Evaluating next-generation intensity–duration–frequency curves for design flood estimates in the snow-dominated western United States

Journal Article · 13 December 2019 · Hydrological Processes

DOI: https://doi.org/10.1002/hyp.13673 · OSTI ID:1600720

Yan, Hongxiang  ^[1]; Sun, Ning ^[1]; Wigmosta, Mark S. ^[2]; Leung, Ruby L. ^[1]; Hou, Zhangshuan  ^[1]; Coleman, Andre ^[1]; Skaggs, Richard L. ^[1]

---

1. Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
2. Pacific Northwest National Lab. (PNNL), Richland, WA (United States); Univ. of Washington, Seattle, WA (United States)

Civil infrastructure such as culverts and bridges are commonly designed using precipitation-based intensity-duration-frequency (PREC-IDF) curves, which assume that the occurrence of precipitation is in the form of rainfall and immediately available for the rainfall-runoff process. In snow-dominated regions, where most winter precipitation occurs as snow that melts during spring to early summer, the use of standard PREC-IDF curves may lead to substantial underestimation of design floods and high failure risk of infrastructure. In this context, we developed next-generation IDF (NG-IDF) curves that characterize the actual water reaching the land surface (i.e., rainfall plus snowmelt) to enhance standard infrastructure design in snow-dominated regions. This study evaluates the performance of NG-IDF curves coupled with U.S. Department of Agriculture Technical Release 55 hydrologic model in estimating design floods for 246 snowy locations in different hydroclimate regimes of the western United States. Design flood estimates from a well-validated continuous simulation using a physics-based hydrologic model, the Distributed Hydrology Soil Vegetation Model (DHSVM), were used as the performance benchmark. Compared with the benchmark estimates, the standard PREC-IDF curves led to substantial errors in design flood estimates while the NG-IDF curves significantly reduced these errors. For example, the averaged error in the 50-year design flood estimates over the 246 locations was reduced from 31% with the use of PREC-IDF curves to 12% with the use of NG-IDF curves. Despite the different model structures, the single-event NG-IDF approach versus the continuous simulation DHSVM did not exhibit statistically significant differences in 91% of the 246 locations for the 50-year design flood estimates. We find this indicates a satisfactory performance of NG-IDF curves to estimate design flow under the conditions tested in the snow-dominated western United States. This article also presents technical suggestions and the limitations of infrastructure design using NG-IDF curves for regulatory agencies and practicing engineers.

• Research Organization: Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• Grant/Contract Number: AC05-76RL01830
• OSTI ID: 1600720
• Report Number(s): PNNL-SA-143219
• Journal Information: Hydrological Processes, Journal Name: Hydrological Processes Journal Issue: 5 Vol. 34; ISSN 0885-6087
• Publisher: WileyCopyright Statement
• Country of Publication: United States
• Language: English