机制检验与情景推演Mechanism testing and scenario reasoning
区域地球系统模式负责表达陆海气动力过程、生地化循环、生态系统响应和社会经济压力。它把气候变化、岸线工程、陆源输入、潮汐交换和生态修复转化为可计算边界条件，进而检验不同驱动如何改变临界距离和恢复能力。Regional Earth system models represent land-ocean-atmosphere dynamics, biogeochemical cycles, ecosystem responses, and socioeconomic pressure. They turn climate change, shoreline engineering, watershed input, tidal exchange, and restoration into computable boundary conditions, testing how different drivers alter threshold distance and recovery capacity.
临界诊断模块关注状态偏移、临界慢化、方差增大、自相关增强、空间同步和级联传播。空间智能模型把这些判据嵌入网格化诊断，使低氧、赤潮、湿地退化、风暴潮暴露等风险能够在同一空间框架中比较。The tipping-diagnosis module tracks state shifts, critical slowing down, rising variance, stronger autocorrelation, spatial synchrony, and cascading propagation. Spatial intelligence embeds these criteria into gridded diagnosis so that hypoxia, HABs, wetland degradation, and storm-surge exposure can be compared in one spatial frame.
数字孪生系统连接实时数据、区域模式和情景库。可视化界面呈现状态、风险和情景结果，后台计算环境同步处理观测更新、模型同化、风险诊断和适应路径。对于重点区域，数字孪生版本同时表达过程状态、风险受体和可调控变量。The digital twin connects real-time data, regional models, and scenario libraries. Its interface presents system state, risk, and scenario results, while the computational environment handles observation updates, model assimilation, risk diagnosis, and adaptive pathways. For the focus regions, each version represents process state, risk receptors, and controllable variables together.
模式链条从物理过程开始，向生地化循环、生态响应和社会经济暴露延伸。区域地球系统模式刻画水动力与物质通量，AI 代理模型提高情景筛选效率，数字孪生把观测更新和管理情景连接起来。三者共同形成可迭代的计算体系。The model chain starts with physical processes and extends to biogeochemical cycling, ecosystem response, and socioeconomic exposure. Regional Earth system models represent hydrodynamics and material fluxes, AI surrogates accelerate scenario screening, and the digital twin links observation updates with management scenarios. Together they form an iterative computational system.
版本记录输入数据、边界条件、参数化方案、同化方法、验证指标、适用尺度和已知限制。跨团队比较和迭代从这些可复现实验信息开始，也由这些信息判断一个模式版本适合解释机制、预测趋势还是支持情景比较。Each version records input data, boundary conditions, parameterization, assimilation methods, validation metrics, scales, and known limitations. Cross-team comparison and iteration begin with these reproducible experiment details, which also indicate whether a version is suited to mechanism explanation, trajectory prediction, or scenario comparison.
版本信息Version information
- 版本号、发布时间、覆盖区域、核心模块和依赖数据Version, release date, region, core modules, and dependent data.
- 临界指标、验证样本、情景集合、运行环境和代码状态Tipping metrics, validation samples, scenario ensemble, runtime environment, and code status.
- 关联论文、数据产品、可视化图件和技术指南Related papers, data products, visual maps, and technical guides.
重点区域中的计算对象Computational objects in focus regions
环渤海版本强调水交换、岸线工程和底层耗氧；长三角版本强调羽状流、层化和低氧阈值；海峡西岸版本强调湾口链、沿岸流和赤潮窗口；粤港澳大湾区版本强调珠江口低氧、风暴潮和基础设施暴露。四个版本共享同一诊断逻辑，但参数、边界条件和风险受体各不相同。The Bohai Rim version emphasizes exchange, shoreline engineering, and bottom oxygen demand; the Yangtze Delta version emphasizes plume dynamics, stratification, and hypoxia thresholds; the Western Taiwan Strait version emphasizes bay chains, coastal currents, and HAB windows; the Greater Bay Area version emphasizes Pearl River hypoxia, storm surge, and infrastructure exposure. The four versions share one diagnostic logic while using different parameters, boundary conditions, and risk receptors.