H079
Physics-Informed Machine Learning in Hydrology and Land Surface Processes

Session ID#: 160601

Session Description:
Machine learning (ML) in hydrology and land surface processes, e.g., energy balance or ecosystem processes, has made substantial progress in the past several years, but it is not without its limitations regarding the scope of prediction, interpretability and relatively underexplored theoretical contributions. This session focuses on integrating machine learning approaches with process-based models to address these issues. Topics include, but are not limited to: 1) Approaches integrating machine learning and process-based modeling components; 2) Computational approaches in physics-informed neural networks or learnable models for hydrology, hydrodynamics or land surface processes; 3) Large-scale applications of scientific machine learning and physics-informed ML methods; 4) The incorporation of ML methods with Earth System Models, especially those focusing on land surface and hydrologic components; 5) Gaining physical knowledge and insights using ML; and 6) Interpretative machine learning; 7) Benchmark datasets and problems; 8) Interactable, explainable AI and trust.
Index Terms:

1804 Catchment [HYDROLOGY]
1847 Modeling [HYDROLOGY]
1860 Streamflow [HYDROLOGY]
1942 Machine learning [INFORMATICS]

Primary Convener:  Chaopeng Shen, Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States
Conveners:  Dipankar Dwivedi, Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, United States, Hoshin Gupta, Hydrology and Atmospheric Sciences, The University of Arizona, Tucson, AZ, United States and Xingyuan Chen, Pacific Northwest National Laboratory, Richland, WA, United States
 
How to select an objective function using information theory (1047356)
Timothy O Hodson, U.S. Geological Survey, Urbana, IL, United States, Tyler J Smith, Clarkson University, Department of Civil & Environmental Engineering, Potsdam, NY, United States, Lucy Amanda Marshall, University of New South Wales, School of Civil and Environmental Engineering, Sydney, NSW, Australia and Thomas Mark Over, U.S. Geological Survey, Urbana, Illinois, United States
 
Using Temporal Graph Neural Networks to Leverage High Fidelity Observation Data for Improving Generalizability of Hydrological Models (1053219)
Raya Horesh1, Muneeza Azmat2, Malvern Madondo2, Fearghal O'Donncha3, Arun Bawa4, Kelsey Dipietro5, Raghavan Srinivasan6 and Michael Jacobs7, (1)IBM Research USA, Yorktown Heights, NY, United States, (2)IBM Research USA, Yorktown Heights, United States, (3)IBM Research - Ireland, Cognitive IoT, Dublin, Ireland, (4)Texas A&M University, Temple, United States, (5)Sandia National Laboratories, Jill Hruby Postdoctoral Fellow, Albuquerque, NM, United States, (6)Texas A&M University, Director of Spatial Sciences Laboratory, Agrilife Research, College Station, TX, United States, (7)IBM Corporate Social Responsibility, Armonk, United States
 
Using machine learning to identify hydrologic signatures with an encoder-decoder framework (1054752)
Hilary K McMillan, San Diego State University, San Diego, CA, United States and Tom Botterill, Independent Researcher, San Diego, United States
 
The value of large-scale climatic indexes for forecasting the severity of widespread flooding using dilated convolutional neural networks (1058054)
Larisa Tarasova, Columbia University, Department Earth and Environmental Engineering, New York, United States; Helmholtz Centre for Environmental Research - UFZ, Catchment Hydrology, Halle (Saale), Germany and Upmanu Lall, Columbia University, Columbia Water Center and Department of Earth and Environmental Engineering, New York, United States; Columbia University, Department of Earth and Environmental Engineering, New York, United States
 
Developing Catchment-Scale Physical-Conceptual Rainfall-Runoff Models Using Machine-Learning (1063673)
Yuan-Heng Wang, University of Arizona, Department of Hydrology and Atmospheric Sciences, Tucson, AZ, United States and Hoshin Gupta, Hydrology and Atmospheric Sciences, The University of Arizona, Tucson, AZ, United States
 
Effect of Deep Learning on the Prediction of Downward Longwave Radiative Flux in the Arctic (1082692)
Dae-Hui Kim, Yonsei University, Seoul, Korea, Republic of (South) and Hyun Mee Kim, Yonsei University, Seoul, South Korea
 
A Physical-Machine Learning Hybrid Model for Predicting Peak Streamflow During Extreme Flooding (1083829)
Feyera A Hirpa1, Wenli Zhao2, Yu Zhang1, Sam Lamont1, Zhuo Liu1, Yi Liu1, Yawen Shen1, Hoda El Safty1, Brett Baker1, Ting Li1, Liping Liu1 and Shabaz Patel1, (1)One Concern, Inc., Menlo Park, CA, United States, (2)Columbia University, Department of Earth and Environmental Engineering, New York, NY, United States
 
Customized Deep Learning for Precipitation Bias Correction and Downscaling (1110812)
Di Tian, Auburn University, Auburn, AL, United States and Fang Wang, Auburn University, Crop, Soil and Environmental Sciences, Auburn, AL, United States
 
Neural ODEs in Large-Sample Hydrology: Using Hybrid Models to Learn General Hydrologic Mechanisms from Multiple Catchments (1112312)
Marvin Hoge, Marco Baity Jesi, Fabrizio Fenicia, Andreas Scheidegger and Carlo Albert, EAWAG Swiss Federal Institute of Aquatic Science and Technology, Duebendorf, Switzerland
 
A physics informed machine learning approach to estimate surface soil moisture (1114576)
Abhilash Singh and Kumar Gaurav, Indian Institute of Science Education and Research Bhopal, Earth and Environmental Sciences, Bhopal, India
 
Incorporating Rainfall Loss Process with Machine Learning Models for Prediction in Ungauged Basins (1119628)
Pin-Ching Li and Venkatesh Merwade, Purdue University, West Lafayette, IN, United States
 
Benchmarking deep learning model estimates of hydrologic pathway contribution to streams and springs using conceptual and physically-based models (1122615)
Admin Husic, University of Kansas, Civil, Environmental and Architectural Engineering, Lawrence, KS, United States, Nabil m.hussain al Aamery, University of Kentucky, Lexington, KY, United States and James Fox, University of Kentucky, Department of Civil Engineering, Lexington, KY, United States
 
Quantile-Based Encoder-Decoder Deep Learning Models for Multi-Step Ahead Hydrological Forecasting (1122936)
Mohammad Sina Jahangir, John You and John Quilty, University of Waterloo, Civil and Environmental Engineering, Waterloo, ON, Canada
 
Migrating Hydrologic Data Based on Machine Learning to Enhance Streamflow Prediction in Transboundary Catchment (1135325)
Kai Ma, Ziyue Xu and Daming He, Yunnan University, Yunnan Key Laboratory of International Rivers and Trans-boundary Eco–security, Kunming, China
 
A physics-informed machine learning model for streamflow prediction (1136287)
Liang Zhang1, Dino G. Bellugi2, Shufan Li3, Ayush Kamat4, Justin Kadi5, Edom Moges1, Galen Gorski6, Omar Wani7 and Laurel Larsen8, (1)University of California Berkeley, Berkeley, CA, United States, (2)University of California, Berkeley, Department of Geography, Berkeley, CA, United States, (3)University of California Berkeley, Berkeley, United States, (4)University of California, Berkeley, Department of Electrical Engineering and Computer Science, Berkeley, CA, United States, (5)University of California, Berkeley, Division of Computing, Data Science, and Society, Berkeley, CA, United States, (6)US Geological Survey, Reston, United States, (7)California Institute of Technology, Division of Geological and Planetary Sciences, Pasadena, United States, (8)University of California, Department of Geography, Department of Civil and Environmental Engineering, Berkeley, United States
 
Towards the Hydrological Interpretability of Long-Short Term Memory Machine Learning Modeling (1137626)
Luis De la Fuente1, Mohammad Reza Ehsani2, Hoshin Gupta3 and Laura E Condon1, (1)University of Arizona, Hydrology and Atmospheric Sciences, Tucson, AZ, United States, (2)University of Arizona, Department of Hydrology and Atmospheric Sciences, Tucson, AZ, United States, (3)Hydrology and Atmospheric Sciences, The University of Arizona, Tucson, AZ, United States
 
Streamflow Modelling using Hybrid Long-Short Term Memory Network over the Indian Catchments (1140235)
Bhanu Magotra, Manabendra Saharia and Dhanya C.T, Indian Institute of Technology Delhi, Department of Civil Engineering, New Delhi, India
 
Imitation or Identification: On the Causal Consistency of Deep Learning to Extrapolate into Future Climate (1142355)
Alexander J Winkler1, Johannes Gensheimer2, Christian Reimers2, Albrecht Schall2, Reda El Ghawi2 and Basil Kraft3, (1)Max Planck Institute for Biogeochemistry, Department Biogeochemical Integration, Jena, Germany, (2)Max Planck Institute for Biogeochemistry, Jena, Germany, (3)Max-Planck-Institute for Biogeochemistry, Department of Biogeochemical Integration, Jena, Germany
 
Peak Anomaly Detection using Critical Zone Time Series Data: Knowledge-Engineering and Deep-Learning (1142799)
Byung Suk Lee, University of Vermont, Computer Science, Burlington, VT, United States, John Clay Kaufmann, University of Vermont, Computer Science, Burlington, United States, James B Shanley, USGS, Montpelier, VT, United States, Donna M Rizzo, University of Vermont, Department of Civil and Environmental Engineering, Burlington, VT, United States and Julia N Perdrial, University of Vermont, Department of Geography & Geoscience, Burlington, United States
 
Let me autodiff this thing: similarities and synergies between physical models and neural networks (Invited) (1143760)
Pierre Gentine, Columbia University of New York, Palisades, United States, Jianing Fang, Johns Hopkins University, Department of Earth and Planetary Sciences, Baltimore, United States and Markus Reichstein, Max Planck Institute for Biogeochemistry, Department Biogeochemical Integration, Jena, Germany
 
A high-quality deep-learning-based historical and near-real-time snow water equivalent product fusing satellite snow cover and in-situ data (1149170)
Wen-Ping Tsai1, Yalan Song2, Jiangtao Liu1 and Chaopeng Shen3, (1)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, PA, United States, (2)Pennsylvania State University Main Campus, University Park, PA, United States, (3)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States
 
An objective estimate of water stress - going beyond PDSI (1152282)
Wenli Zhao1,2, Alexander J Winkler3, Markus Reichstein3, Rene Orth2 and Pierre Gentine1, (1)Columbia University, Department of Earth and Environmental Engineering, New York, NY, United States, (2)Max Planck Institute for Biogeochemistry, Jena, Germany, (3)Max Planck Institute for Biogeochemistry, Department Biogeochemical Integration, Jena, Germany
 
Complexity-based robust hydrologic prediction: extension of statistical learning theory to conceptual hydrological models (1158039)
Saket Pande, Delft University of Technology, Delft, Netherlands
 
Deep Learning at Scale for Watershed Model Calibration (1159262)
Maruti Kumar Mudunuru1, Kyongho Son2, Peishi Jiang1, Glenn Hammond2 and Xingyuan Chen1, (1)Pacific Northwest National Laboratory, Richland, WA, United States, (2)Pacific Northwest National Laboratory, Richland, United States
 
Solving River Dynamics at the River-ocean Interface using a Physics-informed Deep Learning Based Data Assimilation Approach (1165366)
Dongyu Feng1, Zeli Tan1 and Qizhi He2, (1)Pacific Northwest National Laboratory, Richland, WA, United States, (2)University of Minnesota, Minneapolis, United States
 
Towards calibration of physics-based integrated surface-groundwater models using deep-learning based surrogates and simulation based inference (1166866)
Andrew Bennett1, Benjamin Horowitz2, Elena Leonarduzzi3, Yueling Ma4, Amanda Triplett5, Robert Hull6, Laura E Condon5, Reed M Maxwell7 and Peter Melchior8, (1)University of Arizona, Department of Hydrology & Atmospheric Sciences, Tucson, AZ, United States, (2)Princeton University, Department of Astronomy, Princeton, United States, (3)WSL Swiss Federal Institute for Forest, Snow and Landscape Research, Birmensdorf, Switzerland, (4)Princeton University, Princeton, United States, (5)University of Arizona, Hydrology and Atmospheric Sciences, Tucson, AZ, United States, (6)University of Arizona, Hydrology and Atmospheric Sciences, Tucson, United States, (7)Princeton University, Civil & Environmental Engineering, High Meadows Environmental Institute, Princeton, United States, (8)Princeton University, Astrophysical Sciences and the Center for Statistics and Machine Learning, Princeton, NJ, United States
 
One Model to Rule Them All: A Ubiquitous Data Modeling Approach to Improved Temporal-Spatial Learning of Physics Guided Machine Learning Models in Hydrology (1170303)
Aggrey Muhebwa, University of Massachusetts Amherst, Amherst, MA, United States, Colin J Gleason, University of Massachusetts Amherst, Civil & Environmental Engineering, Amherst, MA, United States, Dongmei Feng, University of Cincinnati, Cincinnati, United States and Jay Taneja, University of Massachusetts Amherst, Amherst, United States
 
The U.S. Geological Survey Flow Photo Explorer Tool: An Integrated Database and Machine Learning Platform for Estimating Streamflow from Timelapse Imagery (1173823)
Jennifer Fair1, Benjamin H Letcher2, Jeff Walker3, Amrita Gupta4, Phillip Goodling5, Todd L. Dubreuil2 and Michael J Hayden2, (1)USGS New England Water Science Center - Massachusetts Office, Northborough, MA, United States, (2)U.S. Geological Survey, Eastern Ecological Science Center, Turners Falls, MA, United States, (3)Walker Environmental Research, Brunswick, ME, United States, (4)Conservation Science Partners, Truckee, CA, United States, (5)U.S. Geological Survey, Maryland/Delaware/District of Columbia Water Science Center, Baltimore, MD, United States
 
Deep Learning-based Models for Estimating River Channel Width (1173960)
Mohammad Erfani, University of South Carolina, Civil and Environmental Engineering, Columbia, United States, Zahra Ghahremani, Boise State University, Department of Geoscience, Boise, United States, Laura Manuel, Tulane University of Louisiana, River Coastal Science and Engineering, New Orleans, United States, Shuyu Y Chang, Pennsylvania State University Main Campus, Geography, University Park, PA, United States, Erfan Goharian, University of South Carolina, Civil and Environmental Engineering, Columbia, SC, United States, Jennifer L Pierce, Boise State University, Boise, ID, United States, Ehab Meselhe, Tulane University, Department of River-Coastal Science and Engineering, New Orleans, LA, United States, Kimberly J Van Meter, The Pennsylvania State University, Department of Geography, State College, PA, United States, Chaopeng Shen, Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States and Sagy Cohen, University of Alabama, Geography, Tuscaloosa, United States
 
Improving Data-Driven Flow Forecasting in Large Basins using Machine Learning to Route Flows (1180053)
Mostafa Elkurdy1, Alden Keefe Sampson1 and Phillip Butcher2, (1)Upstream Tech, Alameda, United States, (2)Upstream Tech, Portland, United States
 
A Physics-guided Machine Learning Scheme for Predicting Peak Flow in Streams (1187599)
Sudan Pokharel, University of Nebraska Lincoln, Civil and Environmental Engineering, Lincoln, NE, United States, Tirthankar Roy, University of Arizona, Tucson, United States and David Admiraal, University of Nebraska Lincoln, Lincoln, NE, United States
 
From Parameter Learning to Structure Evolution: A Deep Integration of Physical Models and Deep Learning to Improve Hydrologic Modeling at Large Scales (1188753)
Dapeng Feng1, Chaopeng Shen1, Jiangtao Liu2 and Kathryn Lawson2, (1)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States, (2)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, PA, United States
 
Wetland Mapping in Large Spatial Scales via Machine Learning Approaches (1190625)
Bikas Chandra Gupta, Texas A&M University Kingsville, Kingsville, TX, United States and Adnan Rajib, Assistant Professor, Cincinnati, United States
See more of: Hydrology
<< Previous Session Proposal | Next Session Proposal >>