The Grain for Green Program (GGP) is an effective way to control soil erosion and improve the eco-environment in China. How to develop the most cost-effective GGP scheme, which can balance the contradiction among ecology, economy, and food security, is the key point for the sustainable development of the GGP. In this study, Xixian watershed located in the upper reaches of the Huaihe River is taken as the study area. Based on the current land use situation, the distributed hydrological model SWAT (Soil and Water Assessment Tool) was used to simulate the runoff and sediment processes. Then sediment reduction resultants per unit area of GGP, which defined as the sediment reduction coefficients, were obtained by simulating the successive GGP operations in each sub-basin based on the validated SWAT model. Meanwhile, considering the spatial correspondence between GDP (gross domestic product) and land use type, GDP loss coefficients at sub-basin scale were obtained by overlapping the GDP map and current land use map. On this basis, the ecological benefit and economic benefit of the GGP operation was expressed by sediment reduction coefficients and GDP loss coefficients respectively. Finally, the multi-objective genetic algorithm NSGA-II was used to optimize the GGP scheme at sub-basin scale. The results showed that 1) The SWAT model performed high simulation accuracy for runoff and sediment modeling. The Nash–Sutcliffe coefficients were above 0.90 and 0.70, the deterministic coefficients were both greater than 0.80, and the percentage deviation of the total amount is controlled within −20% to 20%, respectively. It can be conclude that the SWAT model can be used to evaluate the impact of the GGP on sediment reduction. 2) The sediment reduction coefficients ranged from 26.70 to 2 675.85 t/km², decreasing gradually from the upper reaches to the lower reaches, which indicated that implementation of GGP per unit area can reduce sediment more effectively in the upstream river source area. 3) The GDP loss coefficient presented spatial differences significantly with the range from −5 756.83 yuan/km² to 136.26 yuan/km², showing that both increased GDP (i.e GDP loss coefficient values were greater than 0) and decreased GDP (i.e GDP loss coefficient values were less than 0) could be observed among sub-basins. Notably, sub-basins where the values of the GDP loss coefficient appeared to be the smallest were mainly concentrated in the main residential areas of cities and towns. That is, the GGP in these sub-basins would prove more costly. 4) The GGP schemes obtained by multi-objective optimization maintained the per capita cultivated land area between 1.04×10⁻³ and 1.54×10⁻³ km², which was significantly higher than the warning level of food security. Meanwhile, the Pareto-optimal set was able to reduce sediment yield by 53.54% to 69.86% of the initial value and still achieve the sustainable soil erosion level, while only losing 30.13% to 37.67% of the initial economic output. The GGP optimization method proposed in this study based on ecological sediment reduction benefits and economic benefits can provide reference and guidance for scientific planning GGP and other soil and water conservation measures.

With the rapid advancement of agricultural intensification, nitrogen fertilizers have become indispensable inputs for sustaining crop productivity worldwide. However, the excessive and often inefficient application of these fertilizers has resulted in significant nitrogen surpluses, contributing to severe non-point source (NPS) pollution, which poses substantial threats to water quality and ecosystem health. This issue is particularly acute in ecologically sensitive regions such as the Three Gorges Reservoir Region in China, where agricultural activities are intensive and environmental vulnerabilities are high. Against this backdrop, this study aims to assess the spatial risk of nitrogen fertilizer-induced non-point source pollution across the TGRR by applying the “source-sink” theory within a geospatial modeling framework. Methodologically, the study introduces an innovative approach to improve the identification of source landscapes. This was achieved by integrating two critical factors: established buffer zones along the Yangtze River mainstream and the spatial intensity of nitrogen fertilizer application. Source landscapes were subsequently classified into five distinct grades (Grade I to V) based on their pollution potential. To evaluate the mobilization and transport risk of nitrogen pollutants, a comprehensive resistance surface was constructed using key natural factors including elevation, slope, rainfall erosivity, land use type, and soil erodibility. The Minimum Cumulative Resistance (MCR) model was then employed to quantify the diffusion resistance faced by nutrients from each source grade, leading to the categorization of the entire region into five corresponding risk levels: extremely high, high, medium, low, and extremely low. The results revealed several key findings: 1) The composition of source landscapes exhibited clear dominance by the mid-to-high grades. Specifically, Grade IV source landscapes were the most extensive, covering 31.78% of the total study area. This indicates that areas with relatively high pollution potential are widespread. In contrast, the extremes were less common; Grade V landscapes covered the smallest area at 10.90%, while Grade I areas were also limited in extent. 2) A distinct spatial clustering of source grades was observed, closely tied to geography and economic activity. Grade I, II, and III source landscapes were predominantly located in areas distal from the main channel of the Yangtze River, often coinciding with mountainous terrain and economically less developed regions where agricultural intensity is lower. Conversely, Grade IV and V source landscapes were highly concentrated in zones immediately adjacent to the Yangtze River mainstream (particularly within 0-40 km buffers) and in other flat valley areas. These zones are characterized by frequent, intensive agricultural activities, higher population density, and greater fertilizer input, creating a high pollution potential near critical water bodies. 3) Analysis at the administrative county level revealed pronounced spatial heterogeneity in overall pollution risk. Badong County exhibited the highest mean MCR value (23,999), signifying that the cumulative resistance to pollutant movement from source areas is greatest here, thus indicating the lowest overall pollution risk among all counties. On the other end of the spectrum, Yubei District registered the lowest mean MCR value (3,058), reflecting minimal landscape resistance and consequently the highest overall pollution risk. 4) Geographically, extremely high and high-risk zones were mainly clustered in the middle and upper reaches of the Yangtze River within the TGRR, aligning with agriculturally intensive areas. Conversely, low and extremely low-risk zones were primarily situated near the administrative boundary between Hubei Province and Chongqing Municipality. 5) Based on this risk stratification, tailored ecological management strategies are proposed. For high and extremely high-risk regions (e.g., Chongqing core urban area, Wanzhou, Beibei), immediate and prioritized interventions are crucial. These should focus on implementing nitrogen fertilizer reduction and efficiency enhancement technologies and strengthening the protection, restoration, and strategic planning of riparian buffer zones to intercept runoff. For low and extremely low-risk areas (e.g., Wushan, Zigui, Badong), the management priority should shift towards preventive conservation. This involves establishing long-term environmental monitoring networks and robust early-warning systems to detect any negative trends in pollution levels, thereby preserving their current favorable environmental status. Overall, this study provides a spatially explicit decision-support framework for the precise prevention and control of nitrogen non-point source pollution in the TGRR, offering valuable insights for sustainable agricultural practices and watershed management in similar regions globally.

Non-point source pollution can often be caused by nitrogen loss from fertilization in the water system. It is very necessary to identify and evaluate the environmental risks of nitrogen application for water pollutant prevention. In this study, an improved model was proposed to minimize the cumulative resistance, according to the "source-sink" theory in landscape ecology. The watershed was also selected above Hongze Lake of the Huaihe River. The source risk of nitrogen non-point source pollution was evaluated by the fertilization environmental risk index when the amount of nitrogen was applied. The key resistance factors were selected as the elevation, slope, land use, terrain moisture index, rainfall erosivity, and soil erodibility. The migration of non-point source pollution was used to construct a comprehensive resistance base. An innovative approach was proposed to construct the sink risk of nitrogen non-point source pollution using flow concentration routing. Finally, the comprehensive risk index of nitrogen non-point source pollution was formed to combine the source and sink risk. The risk level of non-point source pollution was also classified in the study area. The results show the following. 1) The source risk value of nitrogen non-point source pollution was 0-0.81, with an average value of 0.55. There was a widespread situation of excessive fertilization, especially in Shangqiu, Zhoukou, Zhumadian, and Xinyang City within Henan Province. 2) The resistance base shared a spatial trend of gradually decreasing from the southwest to the northeast under various resistance factors. A decreasing trend of sink risk was observed around the main stream. Moreover, the sink risk in the north part of the main stream was significantly lower than that in the southern region, due to the longer flow routing. 3) The comprehensive risk index demonstrated that 71.16% of the entire study area was above the middle-risk level. There was serious nitrogen non-point source pollution in the study area. 4) Extremely high-risk areas were mainly concentrated in the upper reaches of the main stream, such as Xinyang City and the northern part of Zhumadian City. There were large areas of high-risk areas in Zhoukou, Shangqiu, Fuyang, Suzhou, and Bozhou cities, which were located in the north part of the main stream. Low-risk areas were distributed mainly in the mountainous areas of the headwaters of the main and tributary rivers. Specific prevention and control measures were proposed, according to the comprehensive risk of nitrogen non-point source pollution at different levels. The finding can also provide the decision-making basis for the scientific prevention and effective management of agricultural non-point source pollution at the watershed scale.