Modeling Paleofloods in the Tennessee Valley: Calibrating the Naturals Model and Associating Flood Deposits with Flows

Flood frequency estimates for extreme flows outside of the historical record have been studied for over 50 years. However, many approaches used for such estimates lack precision and have large uncertainties. This is a concern, as accurate flood frequency estimates are critical to informing river operation organizations, such as the Tennessee Valley Authority (TVA), for risk assessments and dam safety improvement projects. The TVA manages 49 dams in the Tennessee River system, necessitating extensive monitoring and modeling of the rivers and reservoirs throughout Tennessee and parts of surrounding states. Recent studies have highlighted the use of paleoflood hydrology as a source of data to extend historical records of extreme floods. This method can also be used as a way to verify flows found in historical records. However, paleoflood studies in the United States have centered primarily on dry climates in the western part of the country, not humid regions such as East Tennessee, where this study is located. Like many major rivers in the United States, the rivers studied herein have undergone changes due to the construction of dams and residential development, which increases the difficulty of considering historical floods in frequency analyses. Thus a “Naturals” HEC-RAS model was created to associate historical flood deposits, with flows estimated from the model. The model was calibrated to an observed flood that occurred when only 3 dams were on the river, allowing closer representation of conditions when these paleoflood deposits were created. Rating curves were developed for different locations along the Tennessee River by modifying the Naturals model so that flood deposits identified in the future can be correlated with a flow-rate. This work demonstrates the need for a better understanding of past floods and how their associated flows would route through the current system. These results will be used to support future paleoflood studies for the TVA, which will in turn provide more accurate flood frequency data. This improved flood frequency data can then be used to improve dam safety decisions. Further, this work acts as a proof that paleoflood hydrology studies are feasible in the eastern United States.