Channel multiplicity and width redistribution in braided rivers: a multi-decadal scaling and crossover analysis

Braided rivers are characterized by multiple interacting channels whose number and width continuously reorganize along the river course. How channel multiplicity relates to the partition of total wetted width among individual channels has remained largely unexplored. We address this gap using cross-sectional data from the Brahmaputra–Jamuna River (Bangladesh), explicitly incorporating individual channel widths rather than channel count alone, and test the results against a hydraulic argument in which discharge, and hence width, is assumed to be shared equally among channels. We further show theoretically that this uniform partition assumption is the stable outcome of a broader, conveyance-weighted family of partition rules below a critical exponent, beyond which the system becomes unstable toward width concentration into fewer dominant channels. Total wetted width increases more slowly than channel number, indicating a constrained wetted section, while partition evenness increases steadily, reflecting progressive homogenization. The mean channel width decreases with the channel number at a rate close to that expected under the assumption of equal discharge partition among channels; however, a chronological record of eleven datasets spanning 1976–2007 shows that this behavior is not captured by a single scaling law, with a decay-to-plateau pattern traced progressively further as sensor resolution improves. In the two most densely sampled datasets, the decline in mean width becomes markedly steeper beyond the plateau, consistent with the predicted instability and with the documented legacy of the 1950 Assam earthquake. These findings indicate that the apparent complexity of braided systems reflects channel-width redistribution within a constrained geometry, with genuine scale-dependent structure superposed on a theoretically motivated mean-field description.