Proceedings of the 2022 Mississippi Water Resources Conference

Year: 2022 Authors: Counts R., Davidson G., Yarbrough L.D.

Subsurface erosion due to underseepage during large floods, particularly through piping and the formation of sand boils, poses a significant threat to the integrity of floodplain levees in certain geologic settings. Research using natural geochemical tracers indicates that seepage beneath levees can follow both shallow and deep flowpaths. Though effective, tracking geochemical tracers is somewhat labor intensive and takes time. This project uses ground penetrating radar (GPR) and unmanned aerial system (UAS) imagery, which can be rapidly deployed with real-time viewing of the data as it is acquired, to test their effectiveness in mapping seepage and piping flowpaths. In 2019, the lower Mississippi River valley experienced significant and sustained flooding. During this historic event, GPR profiles and UAS thermal imagery were acquired in between Mississippi River mainline levees and multiple active sand boils north of Vicksburg, MS to identify active seepage flow paths and locate new sand boils in heavily vegetated areas. Despite recent mitigation efforts in the study area, which included constructing a 3-m thick clay pad over the region of older sand boil formation, new sand boils formed on the floodplain surface beyond the edge of the mitigation area during the 2019 flooding. GPR surveys using 160 MHz and 450 MHz hyperstacking antennae and UAS thermal imagery were acquired in between the actively flowing sand boils and the levee with promising results. Although wet, clay-rich soils are among the worst conditions for GPR, the 160 and 450 MHz antenna imaged to depths of 14 m and 6.5 m, respectively, and revealed multiple anomalies that could be seepage flowpaths. Additionally, actively flowing sand boils were apparent in the thermal imagery. Not only can GPR and UAS surveys be rapidly deployed, they can also cover very large areas in short time spans, and our proof-of concept results show these methods could be valuable tools for real-time levee assessments during flood conditions.