The braided Milk River, northern Montana, fails the Leopold–Wolman discharge-gradient test

Abstract

The Milk River, the northernmost tributary to the Missouri–Mississippi River system, exhibits an anomalous sand-bed braiding reach in an otherwise meandering system. Shortly after leaving Alberta and entering Montana the river suddenly changes to braiding and maintains this pattern for 47 km before entering Fresno Reservoir. Measured stream gradient and bankfull discharge in the braiding reach severely fail the Leopold and Wolman [U.S. Geol. Surv. Prof. Pap. 282B (1957) 39] slope–discharge test for differentiating channel patterns. While channel slope has long been regarded as one of the primary variables associated with braiding, our data from the sand-bed Milk River do not support this relationship. Instead, the data show that the braiding reach has a lower channel slope (0.00047) than the meandering reach (0.00055). Coupled with a constant discharge the unit length stream power is comparable between the two reaches. At the morphologic transition between meandering and braiding, a dramatic reduction in channel bank strength occurs where the sampled silt–clay content declines from 65% in the meandering reach to 18% in the braiding. This enables channel widening which is reflected in a 60% reduction in unit area stream power in the braiding reach. Thus, sediment transport capacity declines and channel bars are deposited. During waning flows, these bars are dissected, producing a braiding morphology. We suggest that for sand-bed braiding rivers the silt–clay percentage in the channel banks may be more important than slope. A review of the original Leopold and Wolman [U.S. Geol. Surv. Prof. Pap. 282B (1957) 39] dataset, and many subsequent analyses, reveals that most braided rivers studied were gravel-bed. As a result, causal variables associated with braiding in sand-bed environments may need a thorough evaluation.