Planning & Management

Oral

394292 - Nonstationary Flood Frequency Analysis using Stochastic Storm Transposition and Hydrologic Modeling

Tuesday, June 5
4:00 PM - 5:30 PM
Location: Greenway GH
Co-Authors: Daniel Wright, Madison,WI – University of Wisconsin Madison; Kathleen Holman, Denver, CO – United States Bureau of Reclamation; Zhihua Zhu, Guangzhou, China – Sun Yat-sen University

Floods are the product of complex interactions of factors including rainfall, watershed morphology, and antecedent conditions. Even in stationary conditions, flood frequency analyses using statistical models of river discharge offer few insights into these interactions and how they combine to create flood frequency. For example, the role of soil moisture in flood frequency is still poorly understood, despite its obvious importance. Understanding these interactions becomes critical in nonstationary conditions since some or all factors may be changing and too few discharge observations are available to understand how such changes alter flood frequency. In this research, we develop a framework that couples Stochastic Storm Transposition (SST) with a simple conceptual hydrologic model to analyze nonstationary flood frequency. This framework also allows us to distinguish sources of uncertainty in the model, inputs, and in initial conditions. The open-source SST software RainyDay is used to extend the rainfall record by temporal resampling and geospatial transposition of observed storms to generate a large number of realistic extreme rainfall “scenarios.” We first show the influence of regional climate and changing catchment characteristics on model parameters. We then use RainyDay combined with Stage IV multisensor precipitation data to generate rainfall scenarios for exceedance probabilities ranging from 1 to 0.002. We also derive the distribution of soil moisture from long-term continuous simulations of daily streamflow. Finally, we compare three flood frequency ensembles that consider 1.) rainfall input uncertainty, 2.) nonstationary watershed characteristics, and 3.) antecedent soil moisture. We contrast the results against stream gage-based statistical analyses. The framework is relative simple to apply in other settings, provided that sufficient data are available for hydrologic model calibration.

Guo Yu

Research Assisstant, PhD student
University of Wisconsin Madison

Presentation(s):

Send Email for Guo Yu


Assets

394292 - Nonstationary Flood Frequency Analysis using Stochastic Storm Transposition and Hydrologic Modeling



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Nonstationary Flood Frequency Analysis using Stochastic Storm Transposition and Hydrologic Modeling