RIVER TRAINING STRUCTURES: EFFECTS ON FLOW DYNAMICS, CHANNEL MORPHOLOGY, AND FLOOD LEVELS

The use of river training structures to create a narrower and deeper river channel – including wing dikes, bendway weirs, and chevron dikes – along the Middle Mississippi River (MMR)has been linked to changes in flow conveyance and, in some locations, to large-scale increases in flood levels (up to 4 to 6 m). However, the mechanisms by which wing dikes impact flood conveyance are not fully understood. Previous research suggests river training structures may broadly affect flow conveyance: (1) through direct interaction with flow and/or (2) indirectly by their effects on channel geometry and bedforms.

Explicit assessment of training structure effects on flow conveyance was accomplished by constructing two 2-D hydrodynamic models: (1) a calibrated model of present river conditions (i.e., with wing dikes) and (2) a scenario model without training structures. Modeling results show that river training structures can locally create complex flow patterns, lengthening flow paths and thereby reducing flow velocities by up to -2.0 m/s (80% for a 40-year flow event). The reduction in flow velocity resulted in some losses of flow conveyance and corresponding increases in stage of up to 0.5 m. However, these results suggest that indirect effects are the primary driver of historic decreases in flood flow conveyance and large-scale increases in flood stages along the MMR.

Recent construction of new river training structures within St. Louis Harbor and the availability of pre-construction data provide a unique opportunity to test the impacts of these structures on flow dynamics, flood conveyance, and river channel morphology. Comparison of pre- and post- river-training structure construction bathymetry within St. Louis Harbor shows significant local aggradation (up to 3 m) within portions of the navigation channel instead of the desired incision (deepening). This unintended geomorphic response suggests current screening techniques (physical models) used in the design and placement for river training structure may be inadequate to accurately predict the complex geomorphic response created by the construction of such structures. We are currently testing numerical models to see if they have a better predicative capability.