湿地是城市生态系统的关键组成部分，具有极高的生态、经济与社会价值，也是区域实现可持续发展的重要战略资源。伴随着城市化进程的加快以及人类活动的不断干扰，湿地生态环境面临巨大威胁。为了提高人们对湿地重要性的认知，推动人与湿地的和谐共存，《湿地公约》在全球范围内展开“湿地城市”认证。然而，湿地城市称号有效期仅有6年时间，只有持续推进湿地保护恢复工作，达成湿地城市认证标准，才能继续拥有这张“世界级绿色名片”。因而及时掌握湿地未来可能的变化趋势，对于湿地城市下一步保护与恢复工作的实施具有重大意义。现阶段在湿地及其生态系统服务的多情景模拟、可持续性评估、湿地变化模式的研究与对比分析上仍有不足和缺失。基于此，本研究拟解决三个关键问题，即湿地城市中湿地及其生态系统服务在未来多情景下的时空分布如何？ 湿地城市中湿地可持续性的时空分布如何？不同时段内湿地城市的湿地变化模式如何？

本研究以常熟市、海口市、东营市、银川市、哈尔滨市和常德市6个中国首批湿地城市为研究对象，基于2015-2022年8期Sentinel遥感分类数据，选择适用于湿地城市的模拟模型，构建趋势延续、经济建设、湿地保护与和谐发展4种情景，开展2023-2035年6个时期9种湿地精细类型与6种非湿地类型的时空分布模拟，评估多情景下湿地城市的水源涵养、水质净化、碳储量、洪水调蓄和气候调节5种湿地生态系统服务，综合识别2015-2035年间湿地及其生态系统服务演变过程中的可持续区与不可持续区，最后总结2015-2020年现状阶段、2020-2025年后评估阶段、2015-2030年可持续发展目标（SDGs）实施阶段和2020-2035年远景目标阶段中6个湿地城市的湿地变化模式。主要内容与结论如下：

（1）选择适用于湿地城市变化模拟的空间模型，开展多情景下湿地与非湿地的长时序变化模拟。基于2015-2022年8期Sentinel遥感分类产品，开展了FLUS、PLUS和CLUE-S模型在湿地模拟中的精度对比。构建趋势延续、经济建设、湿地保护与和谐发展4种情景，建立多个最优目标函数，耦合马尔科夫（Markov）与多目标规划（MOP）面积预测模型与适宜空间模拟模型，开展2023-2035年6个湿地城市9种湿地精细类型与6种非湿地类型的多情景变化模拟，并进行长时序的空间变化分析。使用Dynamic World数据集进行情景验证，探究与城市实际发展最贴合的情景。结果显示：①2015-2022年模型精度验证表明，FLUS、PLUS和CLUE-S模型在6个湿地城市的湿地与非湿地模拟中总体精度均值超过了0.85，Kappa均值超过了0.80。其中，FLUS模型的精度评估结果略高于其他模型，且更加稳定，因此选择FLUS模型开展后续研究。②2023-2035年间，湿地保护情景下湿地面积的增加最为显著，到2035年该情景下常熟市、海口市、东营市、银川市、哈尔滨市和常德市的湿地面积占比分别为24.73%、13.73%、29.75%、6.32%、9.16%和10.69%，分别较经济建设情景多出了3.42%、4.91%、10.96%、3.41%、5.03%和5.18%。和谐发展情景下大多数湿地呈现出了平稳恢复的趋势，但经济建设情景下湿地损失十分严重。河流、湖泊、沼泽和养殖池是湿地中的主要恢复类型，尤其集中在大型湖泊、河流与水库周围。③引入2023年Dynamic World数据集进行情景验证发现，和谐发展情景的综合精度评估结果最优，与城市的实际发展趋势更加贴合。

（2）开展不同情景下湿地城市5种湿地生态系统服务评估与长时序空间分布变化分析。基于2015-2035年遥感分类与多情景模拟数据以及气候情景数据，采用InVEST模型与洪水调蓄模型，评估水源涵养、水质净化、气候调节、碳储量和洪水调蓄5种湿地生态系统服务，得到6个湿地城市8个时期5种服务的时空分布。结果显示：①2015-2020年，常熟市、常德市和海口市的湿地生态系统服务上升明显，但银川市5个服务均有不同程度的下降。②2020-2035年，经济建设情景下6个城市的5种湿地生态系统服务均呈现下降趋势，最显著的是水源涵养服务，常熟市、海口市、东营市、银川市、哈尔滨市和常德市在15年间分别减少了30.81%、18.81%、34.62%、93.02%、58.10%和61.45%。而在其他情景中，水质净化和水源涵养是下降最明显的服务，碳储量和洪水调蓄服务则以恢复为主。③从不同城市来看，常熟市、常德市和银川市在4种情景下均有3个及以上的湿地生态系统服务出现下降趋势，尤其是常德市中仅洪水调蓄服务有小幅上升。

（3）评估多情景下湿地的可持续性，识别湿地变化中可持续区与不可持续区的时空分布。基于2015-2035年遥感分类与模型模拟的湿地变化结果以及湿地生态系统服务的评估结果，构建湿地变化与湿地生态系统服务变化的可持续性准则体系，识别2015-2020年现状时期以及2020-2035年远景目标时期中湿地变化的可持续区与不可持续区占比与分布。结果显示：①2015-2020年，常熟市、海口市、东营市、银川市、哈尔滨市和常德市湿地的可持续区占比分别为86.02%、93.50%、88.15%、86.94%、84.98%和90.12%，不可持续区由湿地生态系统服务退化区构成，其中水质净化服务下降最为显著。②2020-2035年，湿地保护情景下可持续区占比是最高的，常熟市、海口市、东营市、银川市、哈尔滨市和常德市分别达到了97.98%、90.86%、89.70%、95.00%、95.96%和91.91%。而经济建设情景下湿地的不可持续区占比达到最大，6个城市分别为10.16%、26.15%、41.67%、20.83%、29.91%和13.19%。③在所有情景中，和谐发展情景是最稳定的情景，可持续区占比虽略小于湿地保护情景，但其不可持续区占比非常低。因此，若按照和谐发展情景进行城市规划，确保湿地面积不减少且有些许增加，反而可能是最优情景。

（4）对比分析不同时段下6个湿地城市的湿地变化模式，提出湿地保护修复的相关建议。将2015-2035年划分为2015-2020年现状阶段、2020-2025年后评估阶段、2015-2030年SDGs阶段和2020-2035年远景目标阶段，总结6个湿地城市在不同阶段下的湿地变化模式，并提出湿地保护修复建议。结果显示：①2015-2020年现状阶段中，常熟市和东营市分别呈现出平原湿地和滨海湿地的恢复模式。海口市和银川市湿地面积略有减少，分别呈现出滨海湿地和山区湿地的受损模式。而哈尔滨市和常德市因湿地生态系统服务下降，分别呈现出平原湿地和山区湿地的退化模式。②2020-2025年后评估阶段中，通过湿地率和湿地保护率判定各个城市是否还满足湿地城市认证标准，结果发现除经济建设情景外，大部分城市在趋势延续、湿地保护与和谐发展情景下均为达标模式，仅有海口市的趋势延续情景和银川市的趋势延续与和谐发展情景无法达标。③2015-2030年SDGs阶段中，银川市在4种情景下均呈现弱可持续的变化模式，海口市、哈尔滨市和常德市仅在湿地保护情景下存在强可持续的变化模式，而常熟市和东营市在大多数情景下均呈现出强可持续的变化模式。④2020-2035年远景目标阶段中，湿地保护情景中除海口市为轻微恢复变化模式外，其他城市均呈现出明显恢复的变化模式。而经济建设情景下除常熟市为轻微受损变化模式外，其他城市均为严重受损或严重退化模式。另外两种情景中，常熟市和常德市以轻微恢复变化模式为主，其他城市均出现了轻微退化或轻微受损的变化模式。⑤近些年，在国际需求和国内政策的牵引下，6个湿地城市持续推进湿地保护修复工作，但未来仍需注意自然湿地的优先恢复、湿地生态系统服务提升以及保护区边界划定，推动一体化湿地智能监管体系建设，让更多高水平高质量的科研学术成果在湿地保护管理中得到有效应用。

本文以湿地城市为典型案例，开展未来多情景下湿地的变化模拟，以及湿地生态系统服务的未来评估，识别湿地在长时序演变过程中的可持续区和不可持续区，总结不同阶段下湿地城市的湿地变化模式。本研究对于6个首批湿地城市未来的湿地保护恢复与合理利用具有重要的现实意义，更为区域生态文明建设与可持续发展提供科学的数据支撑。同时，本研究的方法与思路对于第二批湿地城市与其他城市湿地的科学管理与空间优化具有一定的理论指导意义。

Wetlands are a key component of urban ecosystems, with high ecological, economic and social value, and are also an important strategic resource for achieving sustainable development in the region. Accompanied by the acceleration of urbanization and the continuous interference of human activities, the ecological environment of wetlands is facing great threats. In order to raise people's awareness of the importance of wetlands and promote the harmonious coexistence of human beings and wetlands, the Convention on Wetlands has launched the certification of "Wetland City" around the world. However, the title of Wetland City is only valid for six years, and only by continuously promoting wetland protection and restoration work, and meeting the criteria for Wetland City certification, can these cities continue to have this "world-class green card". Therefore, it is of great significance to grasp the possible future change trend of wetlands in time for the implementation of the next conservation and restoration work in wetland cities. At present, there are still shortcomings and deficiencies in the multi-scenario modelling of wetlands and their ecosystem services, sustainability assessment, research and comparative analysis of wetland change patterns. Based on this, this study aims to address three key scientific questions, namely, what is the spatial and temporal distribution of wetlands and their ecosystem services in wetland cities under multiple future scenarios? What is the spatial and temporal distribution of wetland sustainability in wetland cities? What are the wetland change patterns in wetland cities during different time periods?

In this study, six first batch wetland cities in China, namely Changshu, Haikou, Dongying, Yinchuan, Harbin and Changde, are taken as the research objects. Based on the Sentinel remote sensing classification data of eight periods from 2015 to 2022, the simulation model applicable to wetland cities is selected, and four scenarios, namely, trend continuation, economic construction, wetland protection and harmonious development are constructed to carry out the simulation of the temporal and spatial distribution of nine wetland fine types and six non-wetland types in the six periods from 2023 to 2035. Assess the five wetland ecosystem services of wetland cities under multiple scenarios: water conservation, water purification, carbon storage, flood regulation and climate regulation, comprehensively identify the sustainable and unsustainable areas in the evolution of wetlands and their ecosystem services between 2015 and 2035, and finally summarizes the wetland change patterns of the six wetland cities in 2015-2020 current status stage, 2020-2025 post-assessment stage, 2015-2030 SDGs stage and 2020-2035 visionary goal stage. The main contents and conclusions are as follows:

(1) Select a spatial model suitable for wetland city change simulation and carry out long time-series changes simulation of wetlands and non-wetland under multiple scenarios. Based on the eight Sentinel remote sensing classification products from 2015 to 2022, the accuracy comparison of FLUS, PLUS and CLUE-S in wetland simulation was carried out. Four scenarios of trend continuation, economic construction, wetland protection and harmonious development are constructed, multiple optimal objective functions are established, coupling Markov and MOP area prediction models with suitable spatial simulation model to conduct multi-scenario change simulations of nine wetland fine types and six non-wetland types in the six wetland cities from 2023 to 2035, and conduct long-term spatial and temporal change analyses. Use the Dynamic World dataset for scenario validation to explore the scenario that best fits the actual development of the cities. The results show that: ①The model accuracy validation from 2015 to 2022 indicates that the average OA of the FLUS, PLUS, and CLUE-S models in wetland and non-wetland simulations of the six wetland cities exceed 0.85, and the average Kappa exceed 0.80. The accuracy evaluation results of the FLUS model are slightly higher than other models and more stable, so the FLUS model is selected to carry out the follow-up study. ②From 2023 to 2035, the wetland area under wetland protection scenario increased most significantly, and by 2035, the proportion of wetland area in Changshu, Haikou, Dongying, Yinchuan, Harbin, and Changde cities under this scenario is 24.73%, 13.73%, 29.75%, 6.32%, 9.16%, and 10.69%, respectively, which is 3.42%, 4.91%, 10.96%, 3.41%, 5.03%, and 5.18% higher than economic construction scenario. Most of the wetlands under harmonious development scenario show a trend of stable recovery, but economic construction scenario show a very serious loss of wetlands. Rivers, lakes, marshes and aquaculture ponds are the main types of restoration in wetlands, especially concentrated around large lakes, rivers and reservoirs. ③Scenario validation by introducing the Dynamic World dataset in 2023 found that harmonious development scenario has the best comprehensive accuracy assessment results, which is more in line with the actual development trend of the cities.

(2) Assess five wetland ecosystem services in wetland cities under different scenarios and analyze long-term changes in spatial and temporal distribution. Based on the remote sensing classification and multi-scenario simulation data from 2015 to 2035, as well as climate scenario data, the five wetland ecosystem services of water conservation, water purification, carbon storage, flood regulation and climate regulation are assessed by using the InVEST model and the flood regulation model, and the spatial and temporal distributions of the five services in eight periods of the six wetland cities are obtained. The results show that: ①From 2015 to 2020, the wetland ecosystem services in Changshu, Changde and Haikou increased significantly, but all five services in Yinchuan decreased to different degrees. ②From 2020 to 2035, the five wetland ecosystem services in the six cities under economic construction scenario show a downward trend, with the most significant being the water conservation service, which decreased by 30.81%, 18.81%, 34.62%, 93.02%, 58.10% and 61.45% in Changshu, Haikou, Dongying, Yinchuan, Harbin and Changde, respectively over the 15-year period. In other scenarios, water purification and water conservation are the services with the most obvious decline, while carbon storage and flood regulation services are dominated by recovery. ③From the perspective of different cities, Changshu, Changde and Yinchuan all have three or more wetland ecosystem services that show a decreasing trend in the four scenarios, especially in Changde, where only the flood regulation service have slightly increased.

(3) Assess the sustainability of wetlands under multiple scenarios and identify the spatial and temporal distribution of sustainable and unsustainable areas in wetland changes. Based on the results of wetland changes from remote sensing classification and model simulation from 2015 to 2035, as well as the assessment results of wetland ecosystem services, a sustainability criteria system for wetland changes and wetland ecosystem services changes is constructed, and the proportion and distribution of sustainable and unsustainable areas in wetland changes are identified in the current period from 2015 to 2020 and in the visionary goal period from 2020 to 2035. The results showed that: ①From 2015 to 2020, the proportion of sustainable areas in the wetlands of Changshu, Haikou, Dongying, Yinchuan, Harbin and Changde are 86.02%, 93.50%, 88.15%, 86.94%, 84.98% and 90.12%, respectively, and the unsustainable areas are consisted of degraded areas of wetland ecosystem services, with the most significant decrease in water quality purification service. ②From 2020 to 2035, the proportion of sustainable areas under wetland protection scenario is the highest, reaching 97.98%, 90.86%, 89.70%, 95.00%, 95.96% and 91.91% in Changshu, Haikou, Dongying, Yinchuan, Harbin and Changde, respectively. And the unsustainable areas of wetlands in economic construction scenario reach the largest percentage, which was 10.16%, 26.15%, 41.67%, 20.83%, 29.91% and 13.19% in the six cities, respectively. ③Among all the scenarios, harmonious development scenario is the most stable one, and although the proportion of sustainable areas is slightly smaller than that of wetland protection scenario, its proportion of unsustainable areas is very low. Therefore, if urban planning is carried out according to harmonious development scenario, ensuring that the wetland area does not decrease or slightly increases, it may actually be the optimal scenario.

(4) Compare and analyze the wetland change patterns of the six wetland cities under different time periods and put forward relevant suggestions for wetland protection and restoration. Dividing 2015-2035 into 2015-2020 current status stage, 2020-2025 post-assessment stage, 2015-2030 SDGs stage and 2020-2035 visionary goal stage, the wetland change patterns of the six wetland cities under different stages are summarized, and wetland protection and restoration proposals are put forward. The results show that: ①In the current status stage from 2015 to 2020, Changshu and Dongying show the restoration patterns of plain wetlands and coastal wetlands, respectively. Haikou and Yinchuan, with a slight decrease in wetland area, show the damage patterns of coastal wetland and mountain wetland, respectively. However, Harbin and Changde show degradation patterns of plain wetland and mountain wetland respectively due to the decline of wetland ecosystem services. ②In the post-assessment stage from 2020 to 2025, the wetland rate and wetland protection rate are used to determine whether each city still meets the wetland city certification standards, and it is found that most cities are in the compliance pattern under trend continuation scenario, wetland protection and harmonious development scenarios, except for economic construction scenario. Only trend continuation scenario in Haikou and trend continuation and harmonious development scenario in Yinchuan are unable to meet the standards. ③In the SDGs stage from 2015 to 2030, Yinchuan show a weakly sustainable change pattern under all four scenarios, Haikou, Harbin and Changde have a strongly sustainable change pattern only under wetland protection scenario, while Changshu and Dongying show a strongly sustainable change pattern under most scenarios. ④In the visionary goal stage from 2020 to 2035, all cities in wetland protection scenario show a significantly restoration change pattern except Haikou, which show a slight restoration change pattern. In economic construction scenario, except for Changshu, which is a slight damage change pattern, all other cities were in a severely damage or severely degradation pattern. In the other two scenarios, Changshu and Changde mainly exhibite a slight restoration change pattern, while all other cities show a slight degradation or damage change pattern. ⑤In recent years, the six wetland cities have continued to promote wetland protection and restoration work under the guidance of international demands and domestic policies, but in the future, it is still necessary to pay attention to the priority restoration of natural wetlands, the enhancement of wetland ecosystem services, and the delineation of protected area boundaries, promote the construction of an integrated wetland intelligent monitoring system, and enable more high-level and high-quality scientific research and academic achievements to be effectively applied in wetland protection and management.

Taking wetland cities as typical cases, this paper carries out the simulation of wetland changes under multiple scenarios in the future, as well as the future assessment of wetland ecosystem services, identifies sustainable and unsustainable areas of wetlands in the long time-series of wetland evolution, and summarises the wetland change patterns of wetland cities at different stages. This study is of great practical significance for the future wetland protection, restoration and rational utilization in the six first batch wetland cities, and provides scientific data support for regional ecological civilisation construction and sustainable development. Meanwhile, the methods and ideas of this study are of theoretical significance for the scientific management and spatial optimisation of wetlands in the second batch of wetland cities and other cities.