Sediment supply to rivers, subsequent fluvial transport, and the resulting connectivity on network-scales are often sparsely monitored or subject to major uncertainty. Hence, we propose to adopt stochastic modeling approaches for studying network sediment connectivity. We demonstrate such a stochastic approach for modeling sand connectivity in the major, poorly monitored Se Kong, Se San, and Sre Pok (3S) tributaries of the Mekong River. Specifically, we run many random initializations of the CASCADE modeling framework for sediment connectivity in a Monte Carlo approach in order to quantify how unknown properties of sediment sources translate into uncertainty regarding network sediment connectivity. We identify a reduced ensemble of model realizations that reproduces downstream observations of sediment transport. This ensemble presents an inverse stochastic approximation of the spatial distribution, magnitude, and variability of transport capacity, sediment flux, and bed material grain size in the entire network (i.e., upscaling point observations to the entire network). The approximated magnitude of sediment flux in each tributary is controlled by reaches of low transport capacity (“bottlenecks”). These “bottlenecks” limit the ability of the inverse stochastic approximation to predict sediment transport in the upper parts of the catchment but they allow a clear partitioning of sand deliveries from the 3S to the Mekong, with the Se Kong delivering less (1.9 Mt/yr) and coarser (median grain size: 0.4 mm) sand than the Se San (5.3 Mt/yr, 0.22 mm) and Sre Pok (11 Mt/yr, 0.19 mm).

{"references": ["1.Schmitt, R. J. P., Bizzi, S., Castelletti, A. & Kondolf, G. M. Stochastic modeling of sediment connectivity for reconstructing sand fluxes and origins in the unmonitored Se Kong, Se San, and Sre Pok tributaries of the Mekong River. Journal of Geophysical Research: Earth Surface (in Review)."]}