The study of Zhang et al. built a ship emission inventory based on AIS signals and basic ship-related data for YRD region. The results were compared with several other emission inventories to verify reliability of the proposed approach and demonstrate the impact inland ship emissions on densely populated regions along the river.

The topic is interesting, but the publication has major deficiencies, is difficult to follow. In my view it cannot be published in the present form but will need substantial improvements.

Major issues:

1. The title is completely inappropriate. It’s named ‘The impact of inland ship emissions on air quality’. The title exaggerates the actual research scope seriously. The manuscript is almost entirely about emissions (Introduction, Methods and Discussion) rather than air quality. And the only description of air quality in Line 424-425 seems not correct, because the air pollution such as PM2.5 induced by ship emission also affected by other emission sources and meteorological condition. Therefore, the title should be modified at least. Otherwise, more quantitative study on air quality should be carried out. This study only investigated the Yangzi river- Jiangsu section, which is a short section, also can NOT represent YRD region as well. The title needs to honestly reflect the scope of the work done. The title is too big at present.
2. There are many mistakes in the review of the research status and the current situation of ship emission control. The authors need to re-write the literature review to include the most advanced and high quality research.
3. The control policy of inland ships is different from that of oceangoing ships. Even for oceangoing ships, China's current control policy is not only in the three major control areas, but for the coastal waters of 12 nautical miles. The author has not quoted the latest research on inland river-related ship emissions.
4. This study assumed that the average sulfur content of marine distillates (MD) of 1.5 % in the study area according to Xu et al. (2019), which was based on the data of 2014. However, this study was implemented from 2018 to 2019. The sulfur content for inland ships was totally different, which makes the results wrong.
5. The author mentioned "We set up an AIS receiver in Nanjing University of Information Science and Technology (NUIST) to collect ship information including ship position, speed and heading, ship name, ship length and ship type”. What’s the power threshold for this receiver can receive from ships? What is the percentage of missing signals? With the distance getting farther, there are more missing signals. How to evaluate the emissions of lost ships? More details of the AIS receiver should be presented.
6. Ship emission is not only related to the length of the river. For example, in some major ports, the emission is large while the length may not high. The author should present more proofs to prove the proposed method that estimates the emission to larger regions based on the length of the river per grid cell. Such as plotting the relationship of emission and the length for the observational area or listing some references.
7. It was not clear that how the 3-4% underestimation in Line 327 be calculated, and except for the emission from the auxiliary engine, the emission from the boiler was also not considered. How much did the boiler emission account for?
8. Almost any sources of emission will account for a large share ratio in the grid where they emit. To illustrate the importance of emissions for inland vessels, the author should compare emissions in specific cities or regions (coastal cities, YRD …) rather than the selected river grid cells.
9. The author should explain why there was a big difference between JESI and SEIM, especially for SO2 and PM2.5 in Section 4.1. Additionally, in the latest version of SEIM (Wang et al., 2021. Ship emissions around China under gradually promoted control policies from 2016 to 2019), the emissions over inland rivers are all included.

Smaller issues:

1. Line 183, it should be Figure S1 rather than S2.
2. In Figure 9 and Figure 11, why the sharing ratio of JESI is low in some grid of which most part is river?
3. Line 379, why the author said “The MEIC grid cells cover a larger area than for DECSO and include more emissions than DECSO”.
4. In Line 402, the author said “ship emissions accounted for more than 40 % of the total emissions in February, which shows that the pollution caused by ship activities during Spring Festival is quite significant in a time period of lower emissions in general”. However, according to MEIC, the changes in NOx emission caused by the Spring Festival is not significant. Thus, which share is more reasonable, JESI/(JESI+SEMI) or JESI/DECSO. The author should give more proofs to support this sentence.
5. The format of references should be unified. (Such as in Line 237)

In this paper, the authors present ship emission calculations for the Yangtze River area by combining AIS ship location and other data from the China Classification Society (CCS). By using simple parameterizations based on the specific ship characteristics (type, engine, length, speed, etc.) they compile a monthly inventory for a section of the Yangtze River and compare it with an annual inventory (SEIM) and two monthly inventories (DECSO and MEIC). I enjoyed reading the paper, as it is a significant contribution towards understanding the significance of inland ship emissions on the NOx, SO2 and PM levels in the broader areas around rivers. The methodology has also potential of being used in other areas (e.g. central Europe) with significant ship traffic across densely populated areas. To this end the paper deserves to be published but has to be improved before. See my comments below:

• The title is not appropriate. The authors should use a title describing the exact focus of their study, e.g. “Inland ship emissions across the Yangtze River area in China” or “Inland ship emissions: the case of Yangtze River area in China”, etc.
• The paper should be revised as the flow is not very pleasant and there are several minor expression mistakes. There are many very small phrases. E.g. “…Note the limitations of our method. Some ships would reduce their speed when going downstream…” which does not help the reader at all.
• The authors should consider discussing new developments in ship emission detection and attribution in their introduction. For example, it has been shown that today the AIS data combined with satellite data can give as an indication of the pollution produced by individual ships (see Georgoulias et al., 2020; Riess et al., 2022; Kurchaba et al., 2022). The use of ground-based DOAS methods to infer individual ship emissions from rivers and channels is also possible as shown in Krause et al. (2021).
• When discussing the method and specifically the engine power (EP) calculations, it is not very clear how the authors calculate the related regression between EP and L^2 x v^3. Probably the authors used EP data from CCS to do the regression; however, in the beginning of the paragraph they write “Since the engine power is missing in the AIS data, we develop a method to relate the engine power to the ship type, length and speed. Those parameters are available in the AIS data.” Are EP data available only for a fraction of the ships? Please refine this!
• On top of my previous comment, the authors might include in Table 3 apart from the slope the whole equation (slope + intercept + the corresponding uncertainties). The uncertainties induced by the use of this equations might be incorporated into the discussion in paragraph 3.4 where the authors discuss only two sources of uncertainty.
• The uncertainty discussion is very limited. Please discuss more aspects or integrate this in another paragraph
• The title of section 4 should be changed. E.g. “Importance of inland ship emissions” might be “Contribution of inland ship emissions relative to emissions from other sources”.
• Please comment on the significant difference between JSEI and SEIM compared to the other two inventories.
• In Fig 12, the authors might add in the legend the resolution of the models to make clear that the difference in the bars is mostly related to the different resolutions.