H081
Hydrology as a Physical Science in the Age of Machine Learning: Toward Theory-Driven Discovery
Session ID#: 252254
Session Description:
Hydrology, a distinct geoscience (Eagleson, 1991), encompasses phenomena at scales pore to continental, seconds to decades. In an era increasingly dominated by artificial intelligence and machine learning (AI/ML), the community faces a challenge: develop and refine physical theories that can keep pace with, and complement, the predictive capabilities of AI/ML (Nearing et al., 2021). The proliferation of “Big Data” offers unprecedented opportunities, but underscores the need for rigorous, theory-driven insights spanning relevant spatiotemporal scales.This session offers a forum for hydrologic theory in the age of AI, through: 1) Reconciling observations with theory, 2) Integrating or discovering theory through AI/ML, 3) Diagnosing or detecting model misspecification/errors, and 4) Introducing information-theoretic approaches to hydrology. We seek submissions go beyond minimization of residuals towards deeper physical understanding of the hydrosphere. Failure cases are of interest, as they provide fertile ground for hypothesis testing, model refinement, and theory development.
Co-Sponsor(s):
Index Terms:
0555 Neural networks, fuzzy logic, machine learning [COMPUTATIONAL GEOPHYSICS]
1839 Hydrologic scaling [HYDROLOGY]
1847 Modeling [HYDROLOGY]
4475 Scaling: spatial and temporal [NONLINEAR GEOPHYSICS]
Primary Convener: Jonathan M Frame, University of Alabama, Department of Geological Sciences, Tuscaloosa, AL, United States
Convener: Jasper A Vrugt, UC-Irvine, Irvine, United States
Student/Early Career Convener: Tadd Bindas, Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States
Demonstrating the Feasibility of DL-Based Pluvial Flood Mapping in Urban Settings (1915544)
Keivan Tavakoli, University of Texas at Austin, Austin, TX, United States, Sadra Seyvani, University of Alabama, Tuscaloosa, United States, Adam Smith, Rutgers University, New Brunswick, United States, Mostafa Saberian, Clemson University, Department of Civil Engineering, Clemson, United States, Mohamed Abdelkader, University of Iowa, IIHR—Hydroscience & Engineering, Iowa City, United States, Jonathan M Frame, University of Alabama, Department of Geological Sciences, Tuscaloosa, AL, United States and Marouane Temimi, Stevens Institute of Technology, Department of Civil, Environmental, and Ocean Engineering (CEOE), Hoboken, United States
Historical Streamflow Reconstruction Using Hydro-Transformer Models with High-Confidence Gage Estimates (1958132)
Taye Akinrele1, Pratiksha Chaudhari1, Sonam Lama1, Quinn Lee1, James Halgren1 and Shahram Rahimi2, (1)The University of Alabama, Alabama Water Institute, Tuscaloosa, United States, (2)The University of Alabama, Computer Science Department, Tuscaloosa, United States
Extracting Hydrological Information from LSTM Models Using Cell State Analysis (1968062)
Collins Mba, New Jersey Institute of Technology, Newark, United States, Deon Fernando, University of Alabama, Tuscaloosa, United States, Jonathan M Frame, University of Alabama, Department of Geological Sciences, Tuscaloosa, AL, United States and James Halgren, The University of Alabama, Alabama Water Institute, Tuscaloosa, United States
Building a Streamflow Reanalysis Dataset using Deep Learning-Based Geostatistical Signal Separation (1974712)
Quinn Lee1, James Halgren1, Sonam Lama1, Pratiksha Chaudhari1, Hannah Lemons2 and Joshua Cunningham1, (1)The University of Alabama, Alabama Water Institute, Tuscaloosa, United States, (2)The University of Alabama, Department of Geological Sciences, Tuscaloosa, United States
Explaining and Predicting the Model Performance via Convex Hull Analysis of Internal States of LSTM based NWM Surrogate (1975758)
Deon Fernando, University of Alabama, Tuscaloosa, United States, Jonathan M Frame, University of Alabama, Department of Geological Sciences, Tuscaloosa, AL, United States and Collins Mba, New Jersey Institute of Technology, Newark, United States
Developing a Machine Learning Framework for Seasonal Lake Level Forecasting of the Great Salt Lake (1979006)
Sophia Breuckman, University of Minnesota Twin Cities, Minneapolis, United States, Sifan Abera Koriche, University of Alabama, Tuscaloosa, AL, United States and Steven J Burian, Cooperative Institute for Research to Operations in Hydrology (CIROH), Alabama Water Institute, The University of Alabama, Tuscaloosa, AL, Alabama Water Institute, Tuscaloosa, United States
Neural Operator for Subsurface Soil Moisture Imputation (1982689)
Vidhi Singh1, Dr. Abhilash Singh, Ph.D.2 and Kumar Gaurav2, (1)Indian Institute of Science Education and Research Bhopal, Bhopal, India, (2)Indian Institute of Science Education and Research Bhopal, Earth and Environmental Sciences, Bhopal, India
