Guest Speaker

Dr. Chaopeng Shen, Associate Professor in Civil Engineering at Pennsylvania State University

Bio

Dr. Chaopeng Shen is an Associate Professor in Civil Engineering at The Pennsylvania State University. He received a Ph.D. in environmental engineering from Michigan State University, East Lansing, MI, USA, in 2009. His PhD research focused on computational hydrology, and he developed the hydrologic model Process-based Adaptive Watershed Simulator (PAWS), which was later coupled to the community land model to study the interactions between hydrology and ecosystem. He was a Post-Doctoral Research Associate with the Lawrence Berkeley National Laboratory, Berkeley, CA, USA, from 2011 to 2012, working on high-performance computational geophysics. His recent efforts focused on harnessing the big data and machine learning (ML) opportunities in advancing hydrologic predictions and understanding. As an early advocate for machine learning in geosciences, he promotes differentiable modeling to integrate ML and physics for knowledge discovery and improved modeling of climate change impacts. He has written technical, editorial, review, and collective opinion papers on hydrologic deep learning to highlight emerging opportunities for scientific advances. In addition, his research interests also include floodplain systems, scaling issues, process-based hydrologic modeling, and hydrologic data mining. He is currently an Editor of the Journal of Geophysical Research – Machine Learning and Computation, an Associate Editor of the Water Resources Research, and Chief Special Editor for Frontiers in AI: Water and AI.

Presentation Title

Differentiable modeling for continental and global-scale water sciences

Presentation Abstract

Process-based modeling offers interpretability and physical consistency in many domains of geosciences but struggles to leverage large datasets efficiently. Machine-learning methods, especially deep networks, have strong predictive skills yet are unable to answer specific scientific questions or elucidate physical processes. A recently proposed genre of physics-informed machine learning, called “differentiable” models (https://t.co/qyuAzYPA6Y), connect neural networks (NNs) with process-based equations (priors) to benefit from the best of both NNs and process-based modeling paradigms. We propose that differentiable models are especially suitable for flood forecasting as well as continental- or global-scale water sciences. They can harvest information from big earth observations to produce state-of-the-art predictions that supersede purely data-driven models, enable physical interpretation naturally, extrapolate well (due to physical constraints) in space and time, and leverage progress in modern AI computing architecture and infrastructure. Differentiable models can also synergize with existing hydrologic models, learn from the lessons of the community and distinguish better priors. We demonstrate the power of differentiable modeling using computational examples in rainfall-runoff modeling, flood forecasting, river routing, as well applications in water-related domains such as ecosystem modeling and water quality modeling. Furthermore, we show how differentiable modeling can enable us to ask fundamental questions in water sciences, ecohydrology, and water quality.