CAREER: A Dynamic-Stochastic Approach to Rainfall and Flood Frequency Analysis Across Scales

点击翻译按钮获取中文摘要

Floods arise from the interactions between rainfall and ground conditions such as soil moisture and river channel dynamics. Yet, floods can also be described as a collection of random processes using statistics in addition to physical modeling.The relationship between flood severity and likelihood ("flood frequency analysis") has become a centerpiece in flood risk management practice. Changes in land use due to urbanization and climate variability implies that existing data and methods may no longer be valid. Meanwhile, study of the links between dynamic and statistical behavior of floods has been hindered by the lack of long-term, high-resolution observations of rainfall, soil moisture, and streamflow. This leaves us with limited understanding of how factors such as soil moisture and river channel dynamics modulate extreme rainfall to produce long-term flood frequency. This research presents a hybrid approach that uses recent advances in hydrometeorological observations, data processing approaches, models, and theory to better understand floods in a changing world. It also proposes several innovative education and outreach initiatives that will reach K-12, undergraduate and graduate students, and instructors using a combination of physical-virtual active learning tools, interactive web-based "apps," and lesson modules to reinforce learning and assess outcomes.The research will advance understanding of the role of four-dimensional (surface rainfall rate, time, and x,y spatial coordinates) storm structure as a driver of rainfall and flood frequency. The first objective focuses on regional geospatial rainfall structure and its linkages to rainfall and flood frequencies and trends. The second objective focuses on the physical drivers of flood frequency including rainfall, soil moisture, and their interactions. The principal inveatigatgor will use stochastic-dynamic modeling to establish the cross-scale links between flood severity and frequency that have thus far proven elusive. The proposed work will advance understanding of flood variability at multiple, coupled scales and provides tools that leverage modern computing power, regional high-resolution rainfall observations, and distributed computational watershed models. Using a framework for physically-based flood frequency analysis, this research will provide better insights into underlying drivers and their links to nonstationary flood frequency and severity. This research has an education/outreach component to train a diverse and "data literate" generation of professionals able to create a flood-resilient nation. Activities and teaching modules will also improve data literacy and geospatial understanding of flood hydrometeorology and the water cycle among a range of audiences including UW-Madison students, K-12 science fair attendees, high school science teachers and students, water resources practitioners, and the public. These initiatives are motivated by recent research showing that diverse teaching approaches including group activities, technology, and real world examples can reinforce learning and attract underrepresented groups to STEM. The proposed work will involve graduate students, undergraduates, and high school educators directly in the development of education and outreach products.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.