Identifying geomorphic domains using hierarchically clustered drainage area‐slope scaling

Abstract

AbstractArea‐slope scaling is an important tool for analysing geomorphic processes. In this study, we used hierarchical clustering to sort the area‐slope scaling data for both the numerical landscape created using the LandLab model and real landscapes. Random forest analysis was used to explore the effect of hierarchical clustering. We aimed to illustrate the heterogeneous features of area‐slope clustering and their influencing factors and to test whether they can be used to distinguish geomorphic domains. Our results showed that hierarchical clustering can easily detect the heterogeneous features of area‐slope scaling constrained by tectonic uplift and fluvial erodibility but not those constrained by the diffusion coefficient. The slope gradients over the inflection/transition zone of the colluvial and fluvial domains play a key role in identifying area‐slope scaling from different geomorphic domains. Case studies show that area‐slope scaling of clusters would be affected by different lithology and distance away from the trunk channel. In the high erosion circumstance, the fluvial process caused by tectonics leads to increasing slope gradients over whole geomorphic domains, as well as to retreating inflection between the colluvial and fluvial process domains. In low erosion circumstances or steady state, the fluvial process caused by nearby rivers cannot effectively change the slope gradients but can retreat the inflection. The real landscapes suggest that basins should be carefully selected for area‐slope scaling analysis to ensure that they are representative of the entire study area. Before geomorphological research, more basins of a specific research region need to be selected to conduct hierarchical clustering to select representative basins. Our study could be helpful in the identification of area‐slope domains, the extraction of geomorphic metrics and geomorphic process research.