Fertilizer nitrogen (N) can accumulate in the soil-vadose zone-groundwater continuum as legacy N, exerting a consistent influence on key biogeochemical processes such as crop N uptake, N leaching, and gaseous N emissions. However, the regional stock of legacy N in soils and its contribution to soil N pools remain unknown due to the lack of effective estimation methods, hindering the understanding of its environmental and agricultural effects. Here, we proposed a method for quantifying legacy N stock at the regional scale by identifying an exponential decay pattern in multi-year N retention rates and establishing a validated power function fitting between a parameter related to the initial retention rate and the N application rate. We further proposed a novel N isotope model to constrain the contribution of legacy N to soil N pools. Overall, China's upland croplands have accumulated legacy N of 54.7 +/- 24.9 Tg N in top 30 cm soils over the past six decades (1961-2020). Croplands in southern China are hotspots for legacy N accumulation, likely due to their rich soil organic carbon. Legacy N may account for 43%-89% of soil organic N sequestration. The N isotope model estimates that legacy N constitutes 21.4% +/- 6.9% of total soil N, comparable with the compiled proportion of total N increases due to fertilization (16.0%). These findings suggest that the legacy N from historical fertilizers substantially enhances soil fertility, and may inform biogeochemical parameterization given that the components of soil N pools may differ in stability.

The unidirectional flow of phosphorus (P) from mineral reserves to croplands, freshwaters and ultimately to marine environments surpasses sustainable limits for human development. Addressing the dual challenges of rising food production and mitigating P pollution necessitates spatially explicit assessments. However, significant uncertainties persist regarding the spatial patterns of cropland-P runoff in China, largely due to inadequate methodologies for accurately estimating P runoff that incorporate climate variability, soil characteristics, and P fertilization rates. Here, we conducted a fine-gridded estimation of cropland-P runoff across China, using a data-driven artificial intelligence model in conjunction with county-level P fertilization data. Results indicated that cropland-P runoff in 2018 totaled 109.9 kilotons P yr−1, representing 0.79% of China's mineral P fertilization, aligning with national surveys from 2017. Notably, P runoff fluxes from uplands exhibited an inverse relationship with latitude, highlighting hotspots in regions south of the Yangtze River, driven by lower soil pH, intensive crop ration, and elevated soil organic matter. In rice paddies, hotspots were concentrated in the Pearl River Delta, Yangtze River Delta and western Sichuan basin, linked to higher irrigation and P fertilization. These patterns suggest that mineral-fertilizer-induced cropland-P runoff hotspots mainly occur where high P input coincides with runoff-prone soil, climate and irrigation conditions. This study provides a spatially explicit basis for prioritizing future monitoring and management of cropland-P runoff in these vulnerable regions.

The Huang (Yellow River)-Huai (Huaihe)-Hai (Haihe) region (3H region) is an area with dense population, industry and agriculture in China, and is vital to the national economic development. Water scarcity in this region has become critical especially due to climate change and intensified human activities. It is therefore essential to quantify the impact of water pricing policies on water consumption for optimizing water consumption structure and improving water use efficiency. This paper analyzed the water-saving effect of water pricing policies based on nonlinear fixed-effect models considering users' behaviors, and compared the water price elasticities in domestic and industrial sectors among provincial regions during 2007-2016. The results show that the price elasticity of industrial water demand approached unit elasticity, whereas the price elasticity of domestic water demand is relatively low. Consequently, water pricing policies during 2007-2016 reduced industrial water demand by 55.9% and domestic water demand by only 23.8%. By 2030, water price needs to be raised about 45% to balance the total water demand, as the growth per capita Gross Domestic Product (GDP) and industrial Gross Value Added (GVA) are projected to drive significant rise in water demand (both P < 0.001). This study revealed the limitation of current water pricing policies on regional water conservation, which needs a combination with other measures to systematically promote water use efficiency and alleviate water stress.

Nitrogen use efficiency (NUE) serves as a useful performance indicator to benchmark agricultural nitrogen management. It is determined by genotype-environment-management interaction, making the quantification of the key drivers and the mechanism behind spatiotemporal trends of NUE challenging. In this study, the soil Water Heat Carbon Nitrogen Simulator (WHCNS) model, integrated with multi-scale datasets, was applied to clarify the key drivers of the NUE trends for single rice in China from 1978 to 2019. The national NUE for single rice was 0.31 over the study period and showed a slight increase before 2005 and afterwards a rapid increase of 0.048/10 years, primarily controlled by fertilization management and cultivar shifts. The N application rate determined the total N input, influenced the N uptake and utilization efficiency, and therefore the NUE, while the effect of cultivar shifts on NUE is mainly determined by the increase in the proportion of photosynthetic products in the rice grain and thus the harvest index. The benefits of cultivar improvement (0.034/10 years) which were almost entirely offset by the excessive use of N fertilizers (-0.029/10 years) before 2005, became lower (0.021/10 years) and unstable thereafter in many subregions, indicating the challenge of cultivar selection under low N inputs. To reach the ambitious NUE target of 0.6 by 2050 for rice, improving NUE through better management alone will not be enough. Improvements in rice breeding are therefore urgently needed to meet future NUE challenges in rice production under climate change and population growth.

Gaseous nitrogen (N) losses from nitrification and denitrification (NO + N 2 O + N 2 ) pathways contribute a significant fraction of the total N losses from cropland ecosystems. The N mass balance and process‐based models are commonly applied to estimate the NO + N 2 O + N 2 losses but have suffered from systematic error accumulations or model over‐parameterization, leading to a large uncertainty in estimation, hindering effective management of the global N budget. Here, we proposed a novel N isotope model, which considers fertilizer, ammonia volatilization and harvest after testing steady‐state assumption of soil δ 15 N and N pool for croplands, and justified if it could be successfully applied to constrain NO + N 2 O + N 2 losses from cropland ecosystems. We compiled the first bulk‐soil δ 15 N data set of 0–30 cm soils ( n = 738) from croplands and produced a global map of cropland soil δ 15 N, which is crucial input data for an isotope model to quantify NO + N 2 O + N 2 losses. The results show that the cropland soil δ 15 N ranges from 3.5 to 9.0‰, with a mean value of 6.6 ± 0.8‰ (mean ± standard deviation). The estimated NO + N 2 O + N 2 losses accounted for an average of 17 ± 9% of N outputs and were 35.86 ± 24.17 kg N ha −1 yr −1 in China's rice paddies, with an increasing trend from Central China to South or North China. The estimations were comparable with the results from observation‐constrained denitrification‐decomposition modeling (38.9 ± 4.8 kg N ha −1 yr −1 ) and in good agreement with experimental observations at site scale ( R 2 = 0.58). Our results suggest that soil N isotopes, as a quantitative tracer, provide a valuable alternative approach to constrain the NO + N 2 O + N 2 losses in croplands at large geographic scales.

Agricultural intensification produces indirect emissions beyond ammonia volatilization from activities such as machinery usage, food processing, transportation, storage and energy inputs. Here we integrate an input-output analysis with air quality modelling approaches, showing that attributable mortality from indirect emissions has risen sixfold in China over the past 37 years. Indirect emissions now account for one-quarter of air pollution-related attributable mortality associated with food consumption. We find a marked redistribution of the indirect health burden, with low-income groups experiencing an additional 58% attributable deaths compared with their expected food consumption burdens, which were initially associated with the food consumption of high-income groups. Targeted strategies using abatement approaches could halve the indirect health burden, thereby mitigating the environmental impact of food consumption amid agricultural intensification.

Atmospheric ammonia (NH3) has multiple impacts on the environment, climate change, and human health. China is the largest emitter of NH3 globally, with the dynamic inventory of NH3 emissions remaining uncertain. Here, we use the second national agricultural pollution source censuses, integrated satellite data, 15N isotope source apportionment, and multiple models to better understand those key features of NH3 emissions and its environmental impacts in China. Our results show that the total NH3 emissions were estimated to be 11.2 ± 1.1 million tonnes in 2020, with three emission peaks in April, June, and October, primarily driven by agricultural sources, which contributed 74% of the total emissions. Furthermore, employing a series of quantitative analyses, we estimated the contribution of NH3 emissions to ecosystem impacts. The NH3 emissions have contributed approximately 22% to secondary PM2.5 formation and a 16.6% increase in nitrogen loading of surface waters, while ammonium deposition led to a decrease in soil pH by 0.0032 units and an increase in the terrestrial carbon sink by 44.6 million tonnes in 2020. Reducing agricultural NH3 emissions in China would contribute to the mitigation of air and water pollution challenges, saving damage costs estimated at around 22 billion US dollars due to avoided human and ecosystem health impacts.

Food consumption contributes to the degradation of air quality in regions where food is produced, creating a contrast between the health burden caused by a specific population through its food consumption and that faced by this same population as a consequence of food production activities. Here we explore this inequality within China's food system by linking air-pollution-related health burden from production to consumption, at high levels of spatial and sectorial granularity. We find that low-income groups bear a 70% higher air-pollution-related health burden from food production than from food consumption, while high-income groups benefit from a 29% lower health burden relative to their food consumption. This discrepancy largely stems from a concentration of low-income residents in food production areas, exposed to higher emissions from agriculture. Comprehensive interventions targeting both production and consumption sides can effectively reduce health damages and concurrently mitigate associated inequalities, while singular interventions exhibit limited efficacy.

Denitrification plays a critical role in soil nitrogen (N) cycling, affecting N availability in agroecosystems. However, the challenges in direct measurement of denitrification products (NO, N2O, and N2) hinder our understanding of denitrification N losses patterns across the spatial scale. To address this gap, we constructed a data-model fusion method to map the county-scale denitrification N losses from China's rice fields over the past decade. The estimated denitrification N losses as a percentage of N application from 2009 to 2018 were 11.8 +/- 4.0% for single rice, 12.4 +/- 3.7% for early rice, and 11.6 +/- 3.1% for late rice. The model results showed that the spatial heterogeneity of denitrification N losses is primarily driven by edaphic and climatic factors rather than by management practices. In particular, diffusion and production rates emerged as key contributors to the variation of denitrification N losses. These findings humanize a 38.9 +/- 4.8 kg N ha-1 N loss by denitrification and challenge the common hypothesis that substrate availability drives the pattern of N losses by denitrification in rice fields. The quantification of denitrification N losses has been challenging and limiting our understanding of denitrification N losses patterns in spatial scales. To address this gap, we constructed a "double constraint method" based on multiscale observations to optimize the DNDC model. We quantified the magnitude of denitrification N losses in Chinese rice fields and found that production rate and diffusion rate were more important than substrate concentration in determining the spatial variability of differentiation N losses. This insight provides an improved approach for simulating denitrification process.image

Agricultural activities contribute almost half of the total anthropogenic nitrous oxide (N2O) emissions, but proper assessment of mitigation measures is hampered by large uncertainties during the quantification of cropland N2O emissions and mitigation potentials. This review summarizes the up-to-date datasets and approaches to provide spatially explicit and crop-specific assessment of the global mitigation potentials. Here, we show that global cropland N2O emissions have quadrupled to 1.2 Tg N2O-N year−1 over 1961–2020. The mitigation potential is 0.7 Tg N2O-N without compromising the crop production, with 86% from optimizing nitrogen fertilization, three-quarters (78%) from maize (22%), vegetables, and fruits (16%), other crops (15%), wheat (13%), and rice (12%), and over 80% from South Asia, China, the European Union, other American countries, the United States, and Southeast Asia. More accurate estimation of cropland N2O mitigation potentials requires extending the N2O observation network, improving modeling capacity, quantifying the feasibility of mitigation measures, and seeking additional mitigation measures.

Foreign investments in overseas coal-fired power plants (OCPs) largely impede decarbonization efforts, yet their global carbon dioxide (CO2) emissions have not been sufficiently quantified. Here we analyse investment data from 908 OCPs worldwide, and then reveal current annual emissions of 0.53 GtCO2 yr-1 and a historical cumulative total of 26 GtCO2. Developed nations account for 78% of these cumulative emissions on the basis of investments, while emissions from developing nations have surged from 8% in 1960 to 39% in 2022. Assuming unchanged policies and technologies, OCPs are projected to contribute an additional 15-30 GtCO2 in cumulative emissions by 2060 directly. Furthermore, they could stimulate local coal power growth in emerging economies, potentially adding 6.3-45.0 GtCO2 emissions indirectly. Our study underscores the critical importance of low-carbon policy interventions in emerging countries to curb the power-sector carbon emissions increasingly influenced by international capital. Foreign investments in recent years drive the expansion of fossil fuel electricity generation in emerging economies, yet necessary quantification still lacks. This Article shows how overseas coal-fired power plants could drive future trajectories of CO2 emissions in host countries.

Abstract Atmospheric ammonia (NH3) has multiple impacts on the environment, climate change and human health. China is the largest emitter of NH3 globally, with the dynamic inventory of NH3 emission remaining uncertain. Here, we use the high-resolution secondary national pollution survey, integrated satellite data, 15N isotope source apportionment and multiple models to better understand those key features of NH3 emissions and its impacts in China. Our results show that the total NH3 emissions were estimated to be 12.3 Tg yr-1 in 2017 with three emission peaks in April, June and October. NH3 emissions have contributed approximately 23% to secondary PM2.5 formation, a 19.7% increase in nitrogen loading of surface waters, while ammonium deposition led to a decrease in soil pH by 0.0035 units and an increase in the terrestrial carbon sink by 83.4 Tg C yr-1. Reducing NH3 emissions in China would contribute to the mitigation of air and water pollution challenges, saving damage costs estimated at around 22 billion US dollars due to avoided human and ecosystem health impacts.

Rice production consumes large amounts of fertilizer and irrigation water and contributes to the non-point source pollution (NPSP) by delivering nitrogen (N) and phosphorus (P) through surface and subsurface fluxes. However, due to the lack of spatially detailed datasets we barely understand the temporal variations of nutrient fluxes or what the key drivers are. Here, we developed a data-driven model and a rice-specific dataset of fertilizer and irrigation schemes to quantify N and P loss fluxes from China's rice paddies across 2320 counties during 1979-2018. Both of the surface and subsurface fluxes decreased,-16 % for N and-28 % for P over the past four decades. Instead of a steady decrease, the N surface and subsurface fluxes slightly increase at first, until decrease since 1999 at rates of -1.2 and -4.7 kiloton N yr-2, respectively. Similarly, the P surface and subsurface fluxes decline rates jumped to -0.4 and -0.2 kiloton P yr -2, respectively. N loss via subsurface flux was slightly prevailing compared with surface runoff, their trends were mainly driven by fertilizer application and extreme rainfall, respectively. P was primarily exported via surface runoff, with its trends jointly driven by extreme rainfall and fertilizer application, while the trends of subsurface P fluxes were largely driven by the latter. Consequently, N and P require different mitigation strategies, with N focusing on optimized fertilizer application and P on optimized water scheme. The combination of mitigation strategies on 20 % of the national rice sowing area accounted for 48 % of N and 70 % of P mitigation potentials under future warming. These findings suggest that we should implement regional specific policy interventions for critical regions and pathways, to effectively alleviate the national NPSP.

The impacts of low soil moisture (SM) and high vapour pressure deficit (VPD) on tree's photosynthesis and productivity are ultimately realized by changing water content in the canopy leaves. In this study, variations in canopy water content (CWC) that can be detected from microwave remotely sensed vegetation optical depth (VOD) have been proposed as a promising measure of vegetation water status, and we first reported that the regulation of CWC on productivity stability is universally applicable for global forests. Results of structural equation model (SEM) also confirmed the significant negative effect of CWC on coefficient of variation (CV) of productivity, indicating that the decrease in CWC could inevitably induce the instability of forest productivity under climate change. The most significant decrease (p < 0.01) of CWC is observed primarily in evergreen broadleaf forest in the tropics, implying an increasing instability of the most important carbon sink in terrestrial ecosystem.

Nitrogen (N) is an important nutrient for crop growth. However, the overuse of N fertilizers has led to a series of devastating global environmental issues. Recent studies show that multiple datasets have been created for agricultural N fertilizer application with varied temporal or spatial resolutions, nevertheless, how to synchronize and use these datasets becomes problematic due to the inconsistent temporal coverages, spatial resolutions, and crop-specific allocations. Here we reconstructed a comprehensive dataset for crop-specific N fertilization at 5-arc-min resolution (~10 km by 10 km) during 1961–2020, including N application rate, types, and placements. The N fertilization data was segmented by 21 crop groups, 13 fertilizer types, and 2 fertilization placements. Comparison analysis showed that our dataset is aligned with previous estimates. Our spatiotemporal N fertilization dataset could be used for the land surface models to quantify the effects of agricultural N fertilization practices on food security, climate change, and environmental sustainability.

Changes in land use intensity and types can affect the structure and function of ecosystems, and thus ecosystem services (ESs) as well as their interactions. However, the impacts of changes in land use intensity on ESs remain poorly understood. Through four different land use scenarios, we distinguished the independent contribution of changes in agricultural land use intensity and types to grain production (GP), water purification (WP), and their trade-offs in the Dongting Lake Basin. The results showed that from 1990 to 2015, GP increased across 58.07% of the total area, but WP decreased across 64.81% of the study area. The two ESs simultaneously increased or decreased across 41.93% of the total area. Watersheds covering 48.72% of the study area where GP increased and WP decreased were mainly distributed in areas with increased land use intensity. The other regions where GP decreased and WP increased were mainly distributed in areas with decreased land use intensity. The scenario analysis of GP, WP, and their trade-offs showed that the areas where agricultural land use intensity was the dominant factor were as large as 1.95 times, 2.38 times, and 2.43 times those dominated by land use type respectively, under the same climate conditions. This study highlighted the importance of changes in agricultural land use intensity on ES, which provided further supporting to ES-based land use management.

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.

Cropland ammonia (NH3) emission is a critical driver triggering haze pollution. Many agricultural policies were enforced in past four decades to improve nitrogen (N) use efficiency while maintaining crop yield. Inadvertent reductions of NH3 emissions, which may be induced by such policies, are not well evaluated. Here, we quantify the China's cropland‐NH3 emission change from 1980 to 2050 and its response to policy interventions, using a data‐driven model and a survey‐based dataset of the fertilization scheme. Cropland‐NH3 emission in China doubled from 1.93 to 4.02 Tg NH3‐N in period 1980–1996, and then decreased to 3.50 Tg NH3‐N in 2017. The prevalence of four agricultural policies may avoid ~3.0 Tg NH3‐N in 2017, mainly located in highly fertilized areas. Optimization of fertilizer management and food consumption could mitigate three‐quarters of NH3 emission in 2050 and lower NH3 emission intensity (emission divided by crop production) close to the European Union and the United States. Our findings provide an evidence on the decoupling of cropland‐NH3 from crop production in China and suggest the need to achieve cropland‐NH3 mitigation while sustaining crop yields in other developing economies.

Mitigating soil nitrous oxide (N 2 O) emissions is essential for staying below a 2 °C warming threshold. However, accurate assessments of mitigation potential are limited by uncertainty and variability in direct emission factors (EFs). To assess where and why EFs differ, we created high-resolution maps of crop-specific EFs based on 1,507 georeferenced field observations. Here, using a data-driven approach, we show that EFs vary by two orders of magnitude over space. At global and regional scales, such variation is primarily driven by climatic and edaphic factors rather than the well-recognized management practices. Combining spatially explicit EFs with N surplus information, we conclude that global mitigation potential without compromising crop production is 30% (95% confidence interval, 17–53%) of direct soil emissions of N 2 O, equivalent to the entire direct soil emissions of China and the United States combined. Two-thirds (65%) of the mitigation potential could be achieved on one-fifth of the global harvested area, mainly located in humid subtropical climates and across gleysols and acrisols. These findings highlight the value of a targeted policy approach on global hotspots that could deliver large N 2 O mitigation as well as environmental and food co-benefits.