Waterworks Archive

Characterization of flash floods in Eastern Kentucky

by Nageshwar R. Bhaskar, Mark French, and George K. Kyiamah
Civil Engineering Department
University of Louisville

Flash flood is a term applied to type of flood that is frequently experienced in Eastern Kentucky and is commonly associated with a violent, short-duration, high-intensity convective storm that falls on a steep watershed. The climatic and geomorphic characteristics of watersheds in Eastern Kentucky provide a suitable environment for persistent flash-flood hazards. These floods usually occur only on small headwater streams, minor streams, or inadequately drained areas. The key hydrological factors that control floods, in general, and flash floods, in particular, include temporal and spatial aspects of rainfall, landuse, soils, and geomorphological aspects of the watershed. Currently, the data available to monitor flash floods in Eastern Kentucky watersheds is limited to such information as rainfall data recorded by the National Weather Service (NWS) rain gages (about one gage per fifty square miles) and streamflow data (minimum watershed size about 50 square miles) recorded by United States Geological Survey (USGS) gages. Furthermore, in 1981 the NWS installed additional rain gages under a program called Integrated Flood Observing and Warning System to provide more extensive spatial representation of rainfall for monitoring potential flash flooding in Eastern Kentucky.

Knowledge of past significant flash-flooding events, recognition of the hydrometeorological patterns that produced them, and a careful analysis of observed flood events are important considerations for developing an effective flash-flood characterization system. With this in mind, four watersheds ranging in size from 50 to 90 square miles were selected to study their flash-flood characteristics. From January 1990 to August 1995, thirty flood events were selected for flash-flood identification and characterization. A majority of these events were associated with short-duration and high-intensity rainfall resulting in at least one inch of rainfall and hydrographs that exhibited sharp peak-flow characteristics. Other moderate flood events were also selected to bring out the differences between ordinary flood events and flash floods.

In flash-flood situations, certain portions of the flood hydrograph, such as the rising limb, time to peak, and magnitude of the peak, are more important than others. To recognize this, parameters such as the rising curve gradient, k; flood magnitude ratio, M; and flood response time, Tp are evaluated directly from the observed flood hydrographs. Typically, for a flash flood, one would expect high values of k and M, and a low Tp value. Hence, these parameters can be used to distinguish flash floods from other flood events within a watershed. For each flood event, parameters k, M and Tp are used to develop relative severity factors RK, RM and RT, respectively, using an ordinal scale of assignment. Furthermore, these three factors are added together to obtain an overall index referred to as the flash-flood index, RF. The flash-flood index is expected to be high (greater than 10) for flash floods as compared to other floods, because the severity factors, RK, RM and RT will also be high. Results from this study show that in one of the watersheds, a reported flash-flood event on May 18, 1995 had values RK=4, RM=8 and RT=5 with a flash flood index RF=17. Overall, 14 of the 30 flood events analyzed in this study had flash flood indices greater than 10. Thus, the indexing procedure developed in this study provides a systematic way of distinguishing a flash-flood event from other flood events. Current practice is based primarily on the rainfall event producing flash flooding. The ability to separate flash floods from other floods based on the characteristics of the observed streamflow hydrograph may provide a better understanding of how to model such events. This in turn can be used to enhance flash-flood forecasting techniques.

(This article is based on a recent KWRRI Research Report, and copies are available from the Waterworks editor.)

Last modified: December 5, 1996

Copyright © 1996 Kentucky Water Resources Research Institute
University of Kentucky