粮食安全问题始终是关乎国计民生与社会政治稳定的重大战略问题，而保障水资源可持续利用是实现粮食安全的重要前提。西辽河流域位于我国北方农牧交错带，属于半干旱流域，生态环境较为脆弱，区域生态安全面临严峻挑战，粮食生产与水资源保护之间的矛盾十分突出。目前亟需深入开展相关研究，在确保区域生态安全的前提下，促进流域“水—粮”关系的协调发展。生态系统服务（Ecosystem services, ESs）可为生态安全视角下的流域“水—粮”关系解析与优化调控研究提供新的视角和决策基础。

本文以西辽河流域为研究对象，通过野外调查、遥感监测和模型模拟等多种方法，评估了2000–2020年流域净初级生产力、产水量、水质净化、粮食生产、土壤水蚀、土壤风蚀、生境质量等7项关键ESs，其中产水量、水质净化和土壤水蚀属于与水资源管理相关的生态系统服务（The water-management-related ecosystem services, WMESs，本文中简称为“水服务”），与粮食生产共同用于解析流域“水—粮”关系，除粮食生产外的6项关键ESs均用于流域综合生态安全格局构建。在此基础上，利用约束线、耦合协调度模型等方法深入揭示了流域“水—粮”关系的演变规律与驱动机制，基于 “ESs—生态安全格局（Ecological security patterns, ESPs）—区域可持续发展目标（Regional sustainable development goals, RSDGs）”框架构建了面向碳中和、水安全、土壤安全、生物多样性保护4项RSDGs的综合生态安全格局，通过集成CMIP6情景和土地利用情景探讨了未来不同情景综合生态安全格局约束下的流域“水—粮”关系优化调控策略，在此基础上提出了未来流域景观格局优化配置方案。本研究取得的主要结论有：

（1）流域“水—粮”之间存在空间权衡和约束关系，且耦合协调度有待改善。研究区WMESs的冷热点分布存在显著的空间差异，流域“水—粮”之间存在显著的空间权衡；流域“水—粮”之间表现出显著的约束效应，随着WMESs的增加，粮食生产整体呈现先升高后降低的特征；西北部分区域的“水—粮”关系处于中度失调状态，流域耦合协调度有待进一步改善。

（2）景观格局优化有助于实现流域“水—粮”关系的有效调控，应重点加强耕地和草地的协同管理。流域景观格局演变对“水—粮”关系有显著影响，耕地和草地的分布显著影响WMESs和粮食生产，在景观格局优化配置过程中应重点关注耕地和草地的协同管理而非仅调整其面积，通过减少景观破碎化，提高“水—粮”关系的生态—生产效益。

（3）构建面向多项RSDGs的综合生态安全格局有助于促进生态系统服务及其协同关系。相比于常规视角（多项ESs直接叠加）或面向单一区域可持续发展目标的生态源地，综合视角下生态源地各项生态系统服务较高、且协同关系更强。通过构建面向多项RSDGs的综合生态安全格局，能够促进多项生态系统服务的整体协同，有助于实现多项区域可持续发展目标的优化管理。

（4）研究区综合生态安全格局整体呈现“四区三带多分支”的空间分布特征。“四区”指的是分别位于研究区西北、西南、中部和东部的四个生态源地聚集区，“三带”指分别连接西北—西南地区、西北—东部地区和西南—东部地区的主体生态廊道，“多分支”指的是主体生态廊道的多个廊道分支，集中分布在研究区的中部地区。

（5）综合生态安全格局与“水—粮”关系优化调控之间存在显著的空间关联和权衡。研究区大量生态源地分布在“水—粮”失调区域或基本协调区域周边，许多生态廊道贯穿“水—粮”失调区域或基本协调区域。若在不考虑综合生态安全格局的情况下对流域“水—粮”关系进行优化调控可能危及区域生态安全，应将综合生态安全格局纳入未来情景模拟，探索生态安全视角下的流域“水—粮”关系优化调控。

（6）综合生态安全格局约束下的景观格局配置有助于实现粮食生产改善和WMESs保护的综合效益最大化。单一情景下提高粮食生产与保护WMESs之间均处于低协同、高权衡状态，但通过土地利用情景组合筛选的A_S3B_S2C_S1情景在改善粮食生产和保护WMESs之间表现出显著的高协同性和低权衡性，提供粮食生产和WMESs的综合能力最强。此时A类子流域采取S3情景（有综合生态安全格局约束、中等排放背景下的自然发展情景），B类子流域采取S2情景（有综合生态安全格局约束、高排放背景下的经济发展情景），C类子流域采取S1情景（有综合生态安全格局约束、低排放背景下的生态保护情景）。

Food security has always been a major strategic issue related to national economy, people's livelihood and social and political stability, and ensuring the sustainable use of water resources is an important prerequisite for achieving food security. The West Liao River basin (WLRB) is located in the agro-pastoral ecotone in the north of China. It is a semi-arid basin with fragile ecological environment and severe challenges to regional ecological security. The contradiction between food production (FP) and water resources protection is very prominent. At present, there is an urgent need to conduct in-depth research to promote the coordinated development of the "water-food" relationship while ensuring regional ecological security. Ecosystem services (ESs) can provide a new perspective and decision-making basis for analyzing and optimizing the "water-food" relationship in watersheds from the perspective of ecological security.

This study took the WLRB as the research object and evaluated seven key ESs from 2000 to 2020, including net primary production, water yield (WY), water purification (WP), FP, soil loss by water (SW), soil loss by wind and habitat quality, through various methods such as field investigation, remote sensing monitoring, and model simulation. Among them, WY, WP and SW belong to the water-management-related ecosystem services (WMESs), which were used together with FP to analyze the "water-food" relationship in the basin. The six key ESs were all used to construct the comprehensive ecological security pattern (ESP) of the basin. On this basis, the evolution laws and driving mechanisms of the "water-food" relationship in the basin were deeply revealed using methods such as constraint lines and coupling coordination models. Based on the "ESs-ESPs-regional sustainable development goals (RSDGs)" framework, the comprehensive ESP of four RSDGs, namely carbon neutrality, water security, soil security, and biodiversity protection, were constructed. By integrating CMIP6 and land use scenarios, this paper explored the optimization and regulation strategies of the "water-food" relationship in the basin under the constraints of comprehensive ESP in different scenarios in the future. Based on this, a plan for optimizing the configuration of the future watershed landscape pattern was proposed. The main conclusions obtained from this study were:

(1) There are spatial tradeoffs and constraints between "water-food" in the basin, and the coupling coordination degree needs to be improved. There are significant spatial differences in the distribution of hot and cold spots in WMESs, and there are significant spatial tradeoffs in the relationship of "water-food" in the basin. There was a significant constraint effect in the relationship of "water-food", with the increase of WMESs, FP first increased and then decreased. And the "water-food" relationship in some areas of northwest China is in a moderate imbalance state, and the coupling coordination degree of the basin needs to be further improved.

(2) The optimization of landscape pattern is conducive to the effective control of the "water-food" relationship in the basin, and the synergistic management of cropland and grassland should be strengthened. The evolution of landscape pattern in a basin has a significant impact on the "water-food" relationship, and the distribution of cropland and grassland significantly affects WMESs and FP. In the process of landscape pattern optimization, emphasis should be paid to the pattern of cropland and grassland, but this does not mean that only adjusting the area of cropland and grassland is effective, but the collaborative management of cropland and grassland should be promoted to reduce landscape fragmentation, and to improve the eco-production benefits of the "water-food" relationship.

(3) Building an integrated ESP for multiple RSDGs is conducive to promoting ESs and their synergistic relationships. Compared with the conventional perspective (multiple ESs directly superimposed) or the ecological sources oriented to the single RSDG, the ESs of the ecological sources under the comprehensive perspective are higher and the synergistic relationship is stronger. By building an integrated ecological security pattern for multiple RSDGs, the overall synergy of multiple ESs can be promoted and the optimal management of multiple RSDGs can be achieved.

(4) The comprehensive ESP of the study area presents the spatial distribution characteristics of "four zones, three zones and multiple branches". "Four zones" refers to the four ecological sources gathering areas located in the northwest, southwest, central and eastern parts of the study area respectively; "three zones" refers to the main ecological corridors connecting the northwest - southwest, northwest - east and southwest - east regions respectively; "multi-branch" refers to the multiple corridor branches of the main ecological corridor, which are concentrated in the central part of the study area.

(5) There is a significant spatial correlation and tradeoff between ESP and the optimal regulation of "water-food" relationship. In the study area, a large number of ecological sources are distributed around the "water-food" disordered area or the basic coordination area, and many ecological corridors run through the "water-food" disordered area or the basic coordination area. If the optimal regulation of the "water-food" relationship in the basin without considering the ESPs may endanger the regional ecological security, the ESPs should be included in the future scenario simulation to explore the optimal regulation of the "water-food" relationship in the basin from the perspective of ecological security.

(6) Landscape pattern allocation under the constraints of comprehensive ESP helps to maximize the comprehensive benefits of FP improvement and WMESs protection. In a single scenario, there was low synergy and high tradeoff between FP improvement and WMESs protection in the study area, but the A_S3B_S2C_S1 scenario screened by land use scenario combination showed significantly high synergy and low tradeoff between FP improvement and WMESs protection, and had the strongest integrated ability to provide FP and WMESs protection. At this time, subwatersheds of class A adopts S3 scenario (a natural development scenario with a medium emission background and under the constraints of a comprehensive ESP), subwatersheds of class B adopts S2 scenario (an economic development scenario with a high emission background and under the constraints of a comprehensive ESP), and subwatersheds of class C adopts S1 scenario (an ecological protection scenario with a low emission background and under the constraints of a comprehensive ESP).