Potential and Limitations of Mountain Wetland Detection using Satellite Imagery and Digital Elevation Models

Mountain wetlands play an important role in climate change adaptation and in reconstructing historical climate and geomorphic changes. Despite their ecological significance, wetlands in South Korea are underrepresented in national wetland inventories. This underrepresentation is largely attributed to the challenging accessibility of mountainous terrains, which complicates field surveys, alongside a general lack of awareness about their significance. This study aimed to identify the potential distribution of mountain wetlands within national parks through a combination of geospatial analysis and field validation. Five spatial variables-slope, curvature, topographic position index (TPI), normalized difference vegetation index (NDVI), and normalized difference water index (NDWI)-were extracted from digital elevation models and Sentinel-2 imagery. These variables were utilized to identify potential wetlands based on the attributes of known wetlands. A total of 522 candidate sites meeting minimum area criteria were identified. From these, 61 locations were prioritized for field surveys considering accessibility, size, and vegetation types. Field investigations confirmed 23 previously undocumented mountain wetlands. In the mountainous terrain, swamps with woody vegetation were more common than herbaceous marshes. These swamps are often small and visually resemble surrounding forests, making them challenging to detect solely through satellite imagery. While the GIS-based approach proved effective for initial screening, limitations remained due to factors such as vegetation cover, scale, and seasonal variability.