Relationship of point bar morphology to channel curvature and planform evolution

Abstract Point bars are prominent features in meandering rivers, yet our understanding of the complex interactions among the morphology of the point bars to channel planform and curvature remains incomplete. This study seeks to address the extent to which channel centerline curvature can explain point bar morphology. High-resolution elevation datasets obtained from multibeam sonar and structure-from-motion surveys are used to characterize point bar morphology on twelve point bars within two river systems; the Pearl and Wabash rivers. Spatial series of centerline curvature are compared to channel and point bar characteristics including width, migration rate, centerline longitudinal bed elevation, transverse bed slope, and a shape factor. Results show that the Wabash reach has overall lower channel curvature values and higher rates of migration compared to the Pearl reach. However, the spatial lag between curvature and width and migration rate are similar between reaches. The mode of bend migration differs between the two reaches, where the Pearl reach has bends mostly evolving through translation downstream, and along the Wabash reach, one bend is expanding, two bends are mostly stationary, and three bends are translating downstream. The translating bends in both reaches exhibit bilinear (a pronounced break in slope) and near-horizontal (a pronounced break in slope with a nearly flat bar top) transverse bar profiles and maximum rates of migration downstream of the apex. The Pearl reach has a systematic distribution of transverse bar profiles where linear profiles are found along the bar head and bar tail, and near-horizontal profiles are concentrated along the middle bar. The distribution along the bars of the Wabash reach is less systematic, with fewer near-horizontal profiles overall. Finally, to evaluate the influence of curvature on the channel bed and bar morphology, the field data are compared to synthetic bed elevation data generated using pyRiverBed, a centerline curvature-driven bed evolution model that maintains a constant width. Analyses reveal the synthetic morphologic data agree reasonably well with the characteristics and magnitude of the field observations, with a mean absolute error of 0.9 and 1.2 m for the Pearl and Wabash rivers, respectively. Values of transverse slope were also compared between the field and synthetic data and the synthetic data are consistently under-predicted along the Pearl. Differences observed between the field and synthetic data are the result of variability in the natural river system not captured by the model including channel width, the presence of local bedrock outcrops, and abrupt changes in curvature.