Intensive pesticide use can enhance crop yields while posing risks to non-target organisms across ecological compartments. A watershed-scale, process-based framework that dynamically links crop-specific applications of multiple pesticides (including transformation products) with their fate in soil and water is still lacking. This study introduces the MARINA-Pesticides model, that simulates monthly, sub-basin-scale transport of 30 pesticides and three transformation products from 12 crops to river networks, while accounting for their degradation, partitioning, and transport processes in both croplands and rivers. We apply the MARINA-Pesticides model to the Three Gorges Reservoir Area (TGRA) to quantify pesticide residue concentrations in soil and river and to assess associated ecological risks. In 2020, an estimated 903 tonnes of pesticides were applied to croplands across the TGRA, with 3% exported to aquatic systems, 10% remaining in soils and 87% degraded. Among 30 pesticides, chlorpyrifos, imidacloprid, and carbendazim posed ecological risks in the soil. In total, 11.2 tonnes of parent pesticides and 2 tonnes of three transformation products were exported into rivers, predominantly via surface runoff (91%), with the remainder via soil erosion (9%). Riverine pesticide concentrations peaked during the summer season. In summer, chlorpyrifos, imidacloprid, and fenpropathrin posed a high risk to riverine ecosystems. Our conclusions advance the understanding of the multi-pesticide fate and timing in diverse agricultural watersheds, supporting guidance for soil and water-resource protection.

Recycling organic phosphorus (P) sources may reduce dependence on synthetic P fertilizers in sustainable agriculture. However, the overall lack of understanding of the impact of organic P management practices on crop yields and the environment limits further optimization of organic P strategies. Herein, we first executed a global meta-analysis of over 860 paired observations to validate potential indicators for delineating risk zones for organic P fertilizer management. Then, a combination of machine learning tools and global datasets was used to further identify the risk zones for organic P management practices. Finally, an optimal organic P management strategy was developed considering the total P inputs, types of organic P sources, organic P proportions, and organic P management risk zones. Results indicated that the P activation coefficient (PAC) can be used as a potential indicator for delineating risk zones for organic P management. Hypothesis-oriented path analysis suggests that P inputs under low-soil-PAC conditions drive the preferential allocation of P to the occluded and moderately labile P pools and that organic P fertilizer application can positively affect labile P. P inputs under high-soil-PAC conditions have a more balanced effect on various P fractions of soil, and P inputs through organic fertilizers are primarily stored in the organic P pool. Current organic P management practices do not benefit food production in case of low-soil-PAC cropland and may increase the risk of P runoff in case of high-soil-PAC cropland. A combination of optimal organic P management with risk zones achieved a 13.3 % reduction in global P runoff and a 10.7 % increment in global food production compared with the use of synthetic P fertilizers alone. Our study provides a solution for enhancing the efficient use of organic P resources to create more productive, clean, and sustainable food production systems.

Soil acidification models are useful for evaluating measures to mitigate soil acidification under various agronomic practices. However, the appropriate modeling approaches for simulating the soil acidification process have not been adequately studied across soils with distinct buffering mechanisms. This study evaluated the performance differences between a process-based soil acidification model (VSD+) and four machine learning models, including random forest (RF), support vector machine, extreme gradient boosting and decision tree, in simulating pH dynamics of neutral and acidic soils. Two longterm experimental sites were selected with distinct buffering mechanisms on purple soil as an example for the development, calibration and validation of soil acidification models. Results from the RF importance factor analysis indicated that soil background pH was the primary factor influencing the dynamic changes in purple soil pH, followed by meteorological conditions and agronomic practices. pH was then chosen as an essential input variable to developing machine learning models for simulating soil acidification patterns. Machine learning models achieved higher accuracy in neutral soil than the VSD+ model. The RF model gave the best simulation performance, outperforming other machine learning models at both sites, with the highest R-2 of 0.70 and 0.47 and the lowest MAE of 0.19 and 0.17 for neutral and acidic soils, respectively. In contrast, the VSD+ model exhibited excellent accuracy with acidic soil (R-2 = 0.95, RMSE = 0.05 and MAE = 0.02) compared to the other machine learning models (R-2 = 0.20-0.47, RMSE = 0.15-0.23 and MAE = 0.14-0.20). These findings provide information for selecting the most suitable modeling approach to simulate soil acidification process with distinct buffering mechanisms, supporting informed decision-making for restoring soil health and quality. (c) The Author(s) 2025. Published by Higher Education Press. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0)

Rising atmospheric CO₂ enhances terrestrial productivity, but its capacity to sustain ecosystem functioning under intensifying climate extremes remains uncertain. Using the process-based LPJmL4 model, we simulated net primary production (NPP) for a historical period (1990–2019) and a late-century period (2071–2100) under SSP1-2.6 and SSP5-8.5, with dry and wet conditions classified by the SPEI. Counterfactual simulations with fixed and transient CO2 were conducted to isolate the CO2 effect from the climate effect. Results show that hydroclimatic extremes shift from broadly wetter conditions under SSP1-2.6 to co-occurring dry and wet extremes under SSP5-8.5, with stronger drying upstream and enhanced wetting in the middle and lower basin. Although long-term mean NPP increases by 17–44% under future scenarios, the buffering of CO2 during extremes is minimal. Relative NPP changes during dry and wet conditions stay within ±1% when CO2 rises. This contrast reveals a decoupling between mean-state greening and ecosystem resistance to extremes. Mechanistically, elevated CO₂ shifts ecosystem regulation from water limitation toward energy limitation, strengthening light-mediated responses while increasing transpiration demand. This higher transpiration highlights the water cost of greening and potential risks to regional water availability under intensified warming. These findings show that enhanced productivity in long-term average can mask vulnerability to climate extremes and may lead to overestimation of carbon-sink resilience in humid monsoon regions.

The widespread use of antibiotics in humans and animals raises significant environmental concerns. However, few approaches can simultaneously quantify their transfer from humans and animals and track their fate in soils and rivers. In this study, we developed the MARINA-Antibiotics model (Model to Assess River Inputs of pollutaNts to seAs for Antibiotics) to quantify the sources and concentrations of 30 widely used antibiotics, as well as assess their associated environmental risks, and implemented this model in the Three Gorges Reservoir Area in 2020. The risk of antimicrobial resistance was evaluated by calculating the risk quotient based on the minimum selective concentration. Our findings revealed that 11 tons of antibiotics entered cropland via manure application. High and medium risk quotients indicated potential risks from ciprofloxacin, enrofloxacin, sulfapyridine, tetracycline, doxycycline, and ofloxacin in cropland. In total, 13 tons were discharged into rivers, primarily through point sources (99%). In rivers, degradation and sedimentation accounted for 90% of antibiotic removal. Tetracyclines and fluoroquinolones contributed the highest loads to the sediments. The risk of promoting AMR was low for most antibiotics except for ciprofloxacin and enrofloxacin. This study improves our understanding of antibiotic sources and spatial patterns in watershed environments, providing valuable insights for clean water management.

Most existed crop modelling studies are mainly cereal crops. Vegetables, the most economical and nutrient-dense crops, recieves insufficient attention, particularly on nutrient-uptake predictions. In open-field vegetable systems with shallower roots, shorter lifespan, and higher nutrient requirements, it is even more challenge to minize water pollution from fertilizers. To ensure both food and environment security, there is an urgent need of precise vegetable models to optimize productivity against fertilizer usage.We adapted the WOrld FOod STudies (WOFOST) crop growth simulation model for chili pepper (Capsicum annuum L.) and Chinese cabbage (Brassica rapa L.) to support better fertilizer management under various climate and soil conditions. We conducted field experiments with six various fertilizer strategies (etc., mixed synthetic and organic fertilizers, denitrification products, and slow-control-release fertilizers) in southwestern China from 2019 to 2021. In total about 20 parameters relevant to physiological development, dry matter accumulation, photosynthesis, and nutrient uptake were measured and used in model adaptation.Our study shows that it is possible to model chili pepper’s growth without changing much from the WOFOST-generic model structure. We provide solutions by adapting user-defined developmental stages to mimic the growth from transplanting to fruiting and subsequently ripeness. As for WOFOST-Chinese cabbage, we further modify the phenological module to mimic the special vernalization habits of Chinese cabbage. Additionally, we design a new data re-analyzation method for accurate biomass partitioning predictions. Overall, both WOFOST-Chili and WOFOST-Chinese cabbage models show good model performance on biomass assimilation (rRMSE = 0.23/0.17 for chili/cabbage leaf dry weight; rRMSE = 0.06/0.17 for chili/cabbage storage organ dry weight) and nutrient uptake (rRMSE = 0.46/0.29 for chili/cabbage leaf N amount; rRMSE = 0.12/0.41 for chili/cabbage storage organ N amount). Besides, an improved leaf area index (LAI) simulation is found in WOFOST-Chinese cabbage (rRMSE = 0.11) than WOFOST-Chili (rRMSE = 0.76).These findings improve our understanding of yield-nutrient interactions within crop models, provide insights on expanding application of original-designed-for-field crop models to different vegetable versions, also call for a refined dynamic nutrient simulation flow within soil module to evaluate mitigation effect of expanded fertilizer strategies under climate change.

CONTEXT: Chinese cabbage (Brassica rapa L. ssp. Pekinensis) is a leading open-field leafy vegetable crop in China, with known shallow roots, short lifespan and high nutritional content. To ensure its precise field management in a changing world, an accurate leafy vegetable model is urgently needed to enhance decision-making. OBJECTIVE: We modified the WOrld FOod STudies (WOFOST) crop model to create a version specifically for Chinese cabbage (WOFOST-Chinese cabbage) that quantifies its daily dynamic growth based on fertilizer management, climate, and soil conditions. METHODS: We conducted field experiments in southwestern China from 2019 to 2021 to collect extensive site-specific soil data and specific crop data relevant to eco-physiological processes of Chinese cabbage. We used 2021 field trail observations under optimal growing conditions to parameterize and calibrate the model with integrated image data in a stepwise procedure. The datasets from 2019 and 2020 were used for model validation, while the no-fertilizer dataset was used to test model performance under nutrient-limited conditions. RESULTS AND CONCLUSIONS: Overall, the developed WOFOST-Chinese cabbage is reliable in simulating biomass (rRMSE = 0.13 for total aboveground production), leaf area index (rRMSE = 0.34), and nutrient uptake (rRMSE = 0.15). Additionally, model robustness is increased by the sensitivity analysis of biomass-relevant and nutrient-uptake-relevant parameters. However, this good performance is constrained by a less effective differentiation between functional and non-functional leaves. Besides, model validation provides further improvement directions of model structure under nutrient-limited conditions. SIGNIFICANCE: This study provides the first set of comprehensively calibrated parameters for applying WOFOST to Chinese cabbage in open cropland. The outcome of this study will aid in fertilizer management for open-field Chinese cabbage production across diverse climate and soil conditions. Additionally, it will contribute a new model for comparative multi-model studies of leafy vegetables, addressing climate change impacts at regional, national, and global scales.

It is widely believed that forest plays a vital role in mitigating and adapting climate changes. By adopting afforestation and reforestation, the atmospheric carbon can be captured and stored (sequestered) within plants and soil. Thus, tree expansions are witnessed over the whole world. However, large-scale forestation often requires a high amount of water resources, which may bring up pressures on the local water cycle. In this case, trade-offs between carbon sequestration and ecohydrology remain under discussion yet.China has launched several forestation projects since the last century. These projects spread over the country in both water-sufficient and water-limited areas. However, the changing climate conditions may alter local ecosystems and influence the survival of newly grown vegetations and the potential of carbon sequestration. Therefore, understanding the influences of climate change on carbon sequestration and hydrological response is of importance in national forestation management.In this study, we selected a Dynamic Global Vegetation Model (DGVM), the Lund–Potsdam–Jena Managed Land (LPJmL4), to study climatic influences on the carbon cycle and water cycle in the Yangtze River Basin, China. The model was implemented on a regional scale with increased resolution from 30mins to 5mins. Meanwhile, considering the complex terrain of the Yangtze River Basin, we classified the basin with a climatic zonation scheme to furtherly analyse the influence of climate.Our initial findings indicated that climate effects dominated the variations in the carbon and water cycle. Meanwhile, the future focus of ecological restoration, including forestation and protection, might need to shift from subtropical regions like the Yunnan-Guizhou plateau to western temperate alpine regions. This work can serve as recommendations for guiding national ecological restoration management.

Urban landscapes are high phosphorus (P) consumption areas and consequently generate substantial P-containing urban solid waste (domestic kitchen wastes, animal bones, and municipal sludge), due to large population. However, urbanization can also trap P through cultivated land loss and urban solid waste disposal. Trapped urban P is an overlooked and inaccessible P stock. Here, we studied how urbanization contributes to trapped urban P and how it affects the P cycle. We take China as a case study. Our results showed that China generated a total of 13 (+/- 0.9) Tg urban trapped P between 1992-2019. This amounts to 6 (+/- 0.5) % of the total consumed P and 9 (+/- 0.6) % of the chemical fertilizer P used in China over that period. The loss of cultivated land accounted for 15% of the trapped urban P, and half of this was concentrated in three provinces: Shandong, Henan, and Hebei. This is primarily since nearly one-third of the newly expanded urban areas are located within these provinces. The remaining 85% of trapped urban P was associated with urban solid waste disposal. Our findings call for more actions to preserve fertile cultivated land and promote P recovery from urban solid waste through sound waste classification and recycling systems to minimize P trapped in urban areas.

The rapid increase in the proportion of cash crops and livestock production in the Yangtze River Basin has led to commensurate increases in fertilizer and pesticide inputs. Excessive application of chemical fertilizer, organophosphorus pesticides and inappropriate disposal of agricultural waste induced water pollution and potentially threaten Agriculture Green Development（AGD）. To ensure food security and the food supply capacity of the Yangtze River Basin, it is important to balance green and development, while ensuring the quality of water bodies. Multiple pollutants affect the transfer, adsorption, photolysis and degradation of each other throughout the soil-plant-water system. This paper considers the impact of multi-pollutants on the nitrogen and phosphorus cycles especially for crops, which are related to achieving food security and AGD. It presents prospective on theory, modeling and multi-pollutant control in the Yangtze River Basin for AGD that are of potential value for other developing regions.

Soil contamination, particularly from pesticide residues, presents a significant challenge to the sustainable development of agricultural ecosystems. Identifying the key factors influencing soil pesticide residue risk and implementing effective measures to mitigate their risks at the source are essential. Here, we collected soil samples and conducted a comprehensive survey among local farmers in the Three Gorges Reserve Area, a major agricultural production region in Southwest China. Subsequently, employing a dual analytical approach combining structural equation modeling (SEM) and random forest modeling (RFM), we examined the effects of various factors on pesticide residue accumulation in vegetable ecosystems. Our SEM analysis revealed that soil characteristics (path coefficient 0.85) and cultivation factor (path coefficient 0.84) had the most significant effect on pesticide residue risk, while the farmer factors indirectly influenced pesticide residues by impacting both cultivation factors and soil characteristics. Further exploration using RFM identified the three most influential factors contributing to pesticide residue risk as cation exchange capacity (CEC) (account for 18.84%), cultivation area (account for 14.12%), and clay content (account for 13.01%). Based on these findings, we carried out experimental trials utilizing Integrated Pest Management (IPM) technology, resulting in a significant reduction in soil pesticide residues and notable improvements in crop yields. Therefore, it is recommended that governmental efforts should prioritize enhanced training for vegetable farmers, promotion of eco-friendly plant protection methods, and regulation of agricultural environments to ensure sustainable development.

The upstream cascade dams play an essential role in the nutrient cycle in the Yangtze. However, there is little quantitative information on the effects of upstream damming on nutrient retention in the Three Gorges Reservoir (TGR) in China. Here, we aim to assess the impact of increasing cascade dams in the upstream area of the Yangtze on Dissolved Inorganic Nitrogen and Phosphorus (DIN and DIP) inputs to the TGR and their retention in the TGR and to draw lessons for other large reservoirs. We implemented the Model to Assess River Inputs of Nutrients to seAs (MARINA-Nutrients China-2.0 model). We ran the model with the baseline scenario in which river damming was at the level of 2009 (low) and alternative scenarios with increased damming. Our scenarios differed in nutrient management. Our results indicated that total water storage capacity increased by 98 % in the Yangtze upstream from 2009 to 2022, with 17 new large river dams (>0.5 km(3)) constructed upstream of the Yangtze. As a result of these new dams, the total DIN inputs to the TGR decreased by 15 % (from 768 Gg year(-1) to 651 Gg year(-1)) and DIP inputs decreased by 25 % (from 70 Gg year(-1) to 53 Gg year(-1)). Meanwhile, the molar DIN:DIP ratio in inputs to the TGR increased by 13 % between 2009 and 2022. In the future, DIN and DIP inputs to the TGR are projected to decrease further, while the molar DIN:DIP ratio will increase. The Upper Stem contributed 39 %-50 % of DIN inputs and 63 %-84 % of DIP inputs to the TGR in the past and future. Our results deepen our knowledge of nutrient loadings in mainstream dams caused by increasing cascade dams. More research is needed to understand better the impact of increased nutrient ratios due to dam construction.

BACKGROUNDClimate change significantly impacts global maize production via yield reduction, posing a threat to global food security. Disease-related crop damage reduces quality and yield and results in economic losses. However, the occurrence of diseases caused by climate change, and thus crop yield loss, has not been given much attention.RESULTSThis study aims to investigate the potential impact of six major diseases on maize yield loss over the next 20 to 80 years under climate change. To this end, the Maximum Entropy model was implemented, based on Coupled Model Intercomparison Project 6 data. The results indicated that temperature and precipitation are identified as primary limiting factors for disease onset. Southern corn rust was projected to be the most severe disease in the future; with a few of the combined occurrence of all the selected diseases covered in this study were predicted to progressively worsen over time. Yield losses caused by diseases varied per continent, with North America facing the highest loss, followed by Asia, South America, Europe, Africa, and Oceania.CONCLUSIONThis study provides a basis for regional projections and global control of maize diseases under future climate conditions. (c) 2024 Society of Chemical Industry. Analysis of the occurrence of maize diseases and their resulting yield losses revealed that temperature and precipitation were identified as the main factors influencing disease development, and the combined occurrence of selected diseases is expected to progressively worsen over time. Additionally, yield losses due to diseases varied from continent to continent. This study provides a basis for regional projection and global management of maize diseases under future climate conditions. image

Returning organic nutrient sources (for example, straw and manure) to rice fields is inevitable for coupling crop-livestock production. However, an accurate estimate of net carbon (C) emissions and strategies to mitigate the abundant methane (CH4) emission from rice fields supplied with organic sources remain unclear. Here, using machine learning and a global dataset, we scaled the field findings up to worldwide rice fields to reconcile rice yields and net C emissions. An optimal organic nitrogen (N) management was developed considering total N input, type of organic N source and organic N proportion. A combination of optimal organic N management with intermittent flooding achieved a 21% reduction in net global warming potential and a 9% rise in global rice production compared with the business-as-usual scenario. Our study provides a solution for recycling organic N sources towards a more productive, carbon-neutral and sustainable rice-livestock production system on a global scale.

CONTEXT: Chili pepper (Capsicum annuum L.) is one of the most economically and agriculturally important, and relatively nutrient-dense, vegetables that has, to date, received little attention in model studies relevant to dry matter production and nutrient-uptake predictions. There is an urgent need for models to analyse the potential impacts of climate change, as well as responsive adaptation options, while simultaneously optimising productivity against fertilizer use to reduce nutrient pollution.OBJECTIVE: We adapted the WOrld FOod STudies (WOFOST) crop growth simulation model for chili pepper (WOFOST-Chili) to quantify dry matter production as a function of fertilizer management, climate, and soil conditions.METHODS: We used 2021 field trial data under optimal growing conditions in southwestern China to parameterise and calibrate WOFOST-Chili. The model was tested under no-fertilizer conditions and further validated with data from 2019 and 2020. In addition, a sensitivity analysis over the three consecutive years was performed. RESULTS AND CONCLUSIONS: Overall, the developed WOFOST-Chili model shows good simulations of chili growth dynamics in response to nitrogen (N) fertilization, both on biomass assimilation (rRMSE = 0.07 for total aboveground production; rRMSE = 0.06 for fruit dry weight) and nutrient uptake (rRMSE = 0.46 for leaf N amount; rRMSE = 0.29 for fruit N amount). Additionally, model robustness is increased by the sensitivity analysis of crop initialisation (e.g., biomass and leaf area index at transplanting) and climate-dependent parameters (e.g., temperature sums determining development rate and light use efficiency determining productivity), with the resulting wider simulation range covering more observations. This good performance is only limited by a weaker leaf area index (LAI) simulation (rRMSE = 0.76), which is partially attributed measurement limitations (e.g., equipment, weather conditions and labour/time constraints). Model validation confirms good performance under potential conditions, which is slightly reduced under nutrient-limited conditions.SIGNIFICANCE: These findings improve our understanding of yield-nutrient interactions of chili pepper. They provide insight on expanding the application of crop models originally designed for cereals to non-Gramineae vegetables, while calling for future improvement of model accuracy under different fertilizer application strategies.

Balancing ecology and human development has been a long and wide concern. The upper Yangtze River Basin (UYRB) of China has implemented large important ecological restoration projects since the last century. These restoration practices have changed land use patterns within the UYRB, consequently impacting the local carbon cycle. The most noteworthy project is the Grain for Green Program, which returns cropland to natural vegetation (forest and grassland). Yet the effects of restoration on land use change, carbon sequestration, and associated food production remain unclear. This study utilized remote sensing data and conversion coefficients to analyze the ecological-policy-induced land use changes of the UYRB from 2000 to 2020 and their impacts on terrestrial carbon sequestration. Linear regression, machine learning, and structural equation modeling (SEM) were utilized to evaluate the correlations between environmental and socio-economic factors and the distribution of carbon stocks. The results indicated positive effects of ecological activities on the UYRB, despite decreases in cropland. Over the past 20 years, the UYRB had sequestered carbon by a total amount of 1796 ± 926 Mt C. The spatial distribution of sequestered carbon demonstrated a strong correlation with slopes, followed by temperatures. The SEM results indicated that agricultural production and carbon sequestration were enhanced synergically under land use changes. This further demonstrated the effectiveness of these land policies in achieving a balance between crop productivity and ecology protection. We emphasized the importance of vegetation restoration in achieving carbon neutrality and the necessity to continue these projects. We suggested a more reasonable land management for the future UYRB based on the characteristics of each geographical subregion. This work serves as an example of effective land management to other locations worldwide perusing the harmony of ecological restoration and human development.

Separation of crop and livestock production increases the risk of environmental pollution and the wastage of nutrient resources derived from crop-livestock systems. Integration of crop and livestock production is an important pathway for promoting nutrient cycling and reducing nutrient losses. Research on nutrient management at the basin scale can upscale agricultural production technologies from the farm scale to the basin scale and improve nutrient-use efficiency. Based on production optimization, the environmental threshold of the basin can be used as a bayonet to further reduce environmental nutrient losses. In addition, it is important to achieve higher nutrient efficiency and greater environmental emission reduction of crop-livestock production systems in a large area via nutrient management at the basin scale, which may also support the green development of agriculture. Taking the Yangtze River Basin as an example, this study reviewed the significance of nutrient management based on the integration of crop and livestock production at the basin scale with green development, nutrient management technologies based on the integration of crop and livestock production, and spatial optimization based on the environmental cost of the crop-livestock production system. In addition, the present study focuses on the nutrient management of crop-livestock production systems at the basin scale. Based on the present review, we found that there are a series of nutrient management technologies for crop-livestock systems in the Yangtze River Basin, and the promotion and application of these technologies via a bottom-up approach could further reduce nutrient losses and improve agricultural production efficiency. However, the nutrient losses of the crop-livestock system in some areas are too high and cannot be controlled within the environmental threshold only through technical improvement; it is also necessary to conduct spatial planning for crop-livestock systems via a top-down approach. Future studies on nutrient management of crop-livestock systems at the basin scale should include (1) characteristics and driving factors of nutrient flow and environmental emissions at the basin scale, (2) classification of vulnerable areas of nutrient losses at the basin scale, and (3) evaluation and optimization of crop-livestock systems based on vulnerable areas.

Eliminating both overt and hidden hunger is at the core of the global food and nutrition security agenda. Yet, the collective state of nutrition security at the population level is not known. Here we quantify food-based availability of 11 essential nutrients for 156 countries using a food production-consumption-nutrition model, followed by assessment of the nutrient availability status as a ratio of recommended intake. For the baseline year 2017, global per capita availability was adequate for calorie and protein but in severe deficit for vitamin A and calcium (intake ratios, <0.60, where 1.0 is adequate) and moderate deficit for vitamin B12 (intake ratio, 0.76). At the country level, more than half of the 156 countries were in various degrees of deficit for all nine micronutrients. Disparities across regions or countries were enormous. We explore intervention strategies from an agriculture-food system perspective and discuss the daunting challenges of addressing nutrition security broadly.

Riverine antibiotics have gained significant attention due to their ecological and human health impacts. Here, we assess the distribution of antibiotic concentrations in rivers globally and identify the driving forces behind antibiotic pollution. We have collated a data set of antibiotics detected in rivers globally, encompassing 21 distinct antibiotics from 287 unique rivers, distributed across 128 sub-basins. We also collated data sets of antibiotics in manure and wastewater treatment plants (WWTPs) and analyzed the relationships between riverine antibiotics and anthropogenic indicators. Sulfamethoxazole (SMX) and trimethoprim (TMP) exhibited the highest average concentrations in studied rivers and WWTPs, particularly in Africa. Conversely, tetracyclines (TCs) were most frequently detected in manure. TCs in rivers are strongly correlated to animal production, particularly with cattle and pig farming. In Africa, humans' widespread use of antibiotics, combined with insufficient wastewater treatment, were correlated to high concentrations of sulfonamides(SAs), fluoroquinolones(FQs), and Macrolides (MLs) antibiotics in rivers. In Asia, the application of extensive antibiotics in livestock possibly elevated riverine TCs. Our results provide new insights into the global action required to achieve SDG6 on clean water and sanitation, emphasizing the need for enhanced wastewater treatment and controlled antibiotic use.

The development of phosphate(P)fertilizer industry is closely related to China's food security,P resource utilization,and environmental protection.Nowadays,soil available P rapidly increased due to the large amount chemical P fertilizer application in the past 40 years,especially in the surface soil.It is of great significance to predict and regulate the future demand for P fertilizer in China based on the action of green development of agriculture.The integrated analyses show that in this paper,a large proportion of soil for cereal crops production in China has reached the crop agronomic P threshold(15～25 mg·kg-1),which means above this threshold there is no yield increase response with P application.In particular,the soil available P had considerably exceeded the threshold value for cash crops such as vegetables and fruit trees.Soil available P should be maintained at the agronomic threshold based on the requirement of crop root/rhizosphere biological capacity,nutritional quality and environmental risks.Consequently,it is critical to implement"crop agronomic threshold"oriented P fertilizer management system.Meanwhile,the agricultural green development needs to maximize the recycling and reuse efficiency of P in agricultural waste,which should focus on the recycling process and agronomic utilization.Accordingly,considering the demand for food and other agricultural products in China,this paper re-predicted the future demand for P fertilizer,following the prediction of China's P fertilizer demand in 2007 based on changes in soil P fertility.The consumption of chemical P fertilizer in China will be 10.84 million tons and 7.42 million tons by 2030 and 2050,respectively.Therefore,based on the continued and multiple optimized measures,the overall demand for P fertilizer in China in the short term by 2030 could reduce about 1.5 million tons,in the long term by 2050 could be adjusted to 7.5 million tons per year,more than 30%reduction from the current consumption.