Understanding the spatial–temporal dynamics of wetland land cover (LC) changes and their impacts on ecosystem services (ESs) is essential for wetland conservation and management. This study aimed at quantifying and predicting water provision and habitat quality ecosystem services and analyzing their response to LC changes using the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model. Firstly, LC maps were created for 1987, 1998, 2007, and 2017 using the index-based and supervised approach and were then used to predict LC changes for 2027 and 2047 using the cellular automata and Markov chain (CA–Markov) model. The Shadegan wetland, located in the southwest of Iran, was chosen as the case study. The results showed that, during the study period (1987–2017), the LC with major increasing area was bare land while the LC with major decreasing area was wetland vegetation. In addition, the wetland tended to be a fragmented mosaic with less connectivity. The results of ESs modeling showed that the highest and lowest water yield values, which were almost 68 × 10⁶ m³ and 26 × 10⁶ m³, were observed in 1998 and 2017, respectively. In the meantime, the area with high-quality habitats has reduced (by 33.58%) and has gradually been replaced by moderate- and low-quality habitats, and the same trend will continue in 2027 and 2047. Based on these results, we conclude that changes in wetland LCs have resulted in variations of the wetland ESs. This site-specific study of the InVEST water yield and habitat quality models enables us to explore the applicability of spatially explicit modeling and simulating of wetland ESs to improve wetland conservation and management. The findings on the spatial–temporal analysis of wetland LCs and ESs can promote wetlands decision-making processes and taking effective conservation actions.

了解湿地土地覆盖（LC）变化的时空动态及其对生态系统服务（ESs）的影响对于湿地保护和管理至关重要。这项研究旨在量化和预测水供应和栖息地质量的生态系统服务，并使用生态系统服务和折衷综合评估（InVEST）模型分析其对LC变化的响应。首先，使用基于索引的监督方法为1987、1998、2007和2017创建了LC图，然后使用元胞自动机和马尔可夫链（CA–Markov）模型将其用于预测2027和2047的LC变化。案例研究选择了位于伊朗西南部的Shadegan湿地。结果表明，在研究期间（1987-2017年），面积增加最大的LC为裸地，面积减少最大的LC为湿地植被。此外，湿地往往是零散的马赛克，连通性较低。ESs建模结果表明，最高和最低产水量值几乎为68×10⁶  m ³和26×10 ⁶  m ³分别在1998年和2017年被观察到。同时，拥有高质量栖息地的地区减少了（减少了33.58％），并逐渐被中，低质量栖息地所取代，并且这一趋势将在2027年和2047年继续。基于这些结果，我们得出结论湿地LC的变化导致了湿地ES的变化。这项针对InVEST产水量和栖息地质量模型的针对特定地点的研究使我们能够探索对湿地ES进行空间显式建模和模拟的适用性，以改善湿地的养护和管理。湿地LC和ES的时空分析发现可以促进湿地决策过程并采取有效的保护措施。