state this contribution in the Introduction before the results are presented.
- The 2014–2016 gap means the record is not continuous, and the three stages (2012–2013, 2017–2020, 2021–2023) appear to be defined by data availability rather than by scientific or policy milestones. The 2012–2013 stage contains only ~18 months of data. Please discuss how this discontinuity affects trend interpretation, whether the stage boundaries correspond to identifiable inflection points in emission inventories, and acknowledge the non-continuous coverage, including COVID-affected sampling in 2020 and differing seasonal coverage among stages, as a limitation in the Conclusion.
- The stage-level composition analysis uses 672 valid samples, whereas PMF uses 313 samples with complete common inputs, and the PMF comparison periods (2012; 2019–2020; 2021–2023) are narrower and of unequal length relative to the composition stages. Sect. 2.5 explains why the two datasets differ, but the main component changes in the PMF subset should be confirmed as consistent with those in the full dataset. Please state the three PMF-period sample counts explicitly, comment on the comparability of the unequal-length comparison periods (a single early year versus multi-year later periods), and note this mismatch as a limitation where quantitative factor comparisons are discussed.
- The manuscript already states that particle-phase stoichiometry does not represent gas-phase NH3 and that pH is a model diagnostic, this is appropriate. However, NH4+ excess, reverse-mode ALWC, and model-resolved NH4NO3 all depend on the observed particle-phase ion composition; they support the internal consistency of a nitrate-rich aerosol state but are not fully independent confirmation of formation rates, gas–particle partitioning, or persistence. Without synchronous gas-phase NH3, HNO3, or HCl measurements, the gas–particle partitioning claims for NH4NO3 cannot be directly verified. Please state this relationship explicitly in Sect. 3.3, describe ALWC as a model diagnostic associated with the late-stage aerosol state rather than as evidence that wetter particles caused the nitrate increase, and replace "greater NH4+ availability" with a precise particle-phase stoichiometric description.
- Stage-mean PM2.5 did not decrease monotonically (51.75 → 47.31 → 54.41 μg/m3), with the 2021–2023 mean actually exceeding the 2012–2013 baseline. The manuscript attributes this to compositional changes, but meteorological variability between stages is not systematically addressed. Changes in synoptic circulation, boundary-layer height, and regional transport patterns may also contribute to the stage differences. Please add a concise discussion of meteorological influence in Sect. 3.3, to strengthen the attribution of late-stage PM2.5 levels to emission-driven compositional shifts.
- The manuscript repeatedly invokes "regional transport" as a key process at this coastal background site, yet no backward trajectory analysis (HYSPLIT, FLEXPART), PSCF/CWT analysis, or concentration-weighted trajectory method is presented. Without this, the transport narrative remains unsubstantiated. At minimum, a seasonal trajectory climatology distinguishing continental outflow from marine/clean air masses would help contextualize the stage-level compositional changes and strengthen the background-site framing that is central to the manuscript's novelty claim.
- The port-emission factor accounts for 26–27% of PM2.5 in the later stages, an unusually high contribution for a background site. The factor is identified by Na+, Fe, Ni, and Cu, but these species can also originate from sea salt, crustal material, or other industrial sources. While the manuscript acknowledges that this factor combines port activity, shipping, coastal aerosol, and regional metal inputs, the mixed character should be stated consistently in the Methods, Results, and figure captions. Please provide more rigorous factor validation, including V/Ni ratios, enrichment-factor analysis, or comparison with shipping emission inventories for the Bohai Sea region, and interpret the quantitative contribution cautiously given the factor's mixed nature.
- Only OC and EC are measured; no organic molecular markers, SOA tracers, or online aerosol mass spectrometry data are available. Given that the "secondary nitrate" PMF factor may co-vary with secondary organic aerosol, and that organic nitrate formation could contribute to the observed nitrate enhancement, the absence of organic speciation represents a significant interpretive gap. Please add a concise statement in Sect. 3.3 or the Conclusion discussing this limitation and its potential implications for the source-apportionment results.
- The conclusion that "coordinated control of NOx, NH3, and VOCs" is needed has been stated in numerous prior studies. The study does not perform a formal policy-attribution analysis, so the policy recommendation should be proportional to the evidence. Please leverage the specific background-site perspective to provide more targeted recommendations, for example, quantifying the relative importance of regional transport versus local formation, or identifying which precursor controls would yield the greatest benefit under the observed late-stage thermodynamic conditions, rather than restating generic emission-control conclusions.
Technical corrections
1.Revise the title to reflect the single-site design. "Over North China" may be read as a multi-site regional analysis. Replace it with wording such as "at a coastal background site in North China" to match the spatial scope of the evidence.
2.Use "enrichment factor" or "enrichment ratio" consistently (both currently appear in Sect. 3.1 and Sect. 3.2) and state the calculation once.
3.Define all error bars and statistical summaries directly in the relevant figure captions.
Citation: https://doi.org/10.5194/egusphere-2026-3299-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 39 | 11 | 3 | 53 | 6 | 5 | 2 |
- HTML: 39
- PDF: 11
- XML: 3
- Total: 53
- Supplement: 6
- BibTeX: 5
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a valuable long-term record of PM2.5 composition and source-related changes at Tuoji Island, a coastal background site in the Bohai Strait. The integration of chemical composition, positive matrix factorization (PMF), high-PM2.5 sample analysis, ion balance, and ISORROPIA-II diagnostics is well aligned with the scope of ACP. The long observation period and regional-background setting distinguish the study from many urban or episode-based source-apportionment studies.
The principal conclusions are generally supported by the measurements. EC and sulfate decreased across the observed stages, whereas nitrate became relatively more important. The high-PM2.5 analysis further shows that nitrate was preferentially amplified in the later stage. The revised PMF treatment is substantially more transparent than a conventional factor interpretation alone: the manuscript reports factor-number testing, an unrotated Base solution, Fpeak sensitivity, Bootstrap mapping, DISP results, species-level residuals, and observed–predicted relationships. The Supplement also makes clear that the measurements included major water-soluble ions, OC/EC, and acid-digested trace elements.
I consider the paper suitable for ACP after revision. The remaining issues mainly concern the comparability of the PMF subset, uncertainty in the high-PM2.5 PMF results, the distinction between observational evidence and mechanistic interpretation, and the source-profile constancy assumption underlying the pooled decadal PMF. The revised manuscript should address these points explicitly.
Specific comments
Technical corrections