Riparian vegetation composition and structure reflect the legacy of fine- and coarse-scale patterns of hydrology, geomorphology and ecology. Understanding the dynamics of complex systems requires scaling-up mechanistic patterns and processes. Coupling hydro-geomorphic processes with vegetation development is an integral step in understanding habitat evolution in riparian corridors. We quantified floodplain deposition dynamics and topographic complexity to better understand how riparian forest structure and composition changed through time and over the floodplain. We used LiDAR data and a temporally-rich dataset of aerial photography coupled with field data to analyze landscape-scale patterns with species composition and floodplain dynamics. We examined the significance of two definable large-scale floodplain processes - meander migration and channel abandonment - that set-up the spatial structure of riparian forest dynamics. In addition, we analyzed the potential feedback of riparian vegetation on floodplain formation. Canopy structure and floodplain heterogeneity generally showed a log-linear relationship over time and were highly correlated with species composition changes over time and with a range of past hydrologies (inter- and intra- annual patterns). These data suggest that analysis of landscape formation processes, in particularly during the early stages, improves our understanding of riparian forest dynamics and that channel abandonment can be a significant alternative pathway to cottonwood recruitment. By analyzing the coupled eco-geomorphic processes we can more effectively scale-up riparian vegetation patterns by applying a functional knowledge of the corridor-wide channel dynamics and floodplain hydrology.