February 25, 2014

Uncertainty in PMP and Recurrence Interval-Based Design Storm Estimates from Sparse Data

Gene Bosley, P.E., Tierra Group, Salt Lake City, Utah, USA; Pete Kowalewski, P.E., Tierra Group, Salt Lake City, Utah, USA; and Marybeth Dowell, Tierra Group, Elko, Nevada, USA

ABSTRACT: Dam safety and pollution control considerations require large, infrequent precipitation events (50-year up to Probable Maximum Precipitation (PMP)) to be determined for use in the design of tailings storage facilities (TSF), waste rock storage facilities (WRSF), heap leach facilities, and other process-related infrastructure. New, remote sites in particular lack data with a significant period of record to allow these relatively infrequent events to be determined. Designers often use regional data, attempt to correlate it to the site over short concurrent records, and correct for orographic factors. In some cases, designers adapt U.S. methods or use statistical methods such as Hershfield’s (World Meteorological Organization (WMO), 1986) to determine the PMP. Bias, error, and uncertainty associated with such evaluations are seldom estimated or reported. Using tropical, desert, and mountain site data, common methods for estimating the PMP and other rare events from limited data were investigated. Sites were selected based on theavailability of long, concurrent records of local and regional data or relatively dense gage networks; enabling a comparison of results generated from nearby datasets and sampling shorter periods of record, emulating the results that would be obtained had the total record been shorter or obtained outside the project area. Statistically-derived PMP estimates for U.S. sites are compared to estimates obtained using standard published U.S. National Weather Service (NWS) methods. The results obtained in this study demonstrate the need to reevaluate design storm estimates as data becomes available, not merely at the beginning of a project. Given the sensitivity of facility cost to freeboard and other runoff-related design criteria, revising design storm estimates during operations or approaching closure may achieve significant reductions in capital cost and/or environmental risk.

Download Article