MetNet: A Neural Weather Model for Precipitation Forecasting

By Casper Kaae Sønderby, Lasse Espeholt, Jonathan Heek, Mostafa Dehghani, Avital Oliver, Tim Salimans, Shreya Agrawal, Jason Hickey, Nal Kalchbrenner. 2020

361 citations recorded by the source corpus.

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 km$^2$ 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.

Subfields: Artificial Intelligence, Atmospheric Science, Computer Vision and Pattern Recognition, Environmental Engineering, AI for Physical Sciences

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yoshua Bengio
2. Quoc V. Le
3. Sergey Levine
4. Ruslan Salakhutdinov
5. Mohit Bansal
6. Andrew Zisserman
7. Kyunghyun Cho
8. Jianfeng Gao

In Atmospheric Science

1. Alberto Arribas
2. Shakir Mohamed
3. Kevin Trebing
4. Nal Kalchbrenner
5. Siamak Mehrkanoon

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Andrew Zisserman
3. Jitendra Malik
4. Ming-Hsuan Yang
5. Dacheng Tao
6. Trevor Darrell
7. Xiaogang Wang
8. Andrea Vedaldi

In Environmental Engineering

1. Alberto Arribas
2. Shakir Mohamed
3. Andrew Brock

In AI for Physical Sciences

1. Piotr Wojciech Mirowski
2. Shakir Mohamed
3. Raia Hadsell

Citations in the corpus

• VideoGPT: Video Generation using VQ-VAE and Transformers (2021, neutral): “Video prediction (Srivastava et al., 2015; Finn et al., 2016; Kalchbrenner et al., 2017; Sønderby et al., 2020) can be seen as learning a generative model of future frames conditioned on the past frames.”
• Skilful precipitation nowcasting using deep generative models of radar (2021, methodology): “Finally, [6] use a ConvLSTM encoder and axial attention decoder to perform eight hour pointwise probabilistic prediction by sliding a 64 × 64 window over a much larger area covering the United States.”
• ClimaX: A foundation model for weather and climate (2023, neutral): “Rasp & Thuerey (2021) were the first to use pretraining on climate simulations to achieve good data-driven medium-range weather prediction with a ResNet (He et al., 2016), Weyn et al. (2020) used CNNs on a cubed sphere for global weather prediction, Weyn et al. (2021) forecast weather sub-seasonally”
• SmaAt-UNet: Precipitation Nowcasting using a Small Attention-UNet Architecture (2020, background): “Recent advances in artificial neural network architectures (ANNs) have enabled data-driven based models to bridge the present gap for short-term forecasting [6, 7, 8, 9].”
• Skillful Twelve Hour Precipitation Forecasts using Large Context Neural Networks (2021, methodology): “Furthermore, simple geographical features for longitude, latitude, elevation [27] and temporal features that encode the forecast time t0 are additional sources of inputs for the MetNet-2 models.”