Understanding the hydrological valley landscape: a multi-scenario adaptive framework for delineating valley floors

As critical zones in fluvial geomorphology shaped by hydrological processes, valley floors play an essential role in material exchange and circulation between upland and groundwater bodies. Their accurate delineation is crucial for understanding river morphology, analyzing the distribution of valley floor sediments, and maintaining the riverine landscape ecosystem. However, current methods for delineating valley floors are highly artificial, region-specific and require subjective parameter selection. To address these limitations, we develop a multi-scenario adaptive framework for delineating valley floors. This framework designs several indicators for automatically detecting topographical cross-sectional and longitudinal features, providing a basis for parameter determination in valley floor extraction and achieving geomorphologically adaptive automatic extraction. The framework includes the following components: (1) The initial drainage network was extracted by setting drainage thresholds based on geomorphological texture features obtained using the gray-level co-occurrence matrix (GLCM); (2) The drainage network generated in the previous step was filtered by calculating the average river gradient and setting adaptive parameters, removing drainage networks located in steep valleys; (3) The valley floor extent was adaptively extracted by proposing terrain factors such as slope accumulation and its variation. The experimental results demonstrate that this method applies to the extraction of valley floors in various geomorphological types, exhibiting high precision. This study also explored the correlation between river valleys, geological sedimentation, and surface hydrological processes, finding a significant consistency between sediment distribution and valley floor extent. These findings provide a new perspective for research on geological mapping, and the evolutionary patterns of valley floor morphology.