Capability of the network scale D-CASCADE model in simulating sediment transport dynamics in various riverine contexts

Modeling sediment transport dynamics at the river network scale is challenging. This is due to the uncertainty in quantifying sediment volume entrainment at the reach scale, as well as to the complex way these volumes (that we call here “cascades”) travel to downstream reaches, possibly interacting with other cascades (e.g., sediment input from a tributary). In this context, network-scale (graph-like) numerical models offer a way to integrate these intricate processes, by using simplifying assumptions that allow for simulating sediment routing through networks and potentially for long time periods.

We present the capability of a novel version of the network-scale model D-CASCADE (Dynamic CAtchment Sediment Connectivity And DElivery) in simulating sediment transport in three different (for size and morphological typology) riverine contexts: a mountain stream (the Solda Torrent), a low-anthropized river (the Tagliamento River), and a high-anthropized river (the Po River). Using well-known empirical formulas, the original model simulates the generation and pathways of multi-sized sediment cascades at the reach scale and within a discrete time representation (here daily). In this new version, we add numerical developments regarding the way sediment transport equations are accounted for through the network, and how energy is split among the cascades to respect the sediment mass balance within the simulated time step. These new features allow for a more robust representation of cascade interactions as they move downstream, with the aim of better estimating reach sediment fluxes, velocity, and budgets.

By comparing the model’s results with bedload fluxes measured in different rivers through different techniques (e.g., geophones, the morphological method), we demonstrate the ability of the model to produce a realistic representation of river connectivity both at reach scale (sediment fluxes and budget) and at network scale (sediment traveling velocity and provenance). We also discuss the sensitivity of two core parameters of the model: the active layer depth and the sediment traveling velocity.

The present study demonstrates how D-CASCADE enables the representation of sediment entrainment, transport, and deposition at the reach scale, thereby revealing key aspects of river connectivity functioning at the network scale. It also shows that, despite the scarcity of field data on bedload sediment transport, measurements taken in a few reaches within a network are enough to validate the network functioning generated by D-CASCADE.