以地理模拟与优化系统(GeoSOS)模型进行的升金湖湿地景观模拟与演变预测  

Landscape Simulation and Evolution Prediction of Shengjin Lake Wetland Using Geographic Simulation and Optimization System (GeoSOS) Model

作　　者：程倩[1] 王杰 马靖博 王梦茹 周旭[1] Cheng Qian;Wang Jie;Ma Jingbo;Wang Mengru;Zhou Xu(Anhui University,Hefei 230601,P.R.China)

机构地区：[1]安徽大学,合肥230601

出　　处：《东北林业大学学报》2025年第1期79-85,共7页Journal of Northeast Forestry University

基　　金：国家自然科学基金项目(32171573);安徽省自然科学基金项目(2308085MD114);国家级大学生创新创业训练计划项目(202310357151)。

摘　　要：以升金湖国际重要湿地为研究区,根据2010—2020年Landsat系列遥感影像,分析该湿地景观变化特征及驱动因素,利用地理模拟与优化系统(GeoSOS)模型对2025年该湿地在不同发展情景(自然情景、耕地保护情景和生态保护情景)进行景观模拟与预测,探究该湿地景观的未来演变趋势,并提出相关规划建议。研究表明:(1)2010—2020年间,草地面积略微减少,滩涂和水体面积有部分增加,景观结构趋于复杂化,主要受到年降水(自然因素)及GDP(人文因素)的影响;(2)GeoSOS模型中,人工神经网络-元胞自动机(ANN-CA)算法的模拟精度最高,总体精度达到63.28%,Kappa系数达到0.503,品质因数(FoM)达到0.122,适用于湖泊湿地景观模拟与预测;(3)根据2025年不同发展情景湿地景观预测结果,采取生态保护策略最能有效保障自然湿地面积,维持湿地景观的多样性与稳定性。Using the Shengjin Lake International Important Wetland as the study area, this research analyzes the wetland landscape change characteristics and driving factors based on Landsat series remote sensing images from^(2)010 to 2020. The Geographic Simulation and Optimization System (GeoSOS) model is employed to simulate and predict the wetland landscape in 2025 under different development scenarios (natural scenario, cultivated land protection scenario, and ecological protection scenario), exploring the future evolution trends of the wetland landscape and providing related planning suggestions. The study findings indicate: (1) Between 2010 and 2020, the area of grassland slightly decreased, while the area of shoals and water bodies experienced some increase, leading to increased landscape complexity. This change is primarily influenced by annual precipitation (natural factors) and GDP (human factors). (2) In the GeoSOS model, the Artificial Neural Network-Cellular Automata (ANN-CA) algorithm demonstrated the highest simulation accuracy, with an overall accuracy of 63.28%, a Kappa coefficient of 0.503, and a Figure of Merit (FoM) of 0.122, making it suitable for simulating and predicting lake wetland landscapes. (3) According to the wetland landscape prediction results for 2025 under different development scenarios, implementing ecological protection strategies is the most effective way to safeguard the area of natural wetlands, maintaining the diversity and stability of the wetland landscape.

关 键 词：湿地景观 景观模拟 地理模拟与优化系统模型 地理探测器 发展情景 

分 类 号：P237[天文地球—摄影测量与遥感]