General comments:

This study presents an optimized ammonia emission inventory with a focus on fertilizer application relevant sources over mainland China in 2016. The approach adopted in this study considers the time variability of fertilizer application and thus enables this inventory accurate capture the ammonia emissions over time. To illustrate the capability of this emission inventory against existing emission inventory (MEIC developed by Tsinghua University), the authors evaluate the emission inventory by comparing modelled NH3 concentrations by WRF-Chem with ammonia inventory in this work and MEIC. Further, the model performance is validated with NH3 observations from AMoN-China and IASI NH3 column. This paper is well-written and presents results that would be interesting to the air quality modeling community. However, I have several concerns that the authors should consider when revising the manuscript, as listed below. I recommend this work to be published after the following comments are adequately addressed.

Major comments:

1) While the authors illustrated that large uncertainties remain in the NH3 emissions from agriculture sector thus motivates the improvement in NH3 emission inventory, I think vehicular NH3 emissions co-emitted with nitrogen oxides (NOx) are also important sources for NH3. This additional NH3 source has been always underestimated in emission inventory (such as MEIC). Although this work is focused on agriculture ammonia emissions, I would suggest the authors briefly discuss this in the Introduction because this work is entitled “ammonia emission inventory in China”.

2) In Sec. 2.4, the uncertainty assessment of NH3 emission inventory established in this work based on the Monte Carlo method is missing. Please provide the uncertainties for each sub-sector if possible.

3) The authors employed the WRF-Chem model for performing numerical simulations over eastern China and compared with model outputs driven by MEIC and IASI satellites. However, I could not find the detailed configurations of WRF-Chem simulations (initial and boundary conditions for WRF-Chem, physical parameterizations, chemical mechanisms, etc). I suggest the authors add a section in the supplementary providing the configurations of WRF-Chem.  

4) In the WRF-Chem simulations with NH3 emission inventory developed in this work and MEIC, I wonder whether other air pollutants emissions (such as SO2, NOx, VOCs) are both based upon MEIC. If it is, please clarify it. 

5) For readers not familiar with the China geography and the location of each province, providing a map with province names marked would be valuable. 

6) In Sec. 3.4, it seems the WRF-Chem model with optimized NH3 emission inventory yields better performance in July while large bias still existed in January as compared with the MEIC driven WRF-Chem model. Could you be more specific on the reasons for the discrepancy between the simulated NH3 concentrations and ambient measurements?

7) I think Figure 6(a) is misleading with the erroneous higher concentrations of NH3 in January than July due to the different scales in colormap. Please use the same color scale for both January and July. Otherwise, it is misleading to the readers.

8) Conclusion: This section would be stronger if a discussion on future work that the community should consider toward improving the representation of ammonia emissions in chemical transport models (For example, investigating the impacts of optimized ammonia emission inventory on simulating PM2.5 concentrations and aerosol pH changes). 

Specific comments:

1) Line 17: “differences”-->“variability”. “observed” -->“identified”.

1) Line 84: “huge”-->“considerable”.

2) Line 215: “30 samples” should be “30 sampling stations”.

3) Line 223: “represented”-->“contributed”.

4) Line 223: The 13.1 Tg is derived from Kong et al., (2019). Thus, there is a lack of citation for Kong et al., (2019).

5) Table 2: I suggest the authors include the ammonia emissions quantified in this work in Table 2 as well.

6) Line 277: Please further clarify the reason why located in north China is relevant to elevated ammonia emissions.

7) Line 313: Please rephrase this sentence for read.

8) Line 324: Delete “it”.

9) Line 349: Delete “Thus”.

10) Line 359: The spatial resolution for MEIC is 0.25°×0.25°.

11) Figure S2: Small typo for the caption of subplot for Others.

12) Some figures are hard to read due to the small font size. Please consider improving them. 13) The manuscript is hard to follow in several places but that could be addressed with thorough language editing.

This manuscript presented a new bottom-up estimate of ammonia emissions in China that emphasized some improvements on the estimates of fertilizer application-induced ammonia volatilization. The authors compared their Chinese ammonia emission inventory with a commonly-used emission estimate (MEIC), and evaluated the resulting model simulations using the WRF-Chem air quality model with surface and satellite ammonia measurements.      

The manuscript is in general well conducted and organized, and it meets the scope of ACP. I have some comments below that suggest the authors better clarify the improvements of this ammonia emission inventory relative to previous studies. I think these comments should be addressed before considering publish.

Specific comments:

1) Page 5:

It appears that most of the methods and datasets applied in this NH3 emission inventory have followed previous studies (Huang et al., 2012; Kang et al., 2016; Zhang et al., 2018), except here as described in this section the improvements on fertilizer application-related NH₃ emission estimate. It is important to better identify the fertilizer application timing and proportion throughout the planting growing season. The key improvement in this ammonia emission inventory is depicted in Figure 1 that accounts for the spatial differences in fertilizer application timings for the three main crops (maize, wheat, and rice). From Figure 1, it is not clear what these fertilizer application dates are for each crop and each region. I think this information shall be presented, e.g., as a table in the manuscript or in the Supplement.

In addition, how the fertilizer amounts are distributed in each month according to the application timing needs some description in the text. Did you use a uniform fertilizer application rate for each crop in each month and region?

2) Page 9, Section 2.4:

The results of Monte Carlo calculation were not presented in the manuscript. What are the uncertainties and probability distributions of ammonia emissions from fertilizer and livestock?

3) Page 17, Section 3.4:

The study pointed out that Chinese ammonia emissions are high in summer and low in winter, and confirmed the results using WRF-Chem model simulations and ammonia measurements. This result did not seem to be much improved compared with previous estimates that all suggested higher ammonia emissions in summer than winter. I think that only analyzing the two months (January and July) could not provide sufficient information on the improvements of the ammonia emission inventory. The new fertilizer-induced ammonia emissions (Figure 5) also high values in April and October? Can you also evaluate improvements in these two months? This will provide valuable information to understand ammonia in spring and fall seasons.

4) Page 18, Fig. 6:

The color scale for the left panel of Fig 6 could be misleading. It shows that at many sites, the model results are too high over the North China plain, however, the model results (contours) and measurements (dots) have different color scales. Suggest put them on the same color scale.

5) Page10, Line 220:

Here “9.29-15.54 Gg” should be in unit of “Tg”.

6) Page 11, Line 245:

“replacing complex fertilizers with ABC might reduce fertilizer type-related NH3 emissions”. Should here be “replacing ABC with complex fertilizer”?

7) Page 17, Line 337:

“R2=0.85” should be 0.84 as shown from Table 4. Was the R2 value calculated by integrating January and July measurements?