CS15A - Concurrent Session 15A: Is catchment area the dominant factor for PMF estimation in a supposedly hydrologically homogeneous region?

One of the primary drivers for rehabilitation of dams is to address inadequate spillway capacity. Determination of the inflow design flood (IDF) or design flow for a dam facility is governed by a few considerations, but in many instances it boils down to a balancing act between applicable dam safety rules, relevant industry standards, risk assessment and rehabilitation cost.

Diving one level down into the specifics behind the determination of the IDF determination, it is the estimate of the probable maximum flood (PMF) that consistently serves as a key reference, in particular for high hazard dams, and this holds true whether the evaluation is deterministic, probabilistic, or somewhere in between. It is assumed many times that catchment area will be a key input for estimation of the PMF, and relationships between PMF and catchment area are available in the literature for initial ballpark estimates. Going this route is reasonable, especially if comparing to values from other locations with similar climatic and hydrologic conditions, including land use. However, recent work brings up some questions about the assumption that catchment area is indeed a dominant factor.

This presentation will summarize the results of estimating the PMF for 4 dam facilities, namely Elm Lake Dam, Corsica Dam and Richmond Dam in South Dakota, and Lake Bronson Dam in Minnesota, with catchment areas ranging between 91 and 547 square miles. The study watersheds have common characteristics: low topographic relief, primarily agricultural, and recorded snowmelt peaks usually larger than the observed rainfall-driven floods. We followed the same methodology for PMF estimation at the 4 sites, starting with development and calibration of HEC-HMS hydrologic models, incorporation of updated values for probable maximum precipitation (PMP) -per work by the state of North Dakota in 2021-, and development and application of HEC-RAS models. The PMF estimates to share will suggest that parameters like the shape of the watershed, or the magnitude and associated likelihood of the flood of record, can be equally or more important than catchment area.