H078
Physics-Based and Hybrid (Physics ML) Modeling for Watershed Function

Session ID#: 159767

Session Description:
As Earth system change places increasing stress on the Nation’s water resources, we can no longer rely on historical hydrological and biogeochemical trends and simplified models to guide our understanding of the critical watershed functions that control downstream water supply and quality. Thus, it is incumbent on us to develop a new generation of predictive tools that can address the natural heterogeneity of watershed systems without sacrifice of process fidelity.  These processes include integrated surface and subsurface flow, vegetation-related processes including evapo-transpiration and nutrient uptake, bedrock weathering and soil formation, and biogeochemical processes associated with floodplains and the hyporheic zone.

High performance computing (HPC) for watershed hydrological and biogeochemical function is making high resolution simulations possible over considerably larger spatial extents than was possible in the past.  However, HPC alone is unlikely to be able to scale to full river basin scales, so artificial intelligence (AI) and machine learning (ML) are likely to have an important role to play in making larger scale and longer time duration simulations possible.  Surrogate models for watershed function, for example, can be developed using training on HPC modeling, which can provide much more data at higher resolution than is possible with synoptic observations alone.  In addition, the use of physics-based modeling also provides the opportunity to consider future climate scenarios that may not be present in the historical record (i.e., based on observations alone).

Index Terms:

1804 Catchment [HYDROLOGY]
1836 Hydrological cycles and budgets [HYDROLOGY]
1860 Streamflow [HYDROLOGY]
1879 Watershed [HYDROLOGY]

Primary Convener:  Carl I Steefel, Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, CA, United States
Conveners:  Scott L Painter, Oak Ridge National Laboratory, Watershed Systems Modeling Group, Environmental Sciences Division, Oak Ridge, United States and John D Moulton, Los Alamos National Laboratory, Los Alamos, NM, United States
 
Coupling the Soil and Water Assessment Tool (SWAT) with machine learning algorithms for water quality prediction (1046348)
Yu Li, University of Wisconsin Madison, Madison, WI, United States, Anita Thompson, University of Wisconsin Madison, Madison, WI, UNITED STATES and Zhou Zhang, University of Wisconsin-Madison, Biological Systems Engineering, Madison, United States
 
Learning the streamflow spatial and temporal dynamics through AI/ML (Invited) (1056231)
Alexander Y Sun1, Peishi Jiang2, Xingyuan Chen2 and Zong-Liang Yang3, (1)University of Texas at Austin, Austin, TX, United States, (2)Pacific Northwest National Laboratory, Richland, WA, United States, (3)Univ Texas Austin, Austin, United States
 
Machine learning aided reservoir regulation in a physically based hydrological model (1069590)
Pallav Kumar Shrestha1, Luis Samaniego2, Oldrich Rakovec1, Rohini Kumar2, Chenxi Mi3, Karsten Rinke4 and Stephan Thober1, (1)Helmholtz Centre for Environmental Research - UFZ, Computational Hydrosystems, Leipzig, Germany, (2)Helmholtz Centre for Environmental Research GmbH – UFZ, Computational Hydrosystems, Leipzig, Germany, (3)Helmholtz Center for Environmental Research UFZ Magdeburg, Department of Lake Research, Magdeburg, Germany, (4)Helmholtz Centre for Environmental Research UFZ Magdeburg, Department of Lake Research, Magdeburg, Germany
 
Physics-informed surrogate modeling for supporting climate resilience at groundwater contamination sites (1097258)
Lijing Wang1, Takuya Kurihana2, Aurelien Meray3, Ilijana Mastilovic4, Satyarth Praveen5, Zexuan Xu6, Milad Memarzadeh7, Alexander Lavin8 and Haruko M Wainwright6, (1)Stanford University, Department of Geological Sciences, Stanford, CA, United States, (2)University of Chicago, Computer Science, Chicago, IL, United States, (3)Florida International University, Applied Research Center, Miami, FL, United States, (4)University of Wisconsin Milwaukee, School of Freshwater Science, Milwaukee, United States, (5)Lawrence Berkeley National Laboratory, Berkeley, United States, (6)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (7)NASA Ames Research Center, Moffett Field, United States, (8)Pasteur & ISI, New York, United States
 
Differentiable modeling in Geosciences - Breaking down the imaginary barrier between machine learning and process-based modeling with differentiable modeling (Invited) (1117486)
Chaopeng Shen1, Hoshin Vijai Gupta2, Toshiyuki Bandai3, Daniel Kifer4, Alexandre M Tartakovsky5 and Dapeng Feng1, (1)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States, (2)University of Arizona, Department of Hydrology and Atmospheric Sciences, Tucson, AZ, United States, (3)University of Tokyo, Bunkyo-ku, Japan, (4)Pennsylvania State University, Department of Computer Science & Engineering, University Park, United States, (5)University of Illinois at Urbana Champaign, Civil and Environmental Engineering, Champaign, United States
 
Simulating Operation Behaviors of Cascade Reservoirs Using Physics-Based Machine Learning Models: A Case Study for Gunnison River Basin (1127161)
Bilal Iftikhar1, Sudershan Gangrade2, Dan Lu3, Shih-Chieh Kao2, Scott L Painter4 and Ethan Coon5, (1)Oak Ridge National Laboratory, Environmental Sciences Division, Oak Ridge, United States, (2)Oak Ridge National Laboratory, Environmental Sciences Division, Oak Ridge, TN, United States, (3)Oak Ridge National Laboratory, Oak Ridge, TN, United States, (4)Los Alamos National Laboratory, Los Alamos, United States, (5)Oak Ridge National Laboratory, Watershed Systems Modeling Group, Environmental Sciences Division, Oak Ridge, United States
 
Simulating nitrogen processes in a mountainous watershed using an integrated hydrology and reactive transport model (1131710)
Zexuan Xu1, Sergi Molins2, Lucien Stolze3, Dipankar Dwivedi4, Ilhan Özgen5, Bhavna Arora4, Qina Yan6, Craig Ulrich7, Nicola Falco8, John D Moulton9 and Carl I Steefel8, (1)Lawrence Berkeley National Laboratory, Climate and Ecosystem Sciences Division, Berkeley, CA, United States, (2)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences, Berkeley, CA, United States, (3)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area,, Berkeley, United States, (4)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, United States, (5)Technische Universität Braunschweig, Braunschweig, Germany, (6)University of Illinois at Urbana Champaign, Urbana, IL, United States, (7)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (8)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, CA, United States, (9)Los Alamos National Laboratory, Los Alamos, NM, United States
 
Exploring Interpretability and Performance of an Attention-based Long Short-Term Memory (LSTM) Network for Rainfall-runoff Modeling (1152490)
Soumendra Nath Bhanja1, Dan Lu2, Ethan Coon3 and Scott L Painter3, (1)Oak Ridge National Laboratory, Watershed Systems Modeling Group, Environmental Sciences Division, Oak Ridge, TN, United States, (2)Oak Ridge National Laboratory, Oak Ridge, TN, United States, (3)Oak Ridge National Laboratory, Watershed Systems Modeling Group, Environmental Sciences Division, Oak Ridge, United States
 
Machine Learning Modeling of Snowpack: Combining the Influence of Meteorological Drivers and Surface Properties (1169670)
Utkarsh Mital1, Dipankar Dwivedi1, Ilhan Özgen-Xian1,2, James Bentley Brown3 and Carl I Steefel4, (1)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, United States, (2)Technical University of Braunschweig, Braunschweig, Germany, (3)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (4)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, CA, United States
 
Improving Large-Basin Streamflow Simulation Using a Differentiable, Learnable Routing Model (1174764)
Tadd Bindas1, Chaopeng Shen2, Wen-Ping Tsai2, Jiangtao Liu1, Farshid Rahmani1, Dapeng Feng2 and Yuchen Bian3, (1)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, PA, United States, (2)Pennsylvania State University Main Campus, Department of Civil and Environmental Engineering, University Park, United States, (3)Baidu Research, USA, Sunnyvale, CA, United States
 
WaterNet: a process-based and explainable deep learning framework modeling stream discharge (1185978)
Kuai Fang, Stanford University, Earth System Sciences, Stanford, CA, United States, Zach Perzan, Stanford University, Earth Systems Science, Stanford, CA, United States and Kate Maher, Stanford University, Earth Systems Science, Stanford, United States
 
Upscaling strategy for watershed function using a transfer learning framework (1187842)
Dipankar Dwivedi1, Utkarsh Mital1, Ilhan Özgen-Xian1, Pin Shuai2, Xingyuan Chen2, James Bentley Brown3 and Carl I Steefel4, (1)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, United States, (2)Pacific Northwest National Laboratory, Richland, WA, United States, (3)Lawrence Berkeley National Laboratory, Berkeley, CA, United States, (4)Lawrence Berkeley National Laboratory, Earth and Environmental Sciences Area, Berkeley, CA, United States
 
Spatiotemporal Interpolation of River Discharge using Gauge Observation via Machine Learning (1198323)
Tatsuya Ishikawa, IBM, Research - Tokyo, Tokyo, Jordan, Michiaki Tatsubori, IBM, Research - Tokyo, Oiso, Naka-gun, Japan and Refaldi Intri Dwi Putra, IBM, Research - Tokyo, Tokyo, Japan
See more of: Hydrology
<< Previous Session Proposal | Next Session Proposal >>