MetNet: A Neural Weather Model for Precipitation Forecasting

2020/03/24 by Casper Kaae Sønderby, Sønderby, Casper Kaae, Lasse Espeholt +15 · 21 citations
Earth and Planetary Sciences · Environmental Science · Computer Science · #Meteorological Phenomena and Simulations #Hydrological Forecasting Using AI #Computational Physics and Python Applications

paper · pdf · doi:10.48550/arxiv.2003.12140

Abstract

Weather forecasting is a long standing scientific challenge with direct social and economic impact. The task is suitable for deep neural networks due to vast amounts of continuously collected data and a rich spatial and temporal structure that presents long range dependencies. We introduce MetNet, a neural network that forecasts precipitation up to 8 hours into the future at the high spatial resolution of 1 km² and at the temporal resolution of 2 minutes with a latency in the order of seconds. MetNet takes as input radar and satellite data and forecast lead time and produces a probabilistic precipitation map. The architecture uses axial self-attention to aggregate the global context from a large input patch corresponding to a million square kilometers. We evaluate the performance of MetNet at various precipitation thresholds and find that MetNet outperforms Numerical Weather Prediction at forecasts of up to 7 to 8 hours on the scale of the continental United States.

Citations

• A North American Hourly Assimilation and Model Forecast Cycle: The Rapid Refresh

Cited by

• Long-Range Distillation: Distilling 10,000 Years of Simulated Climate into Long Timestep AI Weather Models
• Huayu: Advanced Real-Time Precipitation Estimation from Geostationary Satellite
• A Diffusion-Based Framework for High-Resolution Precipitation Forecasting over CONUS
• A curated UK rain radar data set for training and benchmarking nowcasting models
• FuXi-Nowcast: Environment-conditioned deep learning for severe convection nowcasting
• DINOv3 as a Frozen Encoder for CRPS-Oriented Probabilistic Rainfall Nowcasting
• Variational Data Assimilation with a Learned Inverse Observation Operator
• Skilful precipitation nowcasting using deep generative models of radar
• VideoGPT: Video Generation using VQ-VAE and Transformers
• Video Prediction of Dynamic Physical Simulations With Pixel-Space Spatiotemporal Transformers
• RainBench: Towards Global Precipitation Forecasting from Satellite Imagery
• A Weather Foundation Model for the Power Grid
• DPSformer: A long-tail-aware model for improving heavy rainfall prediction
• Revisiting the Calibration of Modern Neural Networks
• Accurate medium-range global weather forecasting with 3D neural networks
• Deep multi-stations weather forecasting: explainable recurrent convolutional neural networks
• Generative Models as Distributions of Functions
• A review of radar-based nowcasting of precipitation and applicable machine learning techniques
• IndiaWeatherBench: A Dataset and Benchmark for Data-Driven Regional Weather Forecasting over India
• The AI gambit: leveraging artificial intelligence to combat climate change—opportunities, challenges, and recommendations
• Observation-guided Interpolation Using Graph Neural Networks for High-Resolution Nowcasting in Switzerland

Related

• Skillful Twelve Hour Precipitation Forecasts using Large Context Neural Networks
• Deep Learning for Day Forecasts from Sparse Observations
• An Operational Deep Learning System for Satellite-Based High-Resolution Global Nowcasting
• WeatherFusionNet: Predicting Precipitation from Satellite Data
• Technical overview and architecture of the FastNet Machine Learning weather prediction model, version 1.0