Lake-edge wetland types result from the influence of various hydrogeomorphological (HGM) predictors, such as water level, hydroperiod, and elevation. Climate and land-use changes could induce variations in lake water levels, and subsequently trigger changes in wetland type and area. This study aims to determine whether HGM predictors alone can be used to classify lake-edge wetlands. The resilience of these wetlands to different water level change scenarios was also evaluated. We sampled the plant communities of 37 lake-edge wetlands (peatland, ash and alder swamps) of a medium-size Canadian lake (12.9 km²), and computed 12 HGM predictors using a GIS software and commonly available landscape data. We used canonical redundancy analysis to evaluate the relationship between wetland types, plant communities, and HGM predictors. We then built a random forest model using these predictors to classify wetlands and to evaluate the impacts of small (± 0.5 m) and large (± 2 m) water level changes on wetland type and area. Our results show that elevation and slope were the two main HGM predictors of wetland plant communities and types. The random forest model was robust (89 % accuracy), but performed less well for swamps, likely because alder and ash swamps are similar systems, but at different successional stages. Finally, while small water level changes had minimal impacts on lake-edge wetland type and area (−0.6 %), larger changes triggered substantial gains (+ 13 %) or losses (−20 %) of tree-dominated wetlands. Considering that swamps harbour high species richness, landscape management should be carried out bearing in mind their greater vulnerability.

湖边湿地类型受各种水文地貌 (HGM) 预测因素的影响，例如水位、水周期和高程。气候和土地利用的变化会引起湖水水位的变化，进而引发湿地类型和面积的变化。本研究旨在确定是否可以单独使用 HGM 预测因子对湖边湿地进行分类。还评估了这些湿地对不同水位变化情景的恢复能力。我们对一个中等规模的加拿大湖泊 (12.9 km ² ) 的 37 个湖边湿地（泥炭地、灰烬和桤木沼泽）的植物群落进行了采样。)，并使用 GIS 软件和常用景观数据计算了 12 个 HGM 预测变量。我们使用规范冗余分析来评估湿地类型、植物群落和 HGM 预测因子之间的关系。然后，我们使用这些预测因子构建了一个随机森林模型，以对湿地进行分类，并评估小 (± 0.5 m) 和大 (± 2 m) 水位变化对湿地类型和面积的影响。我们的结果表明，海拔和坡度是湿地植物群落和类型的两个主要 HGM 预测因子。随机森林模型稳健（准确率为 89%），但在沼泽中表现较差，可能是因为桤木和灰沼泽是相似的系统，但处于不同的演替阶段。最后，虽然小的水位变化对湖边湿地类型和面积的影响很小 (-0.6 %)，更大的变化引发了以树木为主的湿地的显着收益 (+ 13 %) 或损失 (-20 %)。考虑到沼泽具有较高的物种丰富度，应在进行景观管理时考虑到它们的更大脆弱性。