The ability to quantify the processes driving geomorphic change in river valley margins is vital to geomorphologists seeking to understand the relative role of transport mechanisms (e.g. fluvial, aeolian, and hillslope processes) in landscape dynamics. High-resolution, repeat topographic data are becoming readily available to geomorphologists. By contrasting digital elevation models derived from repeat surveys, the transport processes driving topographic changes can be inferred, a method termed ‘mechanistic segregation.’ Unfortunately, mechanistic segregation largely relies on subjective and time consuming manual classification, which has implications both for its reproducibility and the practical scale of its application. Here we present a novel computational workflow for the mechanistic segregation of geomorphic transport processes in geospatial datasets. We apply the workflow to seven sites along the Colorado River in the Grand Canyon, where geomorphic transport is driven by a diverse suite of mechanisms. The workflow performs well when compared to field observations, with an overall predictive accuracy of 84% across 113 validation points. The approach most accurately predicts changes due to fluvial processes (100% accuracy) and aeolian processes (96%), with reduced accuracy in predictions of alluvial and colluvial processes (64% and 73%, respectively). Our workflow is designed to be applicable to a diversity of river systems and will likely provide a rapid and objective understanding of the processes driving geomorphic change at the reach and network scales. We anticipate that such an understanding will allow insight into the response of geomorphic transport processes to external forcings, such as shifts in climate, land use, or river regulation, with implications for process-based river management and restoration.