Physics-informed machine learning: case studies for weather and climate modelling

2021/02/15 by Karthik Kashinath, Mohamed Elhafiz Mustafa, Adrian Albert +18 · 42 citations
Earth and Planetary Sciences · Environmental Science · Physics and Astronomy · #Climate variability and models #Meteorological Phenomena and Simulations #Model Reduction and Neural Networks

paper · doi:10.1098/rsta.2020.0093

openalex publication_date 2021/02/15 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/25

Abstract

Machine learning (ML) provides novel and powerful ways of accurately and efficiently recognizing complex patterns, emulating nonlinear dynamics, and predicting the spatio-temporal evolution of weather and climate processes. Off-the-shelf ML models, however, do not necessarily obey the fundamental governing laws of physical systems, nor do they generalize well to scenarios on which they have not been trained. We survey systematic approaches to incorporating physics and domain knowledge into ML models and distill these approaches into broad categories. Through 10 case studies, we show how these approaches have been used successfully for emulating, downscaling, and forecasting weather and climate processes. The accomplishments of these studies include greater physical consistency, reduced training time, improved data efficiency, and better generalization. Finally, we synthesize the lessons learned and identify scientific, diagnostic, computational, and resource challenges for developing truly robust and reliable physics-informed ML models for weather and climate processes. This article is part of the theme issue 'Machine learning for weather and climate modelling'.

Citations

• Lagrangian Coherent Structures
• GENERAL CIRCULATION EXPERIMENTS WITH THE PRIMITIVE EQUATIONS
• Deep Learning Scaling is Predictable, Empirically
• Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
• Turbulence Modeling in the Age of Data
• Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Invariant and Equivariant Graph Networks
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• Making the Black Box More Transparent: Understanding the Physical Implications of Machine Learning
• Deep Learning for Physical Processes: Incorporating Prior Scientific Knowledge
• WeatherBench: A benchmark dataset for data-driven weather forecasting
• Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence
• Physics-Informed Generative Adversarial Networks for Stochastic Differential Equations
• Incorporating Symmetry into Deep Dynamics Models for Improved Generalization
• Neural Operator: Graph Kernel Network for Partial Differential Equations
• Building high accuracy emulators for scientific simulations with deep neural architecture search

Cited by

• Prediction of Ultraslow Magnetic Solitons via Plasmon-induced Transparency by Artificial Neural Networks
• Learning from Interval Targets
• Integrating AI and climate change scenarios for multi-risk assessment in the coastal municipalities of the Veneto region
• Spectropolarimetric Inversion in Four Dimensions with Deep Learning (SPIn4D): II. A Physics-Informed Machine Learning Method for 3D Solar Photosphere Reconstruction
• ADAPT: Lightweight, Long-Range Machine Learning Force Fields Without Graphs
• Climate-Adaptive and Cascade-Constrained Machine Learning Prediction for Sea Surface Height under Greenhouse Warming
• CERA: A Framework for Improved Generalization of Machine Learning Models to Changed Climates
• Scientific Machine Learning Through Physics–Informed Neural Networks: Where we are and What’s Next
• Diffeomorphic Neural Operator Learning
• Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems
• Enhancing material behavior discovery using embedding-oriented Physically-Guided Neural Networks with Internal Variables
• Forecasting chaotic dynamic using hybrid system
• Machine Learning Applications for Fisheries—At Scales from Genomics to Ecosystems
• Synergistic Integration of Flood Inundation Modeling Methods: A Review of Computational, Data‐Driven, Observational and Experimental, and Conceptual Models
• AdjointNet: Constraining machine learning models with physics-based codes
• Physics-Informed Regression Modelling for Vertical Facade Surface Temperature: A Tropical Case Study on Solar-reflective Material
• Tackling Climate Change with Machine Learning
• PGT-I: Scaling Spatiotemporal GNNs with Memory-Efficient Distributed Training
• Physics-Informed Neural Networks with Hard Nonlinear Equality and Inequality Constraints
• Machine Learning in Acoustics: A Review and Open-Source Repository
• Transparent Machine Learning: Training and Refining an Explainable Boosting Machine to Identify Overshooting Tops in Satellite Imagery
• Imposing the Fundamental Dynamical Constraint of Hydrostatic Balance to Improve Global ML Weather Prediction
• Stochastic and Non-local Closure Modeling for Nonlinear Dynamical Systems via Latent Score-based Generative Models
• Efficient inference and learning of a generative model for ENSO predictions from large multi-model datasets
• Artificial Intelligence for Atmospheric Sciences: A Research Roadmap
• A Gravity-informed Spatiotemporal Transformer for Human Activity Intensity Prediction
• VideoPDE: Unified Generative PDE Solving via Video Inpainting Diffusion Models
• Physics-Informed Teleconnection-Aware Transformer for Global Subseasonal-to-Seasonal Forecasting
• Latent Representation Learning of Multi-scale Thermophysics: Application to Dynamics in Shocked Porous Energetic Material
• An efficient and physics-informed regional maize yield estimation scheme by combining data assimilation and machine learning
• Multi-Exit Kolmogorov-Arnold Networks: enhancing accuracy and parsimony
• Toward Knowledge-Guided AI for Inverse Design in Manufacturing: A Perspective on Domain, Physics, and Human-AI Synergy
• Strictly Constrained Generative Modeling via Split Augmented Langevin Sampling
• Improving Coverage in Combined Prediction Sets with Weighted p-values
• Understanding the Capabilities of Molecular Graph Neural Networks in Materials Science Through Multimodal Learning and Physical Context Encoding
• GAIA: A Foundation Model for Operational Atmospheric Dynamics
• Predicting Beyond Training Data via Extrapolation versus Translocation: AI Weather Models and Dubai's Unprecedented 2024 Rainfall
• LatticeVision: Image to Image Networks for Modeling Non-Stationary Spatial Data
• Generalizable neural-network parameterization of mesoscale eddies in idealized and global ocean models
• Mask-PINNs: Mitigating Internal Covariate Shift in Physics-Informed Neural Networks
• Monotone Peridynamic Neural Operator for Nonlinear Material Modeling with Conditionally Unique Solutions
• Artificial intelligence and downscaling global climate model future projections
• Predicting Water Temperature Dynamics of Unmonitored Lakes With Meta‐Transfer Learning
• Hallucination, reliability, and the role of generative AI in science

Related

• Machine learning for climate physics and simulations
• Machine Learning for the Physics of Climate
• Machine Learning Workflows in Climate Modeling: Design Patterns and Insights from Case Studies
• Climate-Invariant Machine Learning
• Leveraging machine learning to enhance climate models: a review