## Introduction

Anthropogenic ammonia (NH3) emissions, primarily from agriculture, have adversely affected environmental quality, including air pollution, soil acidification, eutrophication of water bodies, and led to tremendous damage to human health and ecosystem health1,2. The cost of damage associated with agricultural NH3 emissions was estimated at US dollars (US$)55–114 billion in the European Union (EU) in 2008, with the largest contribution due to increased human mortality from exposure to NH3-containing aerosols3,4. In the United States (US), annual health costs due to NH3 emissions were estimated at US$69–180 billion in 20115.

Mitigating NH3 emissions has attracted much attention in high-income countries. For example, the Gothenburg Protocol was signed in 1999 to control long-range transboundary transport of air pollutants among member countries within the United Nations Economic Commission for Europe. Following the Gothenburg protocol, the EU adopted the first National NH3 Emission Ceilings directive (2001/81/EC) in 20016. The efficacy and costs of NH3 abatement and their climate co-benefits were evaluated in 20157, focusing on European countries. To date, only a few countries have estimated their national NH3 mitigation potential and associated costs and benefits (Table 1).

China is the world’s largest emitter of NH3 (9–13 Tg N year−1 in the 2010s), with over 80% contributed by agriculture8,9. Low fertilizer nitrogen (N) use efficiency (NUE) and poor animal waste management have resulted in enormous NH3 emissions in China8,10. Worse still, regional NH3-related pollution is enhanced due to the increasing decoupling between crop and livestock production systems11. In recent years, frequent smog events with high concentrations of PM2.5 (fine particulate matter < 2.5 µm) in China have triggered both public anxiety and concerns of the Chinese government12. A substantial proportion of PM2.5 pollution was caused by aerosol formation driven by NH3 emissions13,14. Studies have suggested that the current clean air policy for reductions in primary PM2.5, sulfur dioxide (SO2), and nitrogen oxides (NOx) has limitations, and that PM2.5 pollution can be cost-effectively controlled only if NH3 emissions are abated as well as those of SO2 and NOx15,16,17. Studies have also suggested that many NH3 abatement techniques may simultaneously reduce agricultural methane (CH4) and nitrous oxide (N2O) emissions, bringing co-benefits for agricultural greenhouse gas (GHG) mitigation7,18,19,20. However, NH3 emission reduction in China may worsen the adverse impact of acid rain on crops and forests by increasing rainfall acidity21,22.

To date, China has not yet formulated or implemented policies to reduce NH3 emissions23, although there are many available measures to reduce NH3 emissions from agriculture, most of which have been validated and adopted in the EU and North America7,24. Many NH3 abatement measures have not been widely practiced in China and their implementation costs and the impacts on agricultural GHG emissions have not been assessed. Given that poor smallholder farmers still dominate China’s agricultural production and that agricultural N pollution is severe25, it is crucial to identify feasible and cost-effective NH3 abatement measures for Chinese agriculture.

A national systematic assessment of NH3 mitigation potential, and the associated costs and societal benefits, is urgently needed for China to establish cost-effective mitigation strategies and targets. To fill the knowledge gap, this study builds an integrated NH3 mitigation assessment framework (Supplementary Fig. 1) with the combination of Coupled Human And Natural Systems (CHANS), GAINS, Weather Research and Forecasting-Community Multiscale Air Quality (WRF-CMAQ), and exposure–response models to: (1) identify feasible NH3 abatement options and to estimate the agricultural NH3 mitigation potential and the associated implementation costs and societal benefits; (2) determine mitigation priorities and strategies for China; and (3) to explore optimal NH3 mitigation pathways for the next 30 years (2020–2050) using scenario analysis and cost-benefit assessment. We find that the relative NH3 mitigation potential in China is twice that in Europe. The overall societal benefits of agricultural NH3 mitigations in China far exceed the abatement cost and increase when including the synergy with reduction of GHG emissions.

## Results and discussion

### NH3 mitigation potential, abatement costs, and societal benefits

For cropping systems NH3 abatement measures include reductions of urea-based fertilizer, promotion of enhanced efficiency N fertilizer (EENF), and deep placement of fertilizer (Supplementary Table 1). The NH3 mitigation potential of crop production is around 2.0–3.4 Tg N year−1 at an abatement cost of US$1.9–3.4 billion. The three major staple crops in China have the largest NH3 mitigation potential at 460–954 Gg N for maize, 516–684 Gg N for rice, and 446–731 Gg N for wheat. The large reduction potential is mainly due to large sowing areas and poor fertilization practices. The production of vegetables and fruits consumes one-third of total synthetic N fertilizer use in China, and their NH3 mitigation potentials are estimated at 30–55% (269–493 Gg N) and 20–40% (118–235 Gg N), respectively. Unit abatement costs (US$ ha−1 year−1, Table 2) for cash crops (sugar, fruits and, vegetables) are higher than those for staple crops because the production of cash crops is more intensive, requiring higher inputs of manpower, fertilizer, and financial resources26.

### NH3 mitigation pathways in 2020–2050

Scenario analysis and cost-benefit assessment guide the optimization of NH3 mitigation strategies and pathways in the future. In this study, one baseline scenario of business as usual (BAU) and four mitigation scenarios (DIET, NUE, REC, and ALL) toward 2050 are analyzed, which comprise a range of packages of mitigation options (see Table 3 for details) to explore optimal mitigation pathways. The simulated NH3 emission trends (Fig. 2) for the next 30 years (2020–2050) reveal that there would be substantial NH3 mitigation potential with broad welfare benefits (Fig. 3).

Under the BAU scenario total agricultural NH3 emissions in China are estimated to first increase from 10.9 Tg N in 2015 to 12.1 Tg N in 2035 because of a growing and changing food demand for China’s increasing and wealthier population27. The emission would then slightly decrease to 11.9 Tg N following a decrease in China’s population toward 205032. NH3 emission from synthetic N fertilization is expected to remain stable during this period considering the national “Zero-growth Action Plan” for chemical fertilizer use33. The major cause of increased NH3 emission is the rising livestock production to meet the growing demand for animal products both in total and per capita terms27,34.

In contrast, the DIET scenario assumes optimizing human dietary structure (transitioning toward more plant-based diets) to reduce the animal-based food N to 40%, which is in line with Chinese dietary guidelines35. The increased human consumption of plant-based food N will shorten the food chain and decrease food-feed competition from decreased livestock farming. Decreased livestock production (meat, eggs, and milk) in DIET reduces the demand for crop production by 20–30% relative to the BAU scenario, which could reduce NH3 emission by 21% by 2050 (Fig. 2).

Based on the proposed improvement in NUE in crop and livestock production systems by adopting advanced farming practices, or techniques as identified in this study, agricultural NH3 emissions are projected to decline from 11.9 to 8.8 Tg N in 2050 under the NUE scenario (Fig. 3). This scenario could decrease synthetic N fertilizer use by 13 Tg N, decreasing NH3 emission from cropping systems by 39%. In addition, NH3 emission from livestock systems could be reduced by 1.9 Tg N through improved animal feed and waste management.

### Policy implications

To clean up the air, Chinese governments have already made major efforts to reduce anthropogenic SO2 and NOx emissions, which have declined by 41% and 34%, respectively, from their peaks to 201941. Although continuing the stringent policies to reduce SO2 and NOx emissions could further improve air quality, and may partially offset the effects of NH3 mitigation, studies have suggested that current policies are not sufficient or cost-effective in achieving the targets of clean air in China16,23,42. A recent study has found that reducing livestock NH3 emissions would be highly effective in reducing PM2.5 during severe winter haze events43. Our quantitative assessments of the implementation cost and societal effects of NH3 mitigation in China further demonstrate that NH3 mitigation could generate net societal benefits, even though it may worsen regional acid rain. Therefore, coordinated mitigation of multiple air pollutants (SO2, NOx, and NH3) should be implemented to more rationally and effectively achieve the dual benefits of protecting human and ecosystem health in China at both national and regional scales21.

For farm holders, strategic designs of cost-effective mitigation pathways are needed. The aforementioned cheap and easy mitigation options (direct reduction of N fertilizer use and improved animal feeding practices) should be introduced first to pick the low-hanging fruit of NH3 mitigation in China. The remaining mitigation measures (e.g., HA and manure handling systems) are expensive due to the higher requirements of the investments in technologies and infrastructures. It is necessary to increase government support (e.g., technical guidance and training) and subsidies (e.g., enhanced efficiency fertilizers and agricultural machinery) to encourage farmers to adopt these mitigation measures10. Perhaps even more challenging, the government should also promote the reform of China’s land tenure system to facilitate large-scale farming44. Large-scale farms will be a better platform for the implementation of advanced management practices and mechanization (e.g., deep application of fertilizers) and can reduce the abatement cost per unit cropland area10,25.

Livestock production is shifting from small-scale outdoor systems to large-scale intensive indoor systems27, which causes decoupling between croplands for feed production and industrial feedlots10. As a consequence, manure is increasingly dumped or discharged instead of being recycled or reused owing to high transportation costs, resulting in huge NH3 emissions in China11. In 2015 only 30% of livestock manure N was recycled to croplands in China11 while in the EU the proportion was more than 65%45. Relocating feedlots to feed croplands can greatly improve manure recycling and reduce the associated implementation costs where livestock densities being kept within the cropland carrying capacity for manure application11. Financial incentives (e.g., subsidies, discounted interest, technical guidance, taxation exemption, etc.) are required to help farmers develop a region-specific farming structure that facilitates manure recycling, optimizes N management and promotes large-scale operation27.

Furthermore, it should be noted that NH3 mitigation through human dietary changes, also benefits human health46,47. Dietary change is a nontechnical measure with little implementation cost but requires other interventions to change consumers’ preferences. The government can play an essential role in setting up campaigns to promote low-protein diets.

### Limitations and uncertainty

Agricultural NH3 mitigation strategies are linked to the overall N cycle and could affect agricultural production and farmers’ incomes7, which may further influence food security and rural economies. This study did not incorporate the effects of NH3 mitigation on crop yield or animal productivity in the cost-benefit assessment of scenarios owing to the lack of comprehensive Chinese-specific data. In fact, fertilizer N application in China far exceeds the crop demand; NH3 mitigation by improved farming practices would unlikely create N limitation or reduce crop yields19,31. If the yield benefits attributed to NH3 mitigation could be quantified rationally and accurately, it would greatly improve the cost-effectiveness of NH3 mitigation and therefore engage farmers to adopt these measures. Besides, this study does not address the regional difference in China due to the lack of detailed regional data. Given the large differences in regional agricultural structures and environmental conditions, mitigation strategies and targets may vary considerably, which affects the accuracy of current national estimates.

In this paper we limited the climate benefits to non-CO2 GHG (CH4 and N2O) emissions resulting from NH3 mitigation. This is mainly because CO2 emission from agriculture is more related to fossil fuel consumption, such as fertilizer production and transportation48, which is beyond the scope of this study. Furthermore, we did not consider the effects of NH3 mitigation on climate change, including changes in aerosols and carbon sinks in terrestrial ecosystems, owing to limited research and models that target China49,50.

There are complex chemical interactions among SO2, NOx, and NH3 in the atmosphere51. Thus, the future policies to control SO2 and NOx emissions may partially offset the effects of NH3 mitigation, which also bring uncertainties to our estimations. While the projections of NH3 mitigation potential and costs toward 2050 are based on current technologies, future technological advancement, and policy optimization may further reduce the implementation cost to yield a higher NEB. Nevertheless, as the first attempt, this study provides a basis and reference for on-going improvement in NH3 mitigation potential and cost-benefit assessment.

## Methods

### Selection of available mitigation options

To identify feasible and cost-effective NH3 abatement measures for Chinese agriculture, we reviewed all currently available mitigation options in China and other countries. Our criteria for the selection of abatement measures focus on five aspects:

1. (1)

Mitigation efficiency: measures that could significantly reduce NH3 emissions are included, for example, deep manure placement has a very high mitigation potential at 93–99%52.

2. (2)

Implementation cost: measures with lower cost or labor inputs are more acceptable to farmers, for example, reduced use of urea-based fertilizer and lower crude protein diet.

3. (3)

Practical applicability: measures with current limited applicability due to technical, political or obvious social barriers in China, were excluded, for example, soil testing has been ruled out in this study due to high costs for the small farm size and high spatial and temporal variability, although it is an effective measure to optimize fertilizer use in the US and Europe where farm sizes are much larger.

4. (4)

Limitations: measures that likely and significantly reduce agricultural productivity (crop yield or animal productivity) were adopted with caution, for example, the full substitution of synthetic fertilizers by manure decreases the yield of upland crops and lowland rice by 9.6% and 4.1%, respectively53; and low crude protein (LCP) feeding should only be adopted to an applicable level to avoid undermining animal productivity and welfare. Besides, LCP is mainly applicable to indoor animals (pig, poultry, and dairy).

5. (5)

Presence of co-benefit: measures that could reduce both NH3 emission and total GHG (CH4 and N2O) emissions are included, for example, biochar additives could reduce NH3, N2O and CH4 emissions during manure composting54.

Based on the selection criteria and literature review, a total of 27 technical mitigation options for specific crops and animals were included in this study, with a coded version provided in Supplementary Tables 1 and 2. Detailed descriptions of these options are listed in Supplementary Tables 36 and Supplementary Notes 13. Most of these mitigation measures have been validated and adopted in the EU, while some of the measures (e.g., optimum N fertilization techniques) have been validated in China. For the measures that have been validated in China we directly adopted their parameters, whereas for measures that have not been validated in China, we calculated their potential implementation costs based on China-specific parameters such as labor cost, fertilizer prices, machinery cost. Only cost-effective measures and their combinations were selected for the analysis.

Most agricultural NH3 and GHG emissions originate from the same activities (Supplementary Fig. 2) and their emission rates depend on common factors, such as management practice, weather conditions and soil type7. NH3 abatement options can increase or decrease GHG emission20. This study aims to explore the maximum NH3 mitigation potential while achieving the co-benefits of GHG reduction. In this context, optimal combinations of NH3 mitigation options for different crops and animals are proposed in Supplementary Table 7 with their abatement efficiencies.

### Mitigation potential of NH3 emissions

NH3 emissions from agriculture generally are assessed by multiplying the activity level with specific emission factors for each sector. The NH3 mitigation potential is calculated as Eq. (1):

$${\mathrm{\Delta }}E_{NH_3} = \mathop {\sum}\limits_i {{\mathrm{A}}_{i,k}\, * \,{\mathrm{EF}}_{i,k} \times \eta _{i,k} \times X_{i,k}},$$
(1)

where $${\mathrm{\Delta }}E_{NH_3}$$ is the reduction of agricultural NH3 emissions in mainland China; i represents the source type; k means a specific abatement option or the combination of multiple options; $${\mathrm{A}}_{i,k}$$ is the activity data of the source type; $${\mathrm{EF}}_{i,k}$$ is the original emission factor; $$\eta _{i,k}$$ is the NH3 abatement efficacy; $$X_{i,k}$$ is the implementation rate of the abatement technique or options.

Abatement cost of NH3 emissions in this study is defined as direct expenditures (the sum of investment costs and operation costs) for implementation of measures to reduce NH3 emissions from agriculture, while the possible public costs (e.g., subsidy to promote the control policy) are not considered. Here, we mainly refer to the methodology of cost assessment from the GAINS model55 to calculate the abatement costs of implementing various NH3 mitigation measures. China-specific commodity prices were collected mainly from the China Agricultural Products Cost-Benefit Yearbook (2000–2018)26, European cost data were adopted by conversion at market exchange rates where data supply is insufficient. All cost data from the literature were adjusted by the purchasing power parity (PPP) index and measured in constant 2015 US$(e.g., 100 EUR = US$113.49, 100 CNY = US$14.89) by assuming a 2% annual inflation and setting 2015 as the base year for future projection. The calculation of abatement costs is formulated in Eq. (2): $${\mathrm{AC}}_{i,k} = {\mathrm{I}}_{i,k}\, * \,\left[ {\frac{{\left( {1 + r} \right)^{lt_{i,k}} \times r}}{{\left( {1 + r} \right)^{lt_{i,k}}\, - \, 1}}} \right] + {\mathrm{FO}}_{i,k} + {\mathrm{VO}}_{i,k} - {\mathrm{FS}}_{i,k},$$ (2) where ACi,k represents the annual implementation cost; Ii,k refers to the investment cost; r is the discount rate; lti,k represents the lifetime of abatement technique (10–15 years); $${\mathrm{FO}}_{i,k}$$ is the fixed operating cost; $${\mathrm{VO}}_{i,k}$$ is the variable operating costs (e.g., feed, gas, electricity, labor, and water); FSi,k means saving costs from reduced use of N fertilizer. Investment cost Ii,k is calculated as a function of the average farm size ($${\mathrm{AFS}}_i$$) by Eq. (3): $${I}_{i,k} = ci_{i,k}^f \cdot st_i \cdot mp_i \cdot pc_i + \frac{{ci_{i,k}^v}}{{{\mathrm{AFS}}_i}},$$ (3) where $$ci^f,ci^v$$ represents the fixed and variable coefficients derived from Klimont and Winiwarter (Annex: Table A1)56; sti represents manure storage time (in year); mpi represents manure production of a single animal per year; pci represents production cycles per year; parameters used in the function are available in an online GAINS report. Annual fixed operating costs $${\mathrm{FO}}_{i,k}$$ are estimated as the 0.05% of the total investments by Eq. (4) according to GAINS cost calculation21. $${\mathrm{FO}}_{i,k} = {I}_{i,k} \cdot 0.05\%.$$ (4) Variable operating costs $${\mathrm{VO}}_{i,k}$$ are calculated by cost summation of the quantity (Q) of a certain extra supply (e.g. feed, gas, electricity, water, and labor) for a specific abatement option (k), as shown in Eq. (5): $${\mathrm{VO}}_{i,k} = \mathop {\sum}\limits_p {\left( {{Q}_{i,k} \cdot c_{i,k}} \right)},$$ (5) where p represents parameter type (additional feed, gas, electricity, water and labor input); $$c_{i,k}$$ means the unit price of these extra supply, which is mainly derived from the China Agricultural Products Cost-Benefit Yearbook26 and market survey or adjusted by a coefficient if no direct data source could be accessed. The unit labor cost of farmworkers in 2015 is 15.7 Chinese yuan (CNY) per hour according to the national averaged salary for individual persons26,57. Other relevant parameters used in the calculation of FO and VO are obtained from GAINS. The cost-effectiveness of various NH3 mitigation options was calculated following Eq. (6)55,58 to yield MACC curve according to increasing cost-effectiveness, as shown in Fig. 1. $${\mathrm{CE}}_{i,k} = \frac{{{\mathrm{CE}}_{i,k}\, * \,\eta _{i,k} - {\mathrm{CE}}_{i,k - 1}\, * \,\eta _{i,k - 1}}}{{\eta _{i,k} - \eta _{i,k - 1}}},$$ (6) where $$CE_{i,k}$$ is the cost-effectiveness for mitigation option k; $$\eta _{i,k}$$ is the NH3 mitigation efficiency. ### Scenario analysis of future NH3 emissions To explore the feasibility of NH3 mitigation in the future, the CHANS model was employed in this study to make systematic and comprehensive analyses of NH3 sources, emissions, and environmental fates8. A detailed introduction of the model can be found in Zhang et al.8 and Gu et al.59. Taking into consideration the impacts of policy, and other external factors on Chinese agricultural production and consumption, the baseline NH3 emission budgets during 2020–2050 were built in the first place, then four abatement scenarios (DIET, NUE, REC, and ALL) with corresponding packages of mitigation measures (detailed description in Table 3) were integrated into the CHANS model to quantify the new NH3 emission budgets and identify the feasible NH3 reduction potential in China. Human population and the per capita gross domestic product are two important parameters that affect future NH3 emission budgets. These two parameters are assumed to remain the same as the BAU for the four mitigation scenarios while other input drivers and parameters, such as diet structure, NUE, cropping area, animal numbers, will vary with scenarios (Supplementary Fig. 6). Details about the data sources, prediction methods and parameters can be found in Supplementary Tables 818 and Supplementary Note 4. It should be noted that change in human diet structure as a nontechnical measure was also included in the scenario analysis to obtain a more comprehensive assessment of the mitigation potential and pathways. ### Societal benefit assessment of NH3 mitigation Societal benefits ($${\mathrm{SOC}}_{{\mathrm{benefit}}}$$) of NH3 mitigation in this study are defined as the sum of benefits for human health ($${\mathrm{HH}}_{{\mathrm{benefit}}}$$), ecosystem health ($${\mathrm{EH}}_{{\mathrm{benefit}}}$$), GHG mitigation benefit ($${\mathrm{GHG}}_{{\mathrm{benefit}}}$$) minus the cost of damage by increased acidity of precipitation ($${\mathrm{AR}}_{{\mathrm{damage}}}$$, as shown in Eq. (7) $${\mathrm{SOC}}_{{\mathrm{benefit}}} = {\mathrm{HH}}_{{\mathrm{benefit}}} + {\mathrm{EH}}_{{\mathrm{benefit}}} + {\mathrm{GHG}}_{{\mathrm{benefit}}} - {\mathrm{AR}}_{{\mathrm{damage}}}.$$ (7) The human health benefit assessment was performed based on the exposure–response function and the Value of Statistical Life (VSL) as applied in earlier studies both at the global and national scales1. Five categories of diseases causing premature mortality via PM2.5 pollution are considered in this study, namely lower respiratory tract illness, chronic obstructive pulmonary disease (COPD), ischemic heart disease (IHD), lung cancer (LC) and cerebrovascular disease (CEV). The impacts of NH3 mitigation on annual PM2.5 concentration were assessed based on the model simulation of Weather WRF-CMAQ performed by Xu et al.60. A deduced nonlinear function between PM2.5 concentration and NH3 reduction was built in Eq. (8); detailed description of WRF-CMAQ simulation can be found in Xu et al.60 and Supplementary Note 5. Then, an exposure–response function (Eq. (9)) was combined with the health effect function (Eq. (10)) based on Global Burden of Disease61 to estimate the 1-year premature mortality attributable to PM2.5 exposure. Afterward, we used the updated Chinese-specific VSL following the method in Giannadaki et al.1 to derive corresponding economic benefits of NH3 abatement by Eq. (11) in China. $$C_j = C_{2015}\, \times \,(1 - 0.0173\, \times \,e^{2.9532 \times \eta _j}),$$ (8) $${\mathrm{HE}}_{j,q} = \mathop {\sum}\limits_j {e^{\beta _q\, \times \,(C_j - C_o)}\, \times \,{\mathrm{HE}}_{0,q}},$$ (9) $${\mathrm{\Delta }}M_j = \mathop {\sum}\limits_q {\left( {{\mathrm{HE}}_{j,q} - {\mathrm{HE}}_{0,q}} \right)\, \times \,{\mathrm{Pop}}_j},$$ (10) $${\mathrm{HH}}_{{\mathrm{benefit}},j} = VSL_j\, \times \,{\mathrm{\Delta }}M_j,$$ (11) where Cj is the annual average PM2.5 concentration in year j; C2015 is the annual average PM2.5 concentration in year 2015 (50 µg m−3); $$\eta _j$$ is the reduction rate (%) of NH3 emission; q represents the category of diseases (IRL, COPD, IHD, LC, CEV); $$\beta _q$$ is the coefficient in the exposure–response function which refers to the proportion of change in the endpoint of each health effect of the population for unit change in PM2.5 concentration; C0 is the background concentration below which no health impact is assumed (10 µg m−3 as suggested by the WHO62); HE0,q is the baseline health effect (the mortality risk) due to a particular disease category for China estimated by the WHO61; HEj,q is the actual health effect under significant PM2.5 pollution levels; Popj is the population exposed to air population in China; $${\mathrm{\Delta }}M_j$$ is the avoided death toll; VSLj is the Chinese-specific value of a statistical life derived from Giannadaki et al.1; $${\mathrm{HH}}_{{\mathrm{benefit}},j}$$ means the human health benefits by NH3 mitigation. Ecosystem benefits in this study are regarded as the avoided damage cost of decreased acidification and eutrophication of ecosystems due to NH3 mitigation. We assume the unit NH3 damage cost to the ecosystem in the European Nitrogen Assessment21 is also applicable to China after correction for differences in the willingness to pay (WTP) for ecosystem service and PPP in China and EU, as shown in Eq. (12). $${\mathrm{EH}}_{{\mathrm{benefit}},j} = {\mathrm{\Delta }}E_{{\mathrm{NH}}_3,j}\, \times \,\partial _{{\mathrm{EU}}}\, \times \,\frac{{{\mathrm{WTP}}_{{\mathrm{China}}}}}{{{\mathrm{WTP}}_{{\mathrm{EU}}}}}\, \times \,\frac{{{\mathrm{PPP}}_{{\mathrm{China}},j}}}{{{\mathrm{PPP}}_{{\mathrm{EU}},j}}},$$ (12) where $$\partial _{{\mathrm{EU}}}$$ is the estimated unit ecosystem damage cost of NH3 emission in relation to soil acidification and water eutrophication in Europe based on the European Nitrogen Assessment63; WTPChina and WTPEU are the values of the WTP for ecosystem service in China and Europe; PPPChina,j and PPPEU,j stand for the PPP of China and the EU. GHG benefit from NH3 mitigation is regarded as the avoided abatement cost of GHG emissions, as shown in Eqs. (13)–(14). $${\mathrm{GHG}}_{{\mathrm{benefit}},j} = {\mathrm{\Delta }}E_{{\mathrm{GHG}},j}\, * \,{\mathrm{MCost}}_{{\mathrm{GHG}},j},$$ (13) $${\mathrm{\Delta }}E_{{\mathrm{GHG}},j} = \left( {{\mathrm{\Delta }}E_{{\mathrm{directN}}_2{\mathrm{O}},j} + {\mathrm{\Delta }}E_{{\mathrm{indirectN}}_2{\mathrm{O}},j}} \right)\, * \,298 + {\mathrm{\Delta }}E_{{\mathrm{CH}}_4,j}\, * \,34,$$ (14) where $${\mathrm{\Delta }}E_{{\mathrm{GHG}},j}$$ is the estimated reduction in agricultural GHG emissions, presented as kg CO2-eq, using the default values of 298 kg CO2-eq for N2O emissions, and 34 kg CO2-eq for CH4 emissions at a 100-year time horizon64; both reduction of direct and indirect N2O emissions are included, the indirect N2O reduction is calculated as 1% of reduced NH3 deposition according to the IPCC guideline65. $${\mathrm{MCost}}_{{\mathrm{GHG}},j}$$ represents the marginal abatement cost (the carbon price) to reduce one tonne of GHG emissions in$ per tonne CO2-eq, the Chinese-specific (East Asia) value is derived from West et al.66.

Acid rain damage ($${\mathrm{AR}}_{{\mathrm{damage}},j}$$) induced by NH3 mitigation refers to the economic loss of reduced crop yields ($${\mathrm{Crop}}_{{\mathrm{damage}},j}$$) and forestry ($${\mathrm{Forest}}_{{\mathrm{damage}},j}$$) in Eq. (15). Based on the experimental results reported in Feng et al.22 and model simulation of precipitation acidity in Liu et al.21, we estimated the economic cost of increased acid rain under different mitigation scenarios.

$${\mathrm{AR}}_{{\mathrm{damage}},j} = {\mathrm{Crop}}_{{\mathrm{damage}},j} + {\mathrm{Forest}}_{{\mathrm{damage}},j}.$$
(15)

### Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.