Unexpected response of nitrogen deposition to nitrogen oxide controls and implications for land carbon sink

Terrestrial ecosystems in China receive the world’s largest amount of reactive nitrogen (N) deposition. Recent controls on nitrogen oxides (NOx = NO + NO2) emissions in China to tackle air pollution are expected to decrease N deposition, yet the observed N deposition fluxes remain almost stagnant. Here we show that the effectiveness of NOx emission controls for reducing oxidized N (NOy = NOx + its oxidation products) deposition is unforeseen in Eastern China, with one-unit reduction in NOx emission leading to only 55‒76% reductions in NOy-N deposition, as opposed to the high effectiveness (around 100%) in both Southern China and the United States. Using an atmospheric chemical transport model, we demonstrate that this unexpected weakened response of N deposition is attributable to the enhanced atmospheric oxidizing capacity by NOx emissions reductions. The decline in N deposition could bear a penalty on terrestrial carbon sinks and should be taken into account when developing pathways for China’s carbon neutrality.

I still have some major concerns listed as follow.
The main innovative results of the manuscript are a little bit outdated. Some recent studies have reported the lag response of N deposition to N emission reduction both in some regions and whole China. For example, Fu et al., 2020, Zhong et al., 2020, Wen et al., 2020, Xi et al., 2021. Therefore, I would hesitate to recommend this manuscript to be accepted in Nature Communication. Xi et al., 2021. Hysteresis response of wet nitrate deposition to emission reduction in Chinese terrestrial ecosystems. Atmos. Environ. https://authors.elsevier.com/sd/article/S1352-2310(21)00377-0. I suggest the authors to rethink the significance of this study and move forward with new simulation scenarios, i.e. changing meteorological condition with fixed emission, changing emission with fixed meteorological condition. The authors mentioned that this study only considered the effects of anthropogenic emissions on reactive N deposition, while the effect of climatological conditions, such as variability of precipitation, did not be considered. We already knew that the decreased of NOx emission driven by the strict environment policy in China did not decreased wet nitrate deposition or bulk N deposition as expected. An increase in O3 concentrations during 2013-2017 in China has also been reported. The enhancement of the atmospheric oxidation capacity accelerates the conversion of NOX to HNO3. In addition, the reduction of SO2 will reduce the consumption of OH converted to sulfate, which may indirectly increase the oxidation capacity of the atmosphere. High level of NH3 emissions in China provide sufficient precursors for the further conversion of HNO3 to nitrate. The combined pollution is more complicated than we thought. We are eager to know what the roles of climatological conditions play in N deposition under different NOx emission reduction scenarios? What is the threshold value of emission reduction level that the air quality will stay high even with bad climatological conditions?
1. The underlying mechanisms examined here behind the varying responses of nitrogen deposition on NOx reduction in Eastern vs. Southern China have indeed been well researched and documented in previous work, as the authors have also cited -namely, the enhanced oxidant levels due to reduced NOx in Eastern China in the VOC-limited regime (i.e., reduced "titration" effect). I understand that previous works focused on the enhanced ozone levels following NOx reductions, and this paper extended this further to examine the subsequent impacts on HNO3 formation and deposition, but this new result seems incremental compared to existing literature. To be publishable in Nature Communications, more justification is needed to explain how the new result regarding nitrogen deposition, derived from known mechanisms, would substantially revise previous understanding and have new implications on policy making. One possible way, among others, to do so would be to articulate how ecosystem impact assessment done before would be affected by the new result would, and how such assessment should be done in the future. These have been superficially mentioned, but a lot more references and discussion of terrestrial ecosystem-atmosphere nitrogen interactions are warranted (see below also).
2. One novel aspect and important implication of this paper is the potential impacts of future emission control strategies on terrestrial ecosystems (e.g., vegetation productivity, soil biogeochemistry, soil acidification, etc.). Thus far, these are mentioned only peripherally, but it should stand at the center of the main motivation/impact of the paper. At least a paragraph or two in the introduction and discussion are warranted to discuss how the varying responses of nitrogen deposition in Eastern vs. Southern China may affect different ecosystem processes (e.g., primary production, carbon uptake, soil nitrification and denitrification, soil pH, etc.), and suggest possible emission reduction strategies that can maximize benefits for both human health and ecosystem health.
3. As nitrogen deposition influences ecosystems tremendously, while vegetation and soil in ecosystems play major roles regulating surface air quality and atmospheric composition, there can be various feedback mechanisms whereby NOx reductions will lead to changes in nitrogen deposition, thus vegetation and soil, that ultimately affect air quality itself. Such feedback mechanisms were suggested by Zhao et al. (Zhao, Y. H., Zhang, L., Tai, A. P. K., Chen, Y. F., and Pan, Y. P.: Responses of surface ozone air quality to anthropogenic nitrogen deposition in the Northern Hemisphere, Atmos. Chem. Phys., 17, 9781-9796, doi:10.5194/acp-17-9781-2017Phys., 17, 9781-9796, doi:10.5194/acp-17-9781- , 2017, and can be potentially important. However, the current modeling framework in this paper would not allow such feedbacks to be examined because the land cover is mostly prescribed. At least the implications of these feedback mechanisms on the main results of this paper should be discussed. 4. The authors used one model to derive the results. However, multi-model differences are bound to occur, which may or may not annihilate the key conclusions of this paper. It is understandable that the results regarding nitrogen deposition per se are novel and thus hard to compare with previous results since there were none, but the authors should cross-compare their simulated results in other variables that would lead directly to the results regarding nitrogen deposition, e.g., ozone concentrations, HOx concentrations, etc., with previous works (many of which were cited, but not compared). This would give greater credence to the validity of the new results.
This study analyses the mechanism of a non-linear response of reactive nitrogen deposition (in particular in Eastern China) to NOx emission reductions. While the issue of non-linear responses is well known in the atmospheric chemistry community, the interesting aspect is that emission changes are so large in a relatively short time (e.g. a downward emission trend of 20 % or more), making these feedbacks more visible than would be the case under gradually changing conditions. The manuscript is interesting, but there are several loose ends that require to be corroborated, additional testing and analysis.
Specifically: 1) Given spatial and temporal variability, the observational evidence of using only 4 deposition stations over Eastern China for 5 years is very limited. If more data is available also for earlier years, they need to be analysed as well to better understand the trends and variability from 2010-2015. One possible model simulation to demonstrate variability is keeping emissions constant, and analyse the model responses.
2) There is an interesting discrepancy between NO2 column observations and the anthropogenic emissions trends (e.g. Figure 1). This needs to be explored further along with an calculation on how representative NO2 is for NOy dry deposition, and how it is changing. This is relevant for emerging studies that use satellite observations to compute deposition of NOy (and NHx).
3) Why hasn't NH3 satellite column observations not been used for analysis (given the probably important role in buffering some of the excess nitrate produced. 4) The role of atmospheric transport in feedback on deposition between regions has not been discussed. Perturbations similar to MICS-Asia, HTAP etc, where only emissions over East China, South China and the rest of China, would be helpful to explore transport feedbacks 5) Appropriate budget analysis possibly combined with point 4, would be helpful to understand the closure of emissions, chemistry, transport and removal. An analysis of overall NOy lifetime (for China, and its subregions) would be helpful as well.
6) In view of point 5, an appropriate analysis of the chemical budget terms (in particular all the loss reactions NO2+OH; NO2+O3; N2O5 heterogenous) would be useful. The papers main hypothesis is that due to a shift of NO to NO2 (NOx component), and also more OH and other oxidants, the loss of NOy is speeding up. This can only work if subsequently the HNO3 is not forming aerosol nitrate. The paper is ignoring an analysis of this and the analysis of the nitrate (HNO3, and NH4NO3 and other nitrates) must be shown. 7) At several places in the publication the word effectiveness needs to be replaced, possibly by relative response or similar. I do think the work can be of interest to Nature if my major comments above, and detailed comments below are duly taking into account.
Detailed comments l. 34 clarify largest in budget terms-or in terms of deposition fluxes? l. 37 It is not clear what is meant with 'integrating a model with observation'-I think the authors mostly compare (with the exception of dry dep, that is using modelled deposition velocities. Please clarify.
l. 39 what was expected on the basis of what? The word expect suggests that an initial evaluation was made. I sugget 'unforeseen' as more appropriate. In comparing USA, Southern and Easter China, it should be discussed that they have a very different geographical extent and make up of emissions within the regions, which makes such numbers somewhat comparing apples and pears.
l. 43-44 should be considered. I think most if not all models would include such feedbacks through atmospheric chemistry, but I leave it open whether these models have the correct response due to issues with model resolution, transport, chemistry, emissions. It is likely that the model study presented here, was more accurate than e.g used in MICS Asia, HTAP exercises, but that remains to be demonstrated.
l. 54 I recommend to use consistently in this study the units NOx-N and NH3-N for emissions. Currently it is not clear what this, most like NOx expressed as NO2. If the authors want to use these units, the unit need to be 30 Tg Nox(NO2). yr-1.
l. 57 Indeed the Qian paper report stagnant emissions for NH3, but I recommend to be somewhat careful with this, as that paper doesn't provide much information on what basis this assertion was made. A rough cross check could be made by comparing to trends in fertilizer production and imports and legumes import as a proxy for activity data that are ultimately causing NH3 emissions (in the absence of changing management and controls). As this could possibly change the outcomes of the study, it would be good to pay more attention to this aspect. L 69 nitric acid and (neutralized) aerosol nitrate and organic nitrates. List is not complete l. 70 OH react with NO2 (not with NOx in general). O3 reacts with NO to form NO2. O3 reacts with NO2 to form NO3, which under light conditions is rapidly photolysed. Please be more exact.
l. 89 See comment on figure S1. l. 92 social-economic=>just economic is sufficient in this context l. 96 so NH3-N deposition 5.8 Tg N? Please mention this explicitly. Comparing to the earlier mentioned NOx-N emission of 9.13 and NH3-N emissions of 8.2 Tg N, a back of the envelop calculation gives 57 % of NOx deposited on land-and 63.5 % of ammonia N? A budget analysis (including changes could be developed as strong key message of this paper) as the fraction remaining on land.
l. 97 here the authors should be more specific on the form of NO3: as gaseous HNO3 or in the form of aerosol nitrate? The specific form influences deposition strongly and we need to know whether this is part of the response function.
l. 103 see comments to Figure S2 l. 107 a substantial change of NOy lifetime will depend on the ability to form aerosol ammonium (and other ) nitrates. It is not clear what is compared in S2 (gas, aerosol, both?) and I think it is not nitrite/nitrous acid as erroneously in S2.
l. 112 what is the baseline case? Emissions constant or fluctuating over the years? Clarify shortly here, and methods for more details.
l. 116 see comment to Figure 2. The paper really needs to explore better the role of change in residence time and the role of atmospheric transport, including monsoon patterns.
l. 119 I think effectiveness is not the correct word-response ratio is more appropriate.
l. 122 this would point to more NO3 in the form of aerosol nitrate, which would work to prolong lifetime?
l. 143 what is present? l. 166 A thorough chemical budget analysis would pinpoint these hypothesis.
l. 333 this is the response to the earlier question (single year of emission or variable), please summarize in earlier sentence.
l. 357 what is measured. Aerosol nitrate, gaseous nitric acid or both together. Duly discuss possible sampling errors. l. 365 Why was not a similar comparison with NH3 made?  It would probably be more instructive to have a separate panel for 'other China'. What exactly is displayed in 2c? I think it is a scaled response to 30 % emission reduction, not an effectiveness, as part of the response can be due to transport. Also it took me until figure 3 to understand that the green colors over Eastern China are actually higher deposition. Probably it makes sense to first Figure 3 and then 2? Or add a sentence in the figure caption guiding the readers' interpretation. Fig 3: the enhanced response of NO3 deposition over Eastern China, in particular in winter, raises question what the chemical mechanism really is. Conventional wisdom points to heterogenous (N2O5 possibly also NO3 radical) reactions as the dominant driver of NOx removal, but it is not clear why these reaction rates would be enhanced in winter. Unfortunately the paper doesn't do a thorough job in analyzing these chemical mechanism when attempting to understand the drivers of these responses, although there may be clues in Figure 4, where in particular NO3 radical seems to have a strong response. More analysis needed.  : What is going on over Tibet with ratios close to zero "Black stars in figure 2c refer to the four observations in Figure 1?
Fig S1 These are nice plots, but it is difficult to understand the model skills. I recommend to include 2 additional panels that provide a scatter plot+some spatial skill statistics. Also to exclude variability issues it would be better to evaluate skills over a 5 year period rather than a single year. Figure S2 Probably the authors meant nitrate (not nitrite). The way this is presented suggest better model performance. A commonly used statistic is also the fraction of model results with a factor of 0.5 and 2 from observations.  Fu et al., 2020, Zhong et al., 2020, Wen et al., 2020, Xi et al., 2021 Therefore, I would hesitate to recommend this manuscript to be accepted in Nature Communication. 2) In the revised manuscript, we performed the land model simulations using CLM5 and gave associated analysis on the regional ecosystem responses to the changes in N deposition. These new results advance our understanding on the future trends in China's land carbon sinks and the pathways to achieve carbon neutrality goal. Please see Line 298-331 and Line 345-359 in the revised manuscript.  Only meteorological condition variations cannot explain the weakened response of NOy deposition to NOx emissions. This study proposes that the enhanced atmospheric oxidation capacity is a key driving factor. Please see Line 124-135.

Response
2) The major conclusions are unchanged even we use different climatological conditions (e.g., Please see Line 161-167. 3) We have shown that both the decreased NOx and SO2 emissions would increase the atmospheric oxidation capacities in our recent studies (Please see Liu et al., 2018Liu et al., , 2019Liu et al., , and 2020. The decreases of SO2 emissions and associated sulfate formation via the OH + SO2 pathway can increase particulate nitrate fractions in the total nitrate with increased ammonia availability (Please see Liu et al., 2018 andZhai et al., 2021); however, this would decrease NOy deposition because the deposition rate of aerosol nitrate is much lower than gas-phase nitrate. This study reveals that it was the NOx emission reduction that enhanced oxidation capacities and consequently increased the NOy dry deposition rate. 4) Though the emission control strategies to improve air quality under different climatological conditions was not the topic of this study, our results from the view of N deposition indicate that the coordinated controls for NOx and VOCs could reduce nitrate concentrations effectively despite the unfavorable meteorological conditions. Please see Line 283-297. The conclusion is definitely significant and novel, and has major implications on the evaluation of future emission strategies to minimize impacts on terrestrial ecosystems. There are, however, some important questions that need to be addressed before this paper can be published. See below.

Response:
We appreciate the reviewer's useful comments. All the comments are accepted and fully addressed. Please see our revisions in the following. with the inputs of N deposition from WRF-Chem simulations for different cases. The plant N uptake, net primary production (NPP), and net ecosystem production (NEP; used as an indicator of land carbon sink) from CLM5 were analyzed to evaluate the effects of reduced N deposition on terrestrial ecosystems.
2) The new results based on CLM5 demonstrated that the 30% reduction of NOx emission over China decreases N deposition by 1.4 Tg N yr -1 and subsequently decrease the NEP by 5.0 Tg C yr -1 through the limitation of plant N uptake and plant growth. We find that the unexpected weakened response of N deposition would lead to much less carbon sequestration in land ecosystems in the Eastern China compared the ideal case with a 100% relative response of N deposition (-0.56 Tg C yr -1 vs. -0.04 Tg C yr -1 ).
3) Our new analysis indicates that the projected emission reductions for NOx to improve air quality in China diminish ecosystem production. This N-control penalty may represent 5-10% of required land carbon sink to achieve carbon neutral over China by the next mid-century (Wang et al., 2014;Cheng et al., 2020). To our knowledge, it is the first time to project the impact of future controls of N emissions on natural carbon sink. 4) As suggested by the reviewer's next comments, the N deposition-ecosystem interactions are discussed in details (with references like Zhao et al., 2017;Sadiq et al., 2017;Zhou et al., 2018).
Please see Line 360-374 in the revised manuscript.
All these revisions have been updated in the revised text. Please see Line 298-331, Line 345-387, and Line 465-487 in the revised manuscript.
Reviewer #2: 2. One novel aspect and important implication of this paper is the potential impacts of future emission control strategies on terrestrial ecosystems (e.g., vegetation productivity, soil biogeochemistry, soil acidification, etc.). Thus far, these are mentioned only peripherally, but it should stand at the center of the main motivation/impact of the paper. At least a paragraph or two in the introduction and discussion are warranted to discuss how the varying responses of nitrogen deposition in Eastern vs. Southern China may affect different ecosystem processes (e.g., primary production, carbon uptake, soil nitrification and denitrification, soil pH, etc.), and suggest possible emission reduction strategies that can maximize benefits for both human health and ecosystem health.
Response: Accepted. We add substantial results and discussions on how future emission controls would impact different ecosystem processes. Please see the major revisions below and more details in the revised manuscript: 1) As aforementioned, we assess the effects of NOx emission controls on terrestrial production and carbon uptake by coupling the ecosystem model (CLM5) with the simulated N deposition.
2) The results based on CLM5 show that the decrease of NEP induced by the N deposition reduction in the Eastern China (0.73 g C m -2 yr -1 ) is appreciably lower than that in the Southern 3) Our findings along with some previous studies highlight the coordinated controls of NOx and volatile organic compounds to reduce atmospheric oxidation capacity (e.g., ozone pollution) and its harmful effects on human health and ecosystem production; importantly, the new implication is that the N-control penalty on natural carbon sink via plant N uptake should be considered in the pathways towards carbon neutrality, as it is possibly on the order of 5-10% of China's land carbon sink.
Please see detailed revisions at Line 298-331 in the revised manuscript. Reviewer #2: 3. As nitrogen deposition influences ecosystems tremendously, while vegetation and soil in ecosystems play major roles regulating surface air quality and atmospheric composition, there can be various feedback mechanisms whereby NOx reductions will lead to changes in nitrogen deposition, thus vegetation and soil, that ultimately affect air quality itself. Such feedback mechanisms were suggested by Zhao et al. (Zhao, Y. H., Zhang, L., Tai, A. P. K., Chen, Y. F., and Pan, Y. P.: Responses of surface ozone air quality to anthropogenic nitrogen deposition in the Northern Hemisphere, Atmos. Chem. Phys., 17, 9781-9796, doi:10.5194/acp-17-9781-2017, 2017, and can be potentially important. However, the current modeling framework in this paper would not allow such feedbacks to be examined because the land cover is mostly prescribed. At least the implications of these feedback mechanisms on the main results of this paper should be discussed. Response: Accepted. We add one paragraph discussing N deposition-ozone feedbacks mediated by NOx emission controls, as mentioned by Zhao et al. (2017). Considering the complex biosphereatmosphere interactions, we summarize that NOx emission controls can affect air quality (mainly ozone pollution) through the direct effects of NOx-VOCs chemistry on ozone formation and indirectly through the N deposition effects on plant growth (changing biogenic VOCs emissions and ozone dry deposition velocities) and soil NOx emissions (Zhao et al., 2017;Sadiq et al., 2017;and Zhou et al., 2018). Please see Line 360-374 in the revised manuscript.
This study analyses the mechanism of a non-linear response of reactive nitrogen deposition (in particular in Eastern China) to NOx emission reductions. While the issue of non-linear responses is well known in the atmospheric chemistry community, the interesting aspect is that emission changes are so large in a relatively short time (e.g. a downward emission trend of 20 % or more), making these feedbacks more visible than would be the case under gradually changing conditions. The manuscript is interesting, but there are several loose ends that require to be corroborated, additional testing and analysis.

Response:
We accepted all the reviewer's comments and revised the manuscript responsively.
Please see the point-to-point responses in the following. Only meteorological variations cannot explain the weakened responses of NOy deposition. We find that the chemical production of nitric acids driven by NOx emission reductions is a key factor in driving NOy deposition variations. Please see Line 124-135 in the revised manuscript.
Reviewer #3: 2) There is an interesting discrepancy between NO2 column observations and the anthropogenic emissions trends (e.g. Figure 1). This needs to be explored further along with an calculation on how representative NO2 is for NOy dry deposition, and how it is changing. This is relevant for emerging studies that use satellite observations to compute deposition of NOy (and NHx).
Response: Accepted. Our model results show that in response to a 30% reduction of NOx emissions, NO2 columns decreased by 33%, while NOy (including NOx and its oxidation products like nitric acids) dry deposition decreased by less than 20% for the Eastern China because enhanced nitric acids deposition fluxes largely offset the reduction of NOx deposition. So we suggest that using NO2 columns observed by satellites to derive total NOy deposition could bias the tendencies of inter-annual deposition, particularly for the Eastern China. Please see Figure  The emissions in the Eastern China also account for 26% of NOy deposition in the Southern China. These results suggest that the transboundary transport of NOy species can significantly affect the regional N deposition. We also demonstrate that the year-to-year variations of atmospheric transport do not affect our major findings. Please see the results in Line 90-108 and Line 161-167 in the revised manuscript. 2) The overall NOy lifetime is determined by using the ratio of the simulated NOy column concentrations to its deposition at regional scales. We find that the accelerated conversion of NOx to HNO3 results in a higher deposition rate of NOy and a reduction (from 4.5 to 4.1 days) of NOy lifetime over the Eastern China, while the NOy lifetimes are similar between the Base and the 30% NOx reduction cases for both the whole China and the Southern China.
Please see these revisions in the Line 90-108 and Line 198-214 in the revised manuscript. 1) The NOx removal by the photochemical reactions (NO2+OH) and nocturnal reactions (NO2+O3 and N2O5 formation) are analyzed in this study. In line with the recent studies (Wang et al., 2017;Chen et al., 2020;Fu et al., 2020), our simulations show that the gas-phase oxidation of NO2 by OH radical and nighttime reactions mediated by NO3 radical are major two pathways of nitric acid formation in wintertime with relative contributions of 40% vs. 60%, respectively, while the photochemistry pathway is dominant (90%) in summer. Importantly, this study finds that the enhanced chemical formation of nitric acids by NOx reductions is contributed by both the daytime (43%) and the nighttime (57%) reactions during wintertime.
2) A detailed analysis of the nitrate partitioning between gas and aerosol phases is given. The simulations with the reductions of NOx emissions show that about 50% of nitric acid presented in the gas phase, as ammonia is not enough to neutralize them completely to form ammonium nitrate. This is supported by the evidence that the fractions of particulate nitrate in the total nitrate (gas + particles) could be significantly elevated by increased ammonia availability in eastern China in our previous study (Liu et al., 2018) and Zhai et al. (2021). Therefore, the deposition rates of NOy turn to be higher due to increased gas-phase nitrate.
Please see all these revisions in Line 116-121, Line 184-197, and Line 215-234 in the revised manuscript.

Reviewer #3: 7) At several places in the publication the word effectiveness needs to be replaced, possibly by relative response or similar.
Response: Accepted. We replace most of them throughout the manuscript.

Reviewer #3:
I do think the work can be of interest to Nature if my major comments above, and detailed comments below are duly taking into account.
Response: Accepted. We fully addressed these comments. Please see the following responses.

Reviewer #3: Detailed comments l. 34 clarify largest in budget terms-or in terms of deposition fluxes?
Response: Accepted. We reword this sentence as: Terrestrial ecosystems in China receive the world's largest amount of reactive nitrogen (N) deposition. Please see Line 33-34 in the revised manuscript.

l. 37 It is not clear what is meant with 'integrating a model with observation'-I think the authors
mostly compare (with the exception of dry dep, that is using modelled deposition velocities. Please clarify.
Response: Accepted. The national observations are used to indicate the inter-annual tendency of N deposition fluxes and to validate the model. We reword this sentence as: By combining the observations with a chemical-transport model. Please see Line 36-37 in the revised text.

China, it should be discussed that they have a very different geographical extent and make up of emissions within the regions, which makes such numbers somewhat comparing apples and pears.
Response: Accepted. We replace the word 'lower than expected' by 'unforeseen' in this sentence.  Response: Accepted. We remove this sentence. The unexpected N emission-deposition response is demonstrated explicitly for the first time, while some other models in MICS-Asia and HTAP may be capable to consider such feedbacks. Importantly, we provide more results and analysis on ecosystem impacts based on a land model, as suggested by another reviewer.

Currently it is not clear what this, most like NOx expressed as NO2. If the authors want to use these units, the unit need to be 30 Tg Nox(NO2). yr-1.
Response: Accepted. In line with the deposition values mentioned in this study, we express the unit of emissions as Tg N yr -1 for NOx and NH3. Please see Line 53 in the revised manuscript.
l. 57 Indeed the Qian paper report stagnant emissions for NH3, but I recommend to be somewhat careful with this, as that paper doesn't provide much information on what basis this assertion was made. A rough cross check could be made by comparing to trends in fertilizer production and imports and legumes import as a proxy for activity data that are ultimately causing NH3 emissions (in the absence of changing management and controls). As this could possibly change the outcomes of the study, it would be good to pay more attention to this aspect.

l. 92 social-economic=>just economic is sufficient in this context
Response: Accepted. We reword this sentence. Please see Line 95-97 in the revised text. Response: Accepted. 1) We present the exact amount of NHx-N deposition (5.8 Tg N yr -1 ) in this sentence and also note it in Table S1. Response: Accepted. We clarify that the expression as NO3-N in our simulations is the sum of N presented as gaseous and particulate nitrate species and dinitrogen pentoxide.
l. 103 see comments to Figure S2 Response: Accepted. We add statistical indexes for the observation-model comparison along with the scatter plots.
l. 107 a substantial change of NOy lifetime will depend on the ability to form aerosol ammonium (and other) nitrates. It is not clear what is compared in S2 (gas, aerosol, both?) and I think it is not nitrite/nitrous acid as erroneously in S2.
Response: Accepted. We clarity that Figure S2 presented the comparison between the simulated aerosol nitrate concentrations and observations at in-situ stations over China. The observations for nitrite or nitrous acid were not available.

l. 112 what is the baseline case? Emissions constant or fluctuating over the years? Clarify shortly
here, and methods for more details.
Response: Accepted. The definition of the baseline case was given in the main text. We clarity again in this sentence. It was developed using the emissions and meteorological inputs for the year 2015. Please also see Table S1 for the description of all simulation cases. 1) The lifetime of NOy is estimated using the NOy column concentrations and deposition at regional scales. We find that the enhanced chemical conversion of NOx to HNO3 in the Eastern China by NOx reductions results in a larger deposition rate of NOy due to a much larger deposition rate of gaseous HNO3 than NOx species, and consequently a reduced lifetime of NOy from 4.6 days to 4.1 days. These results demonstrate why the relative response of NOy deposition to the NOx emission change is less than 70% in the region. Please see  in the revised manuscript.
2) Though atmospheric transport of N compounds is important for regional N deposition, our results for the response of N deposition to NOx emission changes are robust despite the variations of atmospheric transport including monsoon circulation between 2011and 2015(Li et al., 2016Zhang et al., 2019). Please see 161-167 in the revised manuscript.
l. 119 I think effectiveness is not the correct word-response ratio is more appropriate.
Response: Accepted. We replace most of them by the word 'relative response' throughout the manuscript. The word effectiveness has been used by Tan et al., 2020 to represent the ratio of relative changes in N deposition to the relative changes in its emission.
l. 122 this would point to more NO3 in the form of aerosol nitrate, which would work to prolong lifetime?
Response: We clarify that such increased NO3-N resides largely in the gas phase, pointing to decreased lifetime of NOy due to a larger deposition rate of gas-phase nitrate. The result combining with the following analysis suggests that the increased production rate of NO3-N due to enhanced oxidation capacity relieves the reductions of the regional NO3-N deposition.
Please see Line 184-197 in the revised text.

l. 143 what is present?
Response: Accepted. We have deleted this sentence.

l. 166 A thorough chemical budget analysis would pinpoint these hypothesis.
Response: Accepted. We show more quantitative results on how different chemical mechanisms Response: Accepted. We describe here and in earlier sentences that the baseline simulation was performed using the meteorological reanalysis data and the emission for the year 2015. Please see Table S1 for more details.   So we display as many stations as possible to reflect the long-term variations at regional scales.
Please see Figs. 1 and 2 for more details in the revised manuscript. Response: Accepted. We revise the manuscript as following: 1) The resulted for the Eastern China and the Southern China are presented separately in the panels.
Please see revised Figure 2.
2) As the reviewer suggested, it could be more understandable to use the word response than to use effectiveness. The figure and caption are therefore revised. While atmospheric transport is important for regional N deposition, those response changes of N deposition are the results of the overall reduction (local + regional) of NOx emissions in China. In Figure 2, the meaning of color scales is explained in the figure caption to guide readers.

REVIEWER COMMENTS</B>
Reviewer #1 (Remarks to the Author): Referee comments for NCOMMS-21-21037A-Z: "Unexpected response of nitrogen deposition to nitrogen oxide controls and implications for land carbon sink" The authors took the reviewer's comments into full consideration and revised the manuscript thoroughly. The paper is more clear and improved than previous version. However, I will not recommend the authors to discuss the effect of N deposition on land carbon sink in this paper, which make the topic of this study seems too big and unfocused. Of cause, the effect of N deposition on carbon balance is an interesting topic, but not the key point of this study. Furthermore, the regional simulation of earth system model or dynamic vegetation model also need multi-source input data, parameter localization, and data validation. A preliminary modelling analysis would bring large uncertainty in the estimation for the effects of declined N deposition on carbon sequestration.
For example, what is the land use and land cover data used in the simulation? Did you consider the effect of CO2, land use and land cover change, and N fertilization on NEP in the model? Did you compare the modeling NEP results with the observation results from eddy Flux tower? Line55-56：this reference seems rather old (Tian et al., good paper but not up to date).
Reviewer #2 (Remarks to the Author): The authors have made substantial revisions to the manuscript that have mostly addressed our concerns in an adequate and reasonable manner. In the light of some more recently published work, I recommend however that some more discussion should be warranted before this paper should be accepted. Please see below for details.
In a recent study, Liu et al. (2021Liu et al. ( ) (https://doi.org/10.5194/acp-21-17743-2021 showed that the vegetation responses to changes in nitrogen deposition in China are relatively small, because most of China's terrestrial ecosystems are not very nitrogen-limited. This paper, however, appears to show greater sensitivity of vegetation to changing nitrogen deposition. What could be the sources of differences? Would it have to do with, say, difference in the model used (Liu et al. used CLM4.5 while the authors here used CLM5), or the different treatment of nitrogen limitation as well as nitrogen biogeochemistry in general? That Liu et al. examined rising nitrogen deposition in the light of rising agricultural production, whereas the authors here examined decreasing nitrogen deposition due to emission control, and thus the sensitivity may be nonlinear? A more thorough comparison of this work with Liu et al. (2021), dissecting the differences and similarities, on a par with the other previous related work (e.g., Zhou et al., (2017)) is necessary for a compelling paper to be published.
Reviewer #3 (Remarks to the Author): The authors have largely adequately responded to my previous concerns. This is an interesting research paper, which shows that atmospheric chemistry effects can lead to non-linear responses to emission reductions in China, with unforeseen consequences for climate mitigation objectives. The additional carbon cycle analysis, is a valuable addition to the paper, but may need some additional discussion to cover all possible consequences. Although the English used in the manuscript is in general of a quite good level, but some additional editing would be helpful to remove the last glitches. This also applies to at some places, sloppy use of language that may be somewhat confusing to the reader.
All in all, I recommend this manuscript now for publication in Nature Communications, and I add below some minor suggestions for further improvement of the manuscript.
Minor suggestions>: l. 36 Thanks for revising this wording-here is a suggestion that I think covers better what you did: "Using a chemical transport model to understand observed spatial and temporal deposition trends, we show that … etc" l. 42 NOx reductions =>"Nox emission reductions" l. 42-43 This is a new part of the paper (and requested by one of the reviewers): to me it is not a priori clear why this is a penalty. It is probably related to relatively less nitrogen transport to regions where more nitrogen deposition could lead to an enhanced nitrogen sink.
l. 56 Suggest to use terrestrial instead of natural.
l. 90-100. The model observation correspondence displayed in Fig S1-S2 is objectively not great-and probably cosmetically looks somewhat better by the use of log-log scatter plots. Rather than claiming that this 'generally agreed', I suggest to focus on the given the arguments why this is fit-for-purpose, or is state-of-the-art. My expectation is that this model performance is quite dependent on the specifics of the observational network and the complexities of the pollution chemistry in China. I do suggest to have some additional analysis along with S1/S2 (and not in the main text).
l. 98 mention here what method you used to derive the the relative contribution to deposition in East and South Asia.
l. 108 Great. These results collectively suggest that the base simulation of WRF-CHEM is in line with literature knowledge on large scale transport process within China and between China and surrounding regions.
l. 140 *emissions* of other species constant.
L 142 average for the areas defined in 2c?
l. 144 Nationally=>Nationally aggregrated? l. 290: language. Do you mean enhance the efficacy? Suggest the additional VOC emission control led to a more linear response of NOy deposition to NOx emission controls l. 306 I like the additional carbon cycle simulations, but I also notice that they are not sufficiently described and discussed. I understand from the methods section that a 30 years simulations was performed under near-equilibrium conditions. However, the situation is of course far from equilibrium, and something needs to be said what the likely response to transient N-dep trajectories will be-either by adding 1 additional scenario, or by discussing some literature result. As the emission-deposition responses have a strong regional signature, I expect also some discussion on the covariation of ecosystem location and deposition changes. For instance if for some other reason it will still be attractive to preferentially reduce NOx emissions (without VOC), would a possible response be to also favor reforestation in Eastern China? Maybe this is not a very realistic suggestion, but something needs to said about it (in discussion?).

Response to reviewers' comments
Reviewer #1 (Remarks to the Author):

(comments are marked in italic)
Referee comments for NCOMMS-21-21037A-Z: "Unexpected response of nitrogen deposition to nitrogen oxide controls and implications for land carbon sink" Reviewer #1: The authors took the reviewer's comments into full consideration and revised the manuscript thoroughly. The paper is more clear and improved than previous version. However, I will not recommend the authors to discuss the effect of N deposition on land carbon sink in this paper, which make the topic of this study seems too big and unfocused. Of cause, the effect of N deposition on carbon balance is an interesting topic, but not the key point of this study.

Response:
We appreciate the reviewer's comment. The topic of this study was to unravel the mechanisms in the weakened response of N deposition to nitrogen oxide emission controls in China. The vast majority of our data analysis and the conclusion were focused on that. Interestingly, our study further demonstrates that the N deposition changes in response to emission controls can appreciably modulate ecosystem production (Please see Line 317-328 and 343-365 in the revised manuscript). This implication is strongly recommended by other two reviewers.
Reviewer #1: Furthermore, the regional simulation of earth system model or dynamic vegetation model also need multi-source input data, parameter localization, and data validation. A preliminary modelling analysis would bring large uncertainty in the estimation for the effects of declined N deposition on carbon sequestration.
For example, what is the land use and land cover data used in the simulation? Did you consider the effect of CO2, land use and land cover change, and N fertilization on NEP in the model? Did you compare the modeling NEP results with the observation results from eddy Flux tower?
Response: Accepted. In the revision, we followed CMIP6 protocol to rerun CLM5 in a transient period between 1850 and 2014 and updated the results with such comprehensive simulation (Please see Line 500-508). We provided more details of the input data and model configuration in our land model simulations: 3) Our team has been highly involved in the development of CLM. Dr. Yongjiu Dai, as one of the coauthors in this manuscript, has initialized the development of CLM (Dai et al., 2003). Another coauthor, Dr. Xingjie Lu, has recently improved the representation of carbon and nitrogen cycles in CLM5 simulations (Lu et al., 2020).
So we believe, with our revision, results of CLM5 would provide important reference for future estimates on land carbon sinks. 2) Difference in the emission scenarios. We applied a 30% reduction of anthropogenic NOx emissions across China, resulting in a 12% reduction of the total nitrogen (sum of reduced and oxidized nitrogen) deposition nationwide; while Liu et al. (2021) considered the rising nitrogen deposition by 5-15% in northern China and less than 5% in southern China. The decline in N fluxes may enhance nitrogen limitation. Hence, our modeling cases tend to produce larger changes in ecosystem production.
3) The regional contrast of N deposition responses is in line with previous related works. We show that the fertilization effect of nitrogen deposition was much more significant in southern China than in northern China (Fig. 4) Fig S1-S2 is objectively not great-and probably cosmetically looks somewhat better by the use of log-log scatter plots. Rather than claiming that this 'generally agreed', I suggest to focus on the given the arguments why this is fit-for-purpose, or is state-of-the-art. My expectation is that this model performance is quite dependent on the specifics of the observational network and the complexities of the pollution chemistry in China. I do suggest to have some additional analysis along with S1/S2 (and not in the main text). l. 109 NO3 is aerosol nitrate (NO3-)+HNO3 and N2O5 I think.

Response
Response: Accepted. We clarify that NO3-N is the sum of gas and aerosol NO3and N2O5. Please see Line 114.
l. 140 *emissions* of other species constant.
Response: Accepted. We reword the sentence as 'emissions of other species constant'. Please see Line 146-147.
l. 142 average for the areas defined in 2c?

Response:
The values shown here were the ratios of the changes in region-aggregated deposition to those of region-aggregated emissions for those areas of interest in Fig.  2c. Please see Line 149-151.
Response: Accepted. The results shown here were nationally aggregated deposition. Please see Line 150-151.
l. 290: language. Do you mean enhance the efficacy? Suggest the additional VOC emission control led to a more linear response of NOy deposition to NOx emission controls Response: Accepted. We reword the phrase as 'enhance the efficacy' in this sentence. Please see Line 291.
l. 306 I like the additional carbon cycle simulations, but I also notice that they are not sufficiently described and discussed. I understand from the methods section that a 30 years simulations was performed under near-equilibrium conditions. However, the situation is of course far from equilibrium, and something needs to be said what the likely response to transient N-dep trajectories will be-either by adding 1 additional scenario, or by discussing some literature result. As the emission-deposition responses have a strong regional signature, I expect also some discussion on the covariation of ecosystem location and deposition changes. For instance if for some other reason it will still be attractive to preferentially reduce NOx emissions (without VOC), would a possible response be to also favor reforestation in Eastern China? Maybe this is not a very realistic suggestion, but something needs to said about it (in discussion?).
Response: Accepted. Based on the reviewer's suggestion, we improve our analysis on the carbon cycle simulations as following: We update the modeling results of terrestrial carbon balance from transient simulations in the period of 1850-2014 that incorporate transient nitrogen deposition data from CMIP6 model. Three scenarios using different nitrogen deposition data (baseline N deposition, reduced N deposition, and ideally-reduced N deposition) were applied for the period of present day (2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014). By evaluating the magnitude and the spatial patterns of land carbon sinks among different nitrogen deposition scenarios, we found that the updated results in transient simulations fully support our conclusions that have been derived from original near-equilibrium simulation. Therefore, we replace original equilibrium results with new transient results. (Please see Line 500-514) We also agree that the effects of emission changes on ecosystem production were regionally dependent. The non-linear response of N deposition to NOx emission controls in Eastern China could feed back to ecosystem carbon sinks. Discussion on changing nitrogen fertilization effect on plant growth is included in the text.
(Please see Line 317-328 and Line 343-355) Figure 1: it is it possible to provide in the figure information where the stations are located, as this is important to understand the 'story'. Likewise, can the station codes be added to figure 2 panel c? Also somewhere the station coordinates should be listed.

Figures:
Response: Accepted. The locations of the stations are marked in Fig. 2c. Figure 2c: for which scenario (10/30/50) the results in 1c apply? mention in caption Response: Accepted. Figure 2c presents the results in the 30% reduction case. It is mentioned in the caption.
Supplement: Figure S2: although the authors claim to have corrected nitrite to nitrate, the error is still present.
Response: Accepted. We did correct it this time. Thanks for the reminder.