Most nuclear powerplants in the United States are near rivers, large lakes, or oceans. As evident from the Fukushima Daiichi, Japan, disaster of 2011, these water bodies pose inundation threats. Geologic records can extend knowledge of rare hazards from flooding, storm surges, and tsunamis. This knowledge can aid in assessing the safety of critical structures such as dams and energy plants, for which even remotely possible hazards are pertinent. Quantitative analysis of inundation from geologic records perhaps is most developed for and applied to riverine flood hazards, but because of recent natural disasters, geologic investigations also are now used widely for understanding tsunami hazards and coastal storm surges.

Layered sedimentary deposits commonly give the most complete geologic record of large floods, storm surges, and tsunamis. Sedimentary layers may be preserved for hundreds or thousands of years in suitable depositional environments, thereby providing an archive of rare, high-magnitude events. All inundation hazards discussed in this report—riverine floods, tsunamis, and storm surges—have had long records extracted from sedimentary sequences, many specifically for hazard assessment.

Geologic records commonly are imprecise, so most hazard assessments benefit from evaluation of many sites and rigorous uncertainty assessment. Despite uncertainties, geologic records commonly can improve knowledge of the types and magnitudes of hazards threatening specific sites or regions. New statistical tools and approaches can efficiently incorporate geologic information into frequency assessments. These tools are most developed for riverine flood hazards, but are to some degree transferable to other episodic natural phenomena such as tsunamis and storm surges.

Even with these efficient statistical approaches for examining geologic records, systematic landscape changes may reduce the applicability of retrospective assessments. These non-stationarity issues (such as climate change, sea‑level rise, land-use, dams and flow regulation) may all affect the validity of using past experience—no matter how complete the record—to assess future likelihoods. These issues require site-specific consideration for nearly all hazard assessments drawn from geologic evidence.

A screening of the 104 nuclear powerplants in the United States licensed by the Nuclear Regulatory Commission (at 64 sites) indicates several sites for which paleoflood studies likely would provide additional flood-frequency information. Two sites—Duane Arnold, Iowa, on the Cedar River; and David-Besse, Ohio, on the Toussaint River—have geologic conditions suitable for creating and preserving stratigraphic records of flooding and few upstream dams that may complicate flood-frequency analysis. One site—Crystal River, Florida1, on the Withlacoochee River and only 4 kilometers from the coast—has high potential as a candidate for assessing riverine and marine inundation hazards. Several sites on the Mississippi River have high geologic potential, but upstream dams almost certainly now regulate peak flows. Nevertheless, studies on the Mississippi River to evaluate long-term flood frequency may provide results applicable to a wide spectrum of regional hazard issues. Several sites in the southeastern United States have high geologic potential, and studies at these sites also may be helpful in evaluating hazards from outburst floods from landslide dams (river blockages formed by mass movements), which may be a regional hazard. For all these sites, closer investigation and field reconnaissance would be needed to confirm suitable deposits and settings for a complete paleoflood analysis. Similar screenings may help identify high-potential sites for geologic investigations of tsunami and storm-surge hazards.