Holocene stratigraphy in flood frequency analysis

Abstract

The risk of rare catastrophic flood events is extremely difficult to evaluate by using conventional flood frequency analyses. The true return period may be so long compared to gaging station or historical records that the stratigraphic record of the Holocene (0–10,000 years B.P.) may be the best method for assessing the risk of outstanding floods. Several geologic methods applicable to flood frequency analysis are (1) tree ring analysis of floodplain trees destroyed or scarred by floodwaters, (2) radiocarbon dating of organics in deposits (landforms) eroded by recent floods, (3) radiocarbon dating of organic materials in older flood deposits, (4) pollen analysis and radiocarbon dating of organic materials in lenses of fine‐grained alluvium buried by coarse flood deposits, (5) radiocarbon dating of soil horizons or organic material in slack‐water deposits, (6) radiocarbon or archaeological dating of soils developed in older flood deposits, and (7) dating of landforms eroded or buried by floods based on changes in quantitative soil properties versus time. Using one or more of the above geologic techniques, return periods of some recent catastrophic floods in the United States are: 1976 Big Thompson River, Colorado, 5000 years (?); 1964 Klamath River, California, 100 years (?); 1972 Western Run, Maryland, 2100 years (?); and 1972 Elm Creek, Texas, 400 years (?).