Negative correlation between soil salinity and soil organic carbon variability.

Soil organic carbon (SOC) is vital for terrestrial ecosystems, affecting biogeochemical processes, and soil health. It is known that soil salinity impacts SOC content, yet the specific direction and magnitude of SOC variability in relation to soil salinity remain poorly understood. Analyzing 43,459 mineral soil samples (SOC < 150 g kg-1) collected across different land covers since 1992, we approximate a soil salinity increase from 1 to 5 dS m-1 in croplands would be associated with a decline in mineral soils SOC from 0.14 g kg-1 above the mean predicted SOC ([Formula: see text] = 18.47 g kg-1) to 0.46 g kg-1 below [Formula: see text] (~-430%), while for noncroplands, such decline is sharper, from 0.96 above [Formula: see text] = 35.96 g kg-1 to 4.99 below [Formula: see text] (~-620%). Although salinity's significance in explaining SOC variability is minor (<6%), we estimate a one SD increase in salinity of topsoil samples (0 to 7 cm) correlates with respective [Formula: see text] declines of ~4.4% and ~9.26%, relative to [Formula: see text] and [Formula: see text]. The [Formula: see text] decline in croplands is greatest in vegetation/cropland mosaics while lands covered with evergreen needle-leaved trees are estimated with the highest [Formula: see text] decline in noncroplands. We identify soil nitrogen, land cover, and precipitation Seasonality Index as the most significant parameters in explaining the SOC's variability. The findings provide insights into SOC dynamics under increased soil salinity, improving understanding of SOC stock responses to land degradation and climate warming.

The technological and economic prospects for CO2 utilization and removal.

The capture and use of carbon dioxide to create valuable products might lower the net costs of reducing emissions or removing carbon dioxide from the atmosphere. Here we review ten pathways for the utilization of carbon dioxide. Pathways that involve chemicals, fuels and microalgae might reduce emissions of carbon dioxide but have limited potential for its removal, whereas pathways that involve construction materials can both utilize and remove carbon dioxide. Land-based pathways can increase agricultural output and remove carbon dioxide. Our assessment suggests that each pathway could scale to over 0.5 gigatonnes of carbon dioxide utilization annually. However, barriers to implementation remain substantial and resource constraints prevent the simultaneous deployment of all pathways.

A restatement of the natural science evidence base concerning grassland management, grazing livestock and soil carbon storage.

Approximately a third of all annual greenhouse gas emissions globally are directly or indirectly associated with the food system, and over a half of these are linked to livestock production. In temperate oceanic regions, such as the UK, most meat and dairy is produced in extensive systems based on pasture. There is much interest in the extent to which such grassland may be able to sequester and store more carbon to partially or completely mitigate other greenhouse gas emissions in the system. However, answering this question is difficult due to context-specificity and a complex and sometimes inconsistent evidence base. This paper describes a project that set out to summarize the natural science evidence base relevant to grassland management, grazing livestock and soil carbon storage potential in as policy-neutral terms as possible. It is based on expert appraisal of a systematically assembled evidence base, followed by a wide stakeholders engagement. A series of evidence statements (in the appendix of this paper) are listed and categorized according to the nature of the underlying information, and an annotated bibliography is provided in the electronic supplementary material.

Not all soil carbon is created equal: Labile and stable pools under nitrogen input.

Anthropogenic activities have raised nitrogen (N) input worldwide with profound implications for soil carbon (C) cycling in ecosystems. The specific impacts of N input on soil organic matter (SOM) pools differing in microbial availability remain debatable. For the first time, we used a much-improved approach by effectively combining the 13C natural abundance in SOM with 21 years of C3-C4 vegetation conversion and long-term incubation. This allows to distinguish the impact of N input on SOM pools with various turnover times. We found that N input reduced the mineralization of all SOM pools, with labile pools having greater sensitivity to N than stable ones. The suppression in SOM mineralization was notably higher in the very labile pool (18%-52%) than the labile and stable (11%-47%) and the very stable pool (3%-21%) compared to that in the unfertilized control soil. The very labile C pool made a strong contribution (up to 60%) to total CO2 release and also contributed to 74%-96% of suppressed CO2 with N input. This suppression of SOM mineralization by N was initially attributed to the decreased microbial biomass and soil functions. Over the long-term, the shift in bacterial community toward Proteobacteria and reduction in functional genes for labile C degradation were the primary drivers. In conclusion, the higher the availability of the SOM pools, the stronger the suppression of their mineralization by N input. Labile SOM pools are highly sensitive to N availability and may hold a greater potential for C sequestration under N input at global scale.

Cover crops do not increase soil organic carbon stocks as much as has been claimed: What is the way forward?

When compared to virgin land (forest and grassland), croplands store significantly lower amounts of organic carbon (OC), mainly as a result of soil tillage, and decreased plant inputs to the soil over the whole year. Doubts have been expressed over how much reduced and zero tillage agriculture can increase OC in soils when the whole soil profile is considered. Consequently, cover-crops that are grown in-between crops instead of leaving soils bare appear as the "last man standing" in our quest to enhance cropland OC stocks. Despite the claim by numerous meta-analyses of a mean carbon sequestration rate by cover crops to be as high as 0.32 ± 0.08 ton C ha-1 year-1 , the present analysis showed that all of the 37 existing field studies worldwide only sampled to a depth of 30 cm or less and did not compare treatments on the basis of equivalent soil mass. Thirteen studies presented information on OC content only and not on OC stocks, had inappropriate controls (n = 14), had durations of 3 years or lower (n = 5), considered only one to two data points per treatment (n = 4), or used cover crops as cash crops (i.e., grown longer that in-between two crops) instead of catch crops (n = 2), which in all cases constitutes shortcomings. Of the remaining six trials, four showed non-significant trends, one study displayed a negative impact of cover crops, and one study displayed a positive impact, resulting in a mean OC storage of 0.03 ton ha-1 year-1 . Models and policies should urgently adapt to such new figure. Finally, more is to be done not only to improve the design of cover-crop studies for reaching sound conclusions but also to understand the underlying reasons of the low efficiency of cover crops for improved carbon sequestration into soils, with possible strategies being suggested.

Avoiding emissions versus creating sinks-Effectiveness and attractiveness to climate finance.

The perception of greater impact via new sinks, as opposed to through avoided emissions, has already led some large investors to focus on sink-related projects. This is a flawed perception when applied universally and carries a risk that effective routes to mitigation through avoiding emissions are side-lined. In reality, both emissions avoidance and emissions removal are needed, and both can be a cost-effective means of delivering mitigation.

The role of rice cultivation in changes in atmospheric methane concentration and the Global Methane Pledge.

Resumption of the increase in atmospheric methane (CH4 ) concentrations since 2007 is of global concern and may partly have resulted from emissions from rice cultivation. Estimates of CH4 emissions from rice fields and abatement potential are essential to assess the contribution of improved rice management in achieving the targets of the Global Methane Pledge agreed upon by over 100 countries at COP26. However, the contribution of CH4 emissions from rice fields to the resumed CH4 growth and the global abatement potential remains unclear. In this study, we estimated the global CH4 emissions from rice fields to be 27 ± 6 Tg CH4 year-1 in the recent decade (2008-2017) based on the 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The trend of CH4 emissions from rice cultivation showed an increase followed by no significant change and then, a stabilization over 1990-2020. Consequently, the contribution of CH4 emissions from rice fields to the renewed increase in atmospheric CH4 concentrations since 2007 was minor. We summarized the existing low-cost measures and showed that improved water and straw management could reduce one-third of global CH4 emissions from rice fields. Straw returned as biochar could reduce CH4 emissions by 12 Tg CH4 year-1 , equivalent to 10% of the total reduction of all anthropogenic emissions. We conclude that other sectors than rice cultivation must have contributed to the renewed increase in atmospheric CH4 concentrations, and that optimizing multiple mitigation measures in rice fields could contribute significantly to the abatement goal outlined in the Global Methane Pledge.

Robust paths to net greenhouse gas mitigation and negative emissions via advanced biofuels.

Biofuel and bioenergy systems are integral to most climate stabilization scenarios for displacement of transport sector fossil fuel use and for producing negative emissions via carbon capture and storage (CCS). However, the net greenhouse gas mitigation benefit of such pathways is controversial due to concerns around ecosystem carbon losses from land use change and foregone sequestration benefits from alternative land uses. Here, we couple bottom-up ecosystem simulation with models of cellulosic biofuel production and CCS in order to track ecosystem and supply chain carbon flows for current and future biofuel systems, with comparison to competing land-based biological mitigation schemes. Analyzing three contrasting US case study sites, we show that on land transitioning out of crops or pasture, switchgrass cultivation for cellulosic ethanol production has per-hectare mitigation potential comparable to reforestation and severalfold greater than grassland restoration. In contrast, harvesting and converting existing secondary forest at those sites incurs large initial carbon debt requiring long payback periods. We also highlight how plausible future improvements in energy crop yields and biorefining technology together with CCS would achieve mitigation potential 4 and 15 times greater than forest and grassland restoration, respectively. Finally, we show that recent estimates of induced land use change are small relative to the opportunities for improving system performance that we quantify here. While climate and other ecosystem service benefits cannot be taken for granted from cellulosic biofuel deployment, our scenarios illustrate how conventional and carbon-negative biofuel systems could make a near-term, robust, and distinctive contribution to the climate challenge.

Modelling methane emissions and grain yields for a double-rice system in Southern China with DAYCENT and DNDC models.

Methane (CH4) is an important greenhouse gas that contributes to climate change and one of its major sources is rice cultivation. The main aim of this paper was to compare two well-established biogeochemical models, namely Daily Century (DAYCENT) and DeNitrification-DeComposition (DNDC) for estimating CH4 emissions and grain yields for a double-rice cropping system with tillage practice and/or stubble incorporation in the winter fallow season in Southern China. Both models were calibrated and validated using field measured data from November 2008 to November 2014. The calibrated models performed effectively in estimating the daily CH4 emission pattern (correlation coefficient, r = 0.58-0.63, p < 0.001), but model efficiency (EF) values were higher in stubble incorporation treatments, with and without winter tillage (treatments S and WS) (EF = 0.22-0.28) than that in winter tillage without stubble incorporation treatment (W) (EF = -0.06-0.08). We recommend that algorithms for the impacts of tillage practice on CH4 emission should be improved for both models. DAYCENT and DNDC also estimated rice yields for all treatments without a significant bias. Our results showed that tillage practice in the winter fallow season (treatments WS and W) significantly decreased annual CH4 emissions, by 13-37 % (p < 0.05) for measured values, 15-20 % (p < 0.05) for DAYCENT-simulated values, and 12-32 % (p < 0.05) for DNDC-simulated values, respectively, compared to no-till practice (treatments S), but had no significant impact on grain yields.

1,135 ionomes reveal the global pattern of leaf and seed mineral nutrient and trace element diversity in Arabidopsis thaliana.

Soil is a heterogeneous reservoir of essential elements needed for plant growth and development. Plants have evolved mechanisms to balance their nutritional needs based on availability of nutrients. This has led to genetically based variation in the elemental composition, the 'ionome', of plants, both within and between species. We explore this natural variation using a panel of wild-collected, geographically widespread Arabidopsis thaliana accessions from the 1001 Genomes Project including over 1,135 accessions, and the 19 parental accessions of the Multi-parent Advanced Generation Inter-Cross (MAGIC) panel, all with full-genome sequences available. We present an experimental design pipeline for high-throughput ionomic screenings and analyses with improved normalisation procedures to account for errors and variability in conditions often encountered in large-scale, high-throughput data collection. We report quantification of the complete leaf and seed ionome of the entire collection using this pipeline and a digital tool, Ion Explorer, to interact with the dataset. We describe the pattern of natural ionomic variation across the A. thaliana species and identify several accessions with extreme ionomic profiles. It forms a valuable resource for exploratory genetic mapping studies to identify genes underlying natural variation in leaf and seed ionome and genetic adaptation of plants to soil conditions.

Land-based climate solutions for the United States.

Meeting end-of-century global warming targets requires aggressive action on multiple fronts. Recent reports note the futility of addressing mitigation goals without fully engaging the agricultural sector, yet no available assessments combine both nature-based solutions (reforestation, grassland and wetland protection, and agricultural practice change) and cellulosic bioenergy for a single geographic region. Collectively, these solutions might offer a suite of climate, biodiversity, and other benefits greater than either alone. Nature-based solutions are largely constrained by the duration of carbon accrual in soils and forest biomass; each of these carbon pools will eventually saturate. Bioenergy solutions can last indefinitely but carry significant environmental risk if carelessly deployed. We detail a simplified scenario for the United States that illustrates the benefits of combining approaches. We assign a portion of non-forested former cropland to bioenergy sufficient to meet projected mid-century transportation needs, with the remainder assigned to nature-based solutions such as reforestation. Bottom-up mitigation potentials for the aggregate contributions of crop, grazing, forest, and bioenergy lands are assessed by including in a Monte Carlo model conservative ranges for cost-effective local mitigation capacities, together with ranges for (a) areal extents that avoid double counting and include realistic adoption rates and (b) the projected duration of different carbon sinks. The projected duration illustrates the net effect of eventually saturating soil carbon pools in the case of most strategies, and additionally saturating biomass carbon pools in the case of forest management. Results show a conservative end-of-century mitigation capacity of 110 (57-178) Gt CO2 e for the U.S., ~50% higher than existing estimates that prioritize nature-based or bioenergy solutions separately. Further research is needed to shrink uncertainties, but there is sufficient confidence in the general magnitude and direction of a combined approach to plan for deployment now.

Observation-based global soil heterotrophic respiration indicates underestimated turnover and sequestration of soil carbon by terrestrial ecosystem models.

Soil heterotrophic respiration (Rh ) refers to the flux of CO2 released from soil to atmosphere as a result of organic matter decomposition by soil microbes and fauna. As one of the major fluxes in the global carbon cycle, large uncertainties still exist in the estimation of global Rh , which further limits our current understanding of carbon accumulation in soils. Here, we applied a Random Forest algorithm to create a global data set of soil Rh , by linking 761 field observations with both abiotic and biotic predictors. We estimated that global Rh was 48.8 ± 0.9 Pg C year-1 for 1982-2018, which was 16% less than the ensemble mean (58.6 ± 9.9 Pg C year-1 ) of 16 terrestrial ecosystem models. By integrating our observational Rh with independent soil carbon stock data sets, we obtained a global mean soil carbon turnover time of 38.3 ± 11 year. Using observation-based turnover times as a constraint, we found that terrestrial ecosystem models simulated faster carbon turnovers, leading to a 30% (74 Pg C) underestimation of terrestrial ecosystem carbon accumulation for the past century, which was especially pronounced at high latitudes. This underestimation is equivalent to 45% of the total carbon emissions (164 Pg C) caused by global land-use change at the same time. Our analyses highlight the need to constrain ecosystem models using observation-based and locally adapted Rh values to obtain reliable projections of the carbon sink capacity of terrestrial ecosystems.

How do we best synergize climate mitigation actions to co-benefit biodiversity?

A multitude of actions to protect, sustainably manage and restore natural and modified ecosystems can have co-benefits for both climate mitigation and biodiversity conservation. Reducing greenhouse emissions to limit warming to less than 1.5 or 2°C above preindustrial levels, as outlined in the Paris Agreement, can yield strong co-benefits for land, freshwater and marine biodiversity and reduce amplifying climate feedbacks from ecosystem changes. Not all climate mitigation strategies are equally effective at producing biodiversity co-benefits, some in fact are counterproductive. Moreover, social implications are often overlooked within the climate-biodiversity nexus. Protecting biodiverse and carbon-rich natural environments, ecological restoration of potentially biodiverse and carbon-rich habitats, the deliberate creation of novel habitats, taking into consideration a locally adapted and meaningful (i.e. full consequences considered) mix of these measures, can result in the most robust win-win solutions. These can be further enhanced by avoidance of narrow goals, taking long-term views and minimizing further losses of intact ecosystems. In this review paper, we first discuss various climate mitigation actions that evidence demonstrates can negatively impact biodiversity, resulting in unseen and unintended negative consequences. We then examine climate mitigation actions that co-deliver biodiversity and societal benefits. We give examples of these win-win solutions, categorized as 'protect, restore, manage and create', in different regions of the world that could be expanded, upscaled and used for further innovation.

Actions to halt biodiversity loss generally benefit the climate.

The two most urgent and interlinked environmental challenges humanity faces are climate change and biodiversity loss. We are entering a pivotal decade for both the international biodiversity and climate change agendas with the sharpening of ambitious strategies and targets by the Convention on Biological Diversity and the United Nations Framework Convention on Climate Change. Within their respective Conventions, the biodiversity and climate interlinked challenges have largely been addressed separately. There is evidence that conservation actions that halt, slow or reverse biodiversity loss can simultaneously slow anthropogenic mediated climate change significantly. This review highlights conservation actions which have the largest potential for mitigation of climate change. We note that conservation actions have mainly synergistic benefits and few antagonistic trade-offs with climate change mitigation. Specifically, we identify direct co-benefits in 14 out of the 21 action targets of the draft post-2020 global biodiversity framework of the Convention on Biological Diversity, notwithstanding the many indirect links that can also support both biodiversity conservation and climate change mitigation. These relationships are context and scale-dependent; therefore, we showcase examples of local biodiversity conservation actions that can be incentivized, guided and prioritized by global objectives and targets. The close interlinkages between biodiversity, climate change mitigation, other nature's contributions to people and good quality of life are seldom as integrated as they should be in management and policy. This review aims to re-emphasize the vital relationships between biodiversity conservation actions and climate change mitigation in a timely manner, in support to major Conferences of Parties that are about to negotiate strategic frameworks and international goals for the decades to come.

Deceleration of Cropland-N2O Emissions in China and Future Mitigation Potentials.

Agricultural soils are the largest anthropogenic emission source of nitrous oxide (N2O). National agricultural policies have been implemented to increase crop yield and reduce nitrogen (N) losses to the environment. However, it is difficult to effectively quantify crop-specific and regional N2O mitigation priorities driven by policies, due to lack of long-term, high-resolution crop-specific activity data, and oversimplified models. Here, we quantify the spatiotemporal changes and key drivers of crop-specific cropland-N2O emissions from China between 1980 and 2017, and future N2O mitigation potentials, using a linear mixed-effect model and survey-based data set of agricultural management measures. Cropland-N2O emissions from China tripled from 102.5 to 315.0 Gg N yr-1 between 1980 and 2017, and decelerated since 1998 mainly driven by country-wide deceleration and decrease in N rate and the changes in sowing structure. About 63% of N2O emissions could be reduced in 2050, primarily in the North China Plain and Northeast China Plain; 83% of which is from the production of maize (33%), vegetables (27%), and fruits (23%). The deceleration of N2O emissions highlights that policy interventions and agronomy practices (i.e., optimizing N rate and sowing structure) are potential pathways for further ambitious N2O mitigation in China and other developing countries.

How Modelers Model: the Overlooked Social and Human Dimensions in Model Intercomparison Studies.

There is a growing realization that the complexity of model ensemble studies depends not only on the models used but also on the experience and approach used by modelers to calibrate and validate results, which remain a source of uncertainty. Here, we applied a multi-criteria decision-making method to investigate the rationale applied by modelers in a model ensemble study where 12 process-based different biogeochemical model types were compared across five successive calibration stages. The modelers shared a common level of agreement about the importance of the variables used to initialize their models for calibration. However, we found inconsistency among modelers when judging the importance of input variables across different calibration stages. The level of subjective weighting attributed by modelers to calibration data decreased sequentially as the extent and number of variables provided increased. In this context, the perceived importance attributed to variables such as the fertilization rate, irrigation regime, soil texture, pH, and initial levels of soil organic carbon and nitrogen stocks was statistically different when classified according to model types. The importance attributed to input variables such as experimental duration, gross primary production, and net ecosystem exchange varied significantly according to the length of the modeler's experience. We argue that the gradual access to input data across the five calibration stages negatively influenced the consistency of the interpretations made by the modelers, with cognitive bias in "trial-and-error" calibration routines. Our study highlights that overlooking human and social attributes is critical in the outcomes of modeling and model intercomparison studies. While complexity of the processes captured in the model algorithms and parameterization is important, we contend that (1) the modeler's assumptions on the extent to which parameters should be altered and (2) modeler perceptions of the importance of model parameters are just as critical in obtaining a quality model calibration as numerical or analytical details.

Climate-smart soils.

Soils are integral to the function of all terrestrial ecosystems and to food and fibre production. An overlooked aspect of soils is their potential to mitigate greenhouse gas emissions. Although proven practices exist, the implementation of soil-based greenhouse gas mitigation activities are at an early stage and accurately quantifying emissions and reductions remains a substantial challenge. Emerging research and information technology developments provide the potential for a broader inclusion of soils in greenhouse gas policies. Here we highlight 'state of the art' soil greenhouse gas research, summarize mitigation practices and potentials, identify gaps in data and understanding and suggest ways to close such gaps through new research, technology and collaboration.

Advanced glycation end product intake during pregnancy and offspring allergy outcomes: A Prospective cohort study.

BACKGROUND: Associations have been shown between concurrent assessment of dietary intake of advanced glycation end products (AGEs) and childhood allergic outcomes. We examined the association between maternal AGEs intake and development of offspring asthma, wheeze, atopic dermatitis, allergic rhinitis and food allergies, and sought to determine whether the intake of AGEs was associated with cord sera cytokines/chemokines. METHODS: Pregnant women ≥16 years were recruited in the Healthy Start study, a prospective pre-birth cohort from Colorado (N = 1410). The analysis included 962 dyads with adequate diet (≥2 recalls) and allergy outcome details. AGEs intake was estimated for each mother by matching intakes reported using 24-h dietary recalls during pregnancy to a reference database of commonly consumed foods' AGEs values. Child diagnoses of asthma and allergies up to 8 years were obtained from electronic medical records. Cord sera cytokines and chemokines were analysed in a subset (N = 462) of children. RESULTS: The median [IQR] AGEs intake for the overall sample was 11,919 kU/day [8293, 16,573]. Unadjusted analysis showed a positive association between maternal AGEs intake in pregnancy and rhinitis up to 8 years of age (HR = 1.03; 95% CI: 1.01, 1.06), but the association was attenuated and no longer significant in adjusted models (HR = 1.01; 95% CI: 0.98, 1.04). Both adjusted and unadjusted models showed no associations between AGEs intake in pregnancy and any of the other outcomes (p > .05). There were no significant associations between any cytokine or chemokine measured and AGEs intake or any of the outcomes studied (p > .05). CONCLUSION: The study showed that maternal AGEs intake was not associated with offspring asthma and allergy outcomes. AGEs exposure during pregnancy may not have the same impact on child development as postnatal exposure.

Can cropland management practices lower net greenhouse emissions without compromising yield?

Smart cropland management practices can mitigate greenhouse gas (GHG) emissions while safeguarding food security. However, the integrated effects on net greenhouse gas budget (NGHGB) and grain yield from different management practices remain poorly defined and vary with environmental and application conditions. Here, we conducted a global meta-analysis on 347 observation sets of non-CO2 GHG (CH4 and N2 O) emissions and grain yield, and 412 observations of soil organic carbon sequestration rate (SOCSR). Our results show that for paddy rice, replacing synthetic nitrogen at the rate of 30%-59% with organic fertilizer significantly decreased net GHG emissions (NGHGB: -15.3 ± 3.4 [standard error], SOCSR: -15.8 ± 3.8, non-CO2 GHGs: 0.6 ± 0.1 in Mg CO2 eq ha-1  year-1 ) and improved rice yield (0.4 ± 0.1 in Mg ha-1  year-1 ). In contrast, intermittent irrigation significantly increased net GHG emissions by 11.2 ± 3.1 and decreased rice yield by 0.4 ± 0.1. The reduction in SOC sequestration by intermittent irrigation (15.5 ± 3.3), which was most severe (>20) in alkaline soils (pH > 7.5), completely offset the mitigation in CH4 emissions. Straw return for paddy rice also led to a net increase in GHG emissions (NGHGB: 4.8 ± 1.4) in silt-loam soils, where CH4 emissions (6.3 ± 1.3) were greatly stimulated. For upland cropping systems, mostly by enhancing SOC sequestration, straw return (NGHGB: -3.4 ± 0.8, yield: -0.5 ± 0.6) and no-tillage (NGHGB: -2.9 ± 0.7, yield: -0.1 ± 0.3) were more effective in warm climates. This study highlights the importance of carefully managing croplands to sequester SOC without sacrifice in yield while limiting CH4 emissions from rice paddies.

Climate change may interact with nitrogen fertilizer management leading to different ammonia loss in China's croplands.

Despite research into the response of ammonia (NH3 ) volatilization in farmland to various meteorological factors, the potential impact of future climate change on NH3 volatilization is not fully understood. Based on a database consisting of 1063 observations across China, nonlinear NH3  models considering crop type, meteorological, soil and management variables were established via four machine learning methods, including support vector machine, multi-layer perceptron, gradient boosting machine and random forest (RF). The RF model had the highest R2 of 0.76 and the lowest RMSE of 0.82 kg NH3 -N ha-1 , showing the best simulation capability. Results of model importance indicated that NH3 volatilization was mainly controlled by total input of N fertilizer, followed by meteorological factors, human managements and soil characteristics. The NH3 emissions of China's cereal production (paddy rice, wheat and maize) in 2018 was estimated to be 3.3 Mt NH3 -N. By 2050, NH3 volatilization will increase by 23.1-32.0% under different climate change scenarios (Representative Concentration Pathways, RCPs), and climate change will have the greatest impact on NH3 volatilization in the Yangtze river agro-region of China due to high warming effects. However, the potential increase in NH3 volatilization under future climate change can be mitigated by 26.1-47.5% through various N fertilizer management optimization options.