Utilizing archival data to assess historic changes in flood flow conveyance of the Mississippi River

The assessment of change in river systems requires reference conditions. However, most large navigable waterways in the United States and elsewhere around the world have a wealth of archival data because of past engineering projects and flood control efforts. For example, in the United States, large quantities of archival data are available for the Mississippi River System which can be used to assess historical baseline conditions and change. Maps, charts, surveys, structure-history databases, and other quantitative measurements stretch back 100 to as much as ~200 years. The purpose of his dissertation was to develop a robust methodology from which to use these archival data to establish historic reference conditions in order to quantify and assess the causes of change in flood levels. Since the early 19th century, the Middle and Lower Mississippi River (MMR and LMR) have been intensively modified for flood protection and commercial navigation. In order to quantify the effects of levee expansion, channel modification, and land-cover change upon flood stages, a 1-D unsteady flow model was developed for multiple historical reference conditions ( retro-models ) for three large study reaches (225-315 km each): one along the MMR and two reaches along the LMR. For each reference condition, four 1-D unsteady-flow models were developed. These models include a calibrated model of actual conditions and three scenario models: 1) a model with levees of the next time step, 2) a model with the channel geometry of the next time step, and 3) a model with floodplain roughness (i.e., land cover) of the next time step. Comparison of the model for actual conditions and the scenario models provide a quantitative assessment of levee expansion, channel modification, and land-cover change on stage. Comparison of the modeling results for this investigation showed significant changes in stage for equal discharges between each of the modeled time steps. Changes in stage for the three modeled reaches ranged from -3.1 to 4.4 m. The largest changes in flood stage were found along the MMR. The largest decreases in flood levels were found along the LMR between Obion and Memphis, TN. These results confirm previous hydrologic analyses, but show stage-changes as continuous longitudinal profile and not just at gaging stations. Scenario modeling suggests that the majority (38 to 70%) of the changes in flood stage on the LMR and MMR study reaches can be attributed to changes in channel geometry and hydraulic roughness. Levees were the next largest contributor to changes in flood stage. For time steps with significant levee expansion, these structures increase stage up to 1.0 m. Observed changes in floodplain land cover were associated with little (or none) of the increase in flood stage. These result show changes in channel geometry and roughness related to river engineering tools employed for the facilitation of navigation and flood conveyance are the principal drivers of historic changes in flood stages along the LMR and MMR. The results from the 1-D scenario model assessment suggest wing dikes may broadly affect flow conveyance in two ways: (1) through direct interaction with flow and (2) indirectly by their effects on channel geometry and roughness. Direct effects of wing dikes on flood stage were assessed by constructing two 2-D hydrodynamic models: (1) a calibrated model of actual conditions (i.e., with wing dikes) and (2) a scenario model without wing dikes, for a heavily modified reach of the MMR. Comparison of the model of actual conditions and model without wing dikes revealed that direct effects of wing dikes increase stage, modestly by only 0.1 to 0.5 m, depending on discharge and location. Wing dikes also were shown to increase and decrease channel velocities. Local increases in flow velocities of up to 0.4 m/s were found along portions of the main channel. Decreases in flow velocity by as much as -2.0 m/s were found along the edge of the channel within the wing dike fields. The direct effects of wing dikes on flow conveyance also were observed to decrease with an increase in flow, a result that runs contrary to the total cumulative effects of wing dikes observed empirically. These results suggest that the indirect effects are the likely cause of the historic decreased in flood flow conveyance and large-scale increases in flood stages along portions of the MMR