🔥 Deep learning has been widely explored almost in all research filed. Here is a curated list of papers about deep learning methods in Earth System, especially relating to weather prediction. It also contains frameworks for data-driven Numerical Weather Prediciton (NWP) training, tools to deploy weather prediction, courses and tutorials about Ai4earth and all publicly available weather prediction checkpoints and APIs.
- Awesome-ai4earth
-
- Extreme Weather and Prediction
tropical cyclones, heat wave etc.
- Climate phenomena analysis
- Extreme Weather and Prediction
- [2023-11-07] Creat this project, add some big model for atmospherical modeling!
- Add more paper, datasets, research directions for ai4earth ✨Contributions Wanted
Also check out the project that I am currently working on: EarthVision - A deep learniong framwork for Numerical Weather Prediction, Earth System Data cmpression, Precipitation Prediction)
Date | keywords | Institute | Paper | Publication |
---|---|---|---|---|
2022-02 | FourCastNet | Nvidia | FourCastNet: Accelerating Global High-Resolution Weather Forecasting Using Adaptive Fourier Neural Operators [paper1][paper2] | PASC |
2022-11 | PanguWeather | HuaWei | Accurate medium-range global weather forecasting with 3D neural networks [paper1][paper2] | Nature |
2022-12 | GraphCast | DeepMind | GraphCast: Learning skillful medium-range global weather forecasting | Science |
2023-04 | FengWu | Shanghai AILab | FengWu: Pushing the Skillful Global Medium-range Weather Forecast beyond 10 Days Lead | arxiv |
2023-06 | FuXi | Fudan University | FuXi: A cascade machine learning forecasting system for 15-day global weather forecast | npj |
- FengWu-Adas Towards an End-to-End Artificial Intelligence Driven Global Weather Forecasting System. (arxiv, 12/2023)
- FengWu-4DVar FengWu-4DVar: Coupling the Data-driven Weather Forecasting Model with 4D Variational Assimilation (arxiv, 12/2023)
- NeuralGCM Neural General Circulation Models. (arxiv, 11/2023)
- GenCast Diffusion-based ensemble forecasting for medium-range weather. (arxiv. 12/2023)
- MetNet-3 Deep Learning for Day Forecasts from Sparse Observations. (arxiv, 06/2023)
- ClimaX A foundation model for weather and climate, (ICML, 2023) [project]
- NowcastNet Skilful nowcasting of extreme precipitation with NowcastNet. (nature, 07/2023)
- PostRainBench - A comprehensive benchmark and a new model for precipitation forecasting. (arxiv, 10/2023)
- Anthropogenic fingerprints in daily precipitation revealed by deep learning. (nature, 08/2023)
- CorrDiff Generative residual diffusion modeling for km-scale atmospheric downscaling, (arxiv, 09/2023)
- Two deep learning-based bias-correction pathways improve summer precipitation prediction over China, (Environmental Research Letters, 12/2022) paper
- FuXi-Extreme: Improving extreme rainfall and wind forecasts with diffusion mode. (arxiv, 10/2023)
- The Potential Predictability of Fire Danger Provided by Numerical Weather Prediction (JAMC, 11/2016)
- Machine learning–based observation-constrained projections reveal elevated global socioeconomic risks from wildfire. (nature communication, 22/03/2023)
The following list makes sure that all weather forecasting models are compared apples to apples.
WeatherBench 2: A benchmark for the next generation of data-driven global weather models [project] [LeaderBoard] (08/2023)
You may also find these leaderboards helpful:
- WeatherBench2 Leaderboard - evaluating and comparing various weather forecasting models. displays up-to-date scores of many of the state-of-the-art ML and physics-based models.
If you're interested in the field of AI4Earth, you may find the above list of milestone papers helpful to explore its history and state-of-the-art. However, each direction of AI4Earth offers a unique set of insights and contributions, which are essential to understanding the field as a whole. For a detailed list of papers in various subfields, please refer to the following link (it is possible that there are overlaps between different subfields):
(:exclamation: We would greatly appreciate and welcome your contribution to the following list. ❗)
-
Analyse different NWP models in different fields with respect to different abilities
-
some deeplearning methods in wildfire forecast.
- Era5 single-level - ERA5 is the fifth generation ECMWF reanalysis for the global climate and weather for the past 8 decades. Data is available from 1940 onwards.
- Era5 pressure-level - ERA5 is the fifth generation ECMWF reanalysis for the global climate and weather for the past 8 decades. Data is available from 1940 onwards.
- Era5-Land - ERA5-Land is a reanalysis dataset providing a consistent view of the evolution of land variables over several decades at an enhanced resolution compared to ERA5
Model | Size | Architecture | Access | Date | Origin | Model License1 |
---|---|---|---|---|---|---|
PanguWeather | ~740M | 3D Transformer | ckpt | 2022-11 | Paper | BY-NC-SA 4.0 |
GraphCast | unkonwn | GNN | ckpt | 2022-12 | Paper | Apache 2.0 & BY-NC-SA 4.0 |
FengWu | unkonwn | U-Transformer | ckpt | 2023-06 | Paper | - |
FuXi | unkonwn | Transformer | ckpt | 2023-04 | Paper | - |
- DeepSpeed - DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
- Megatron-DeepSpeed - DeepSpeed version of NVIDIA's Megatron-LM that adds additional support for several features such as MoE model training, Curriculum Learning, 3D Parallelism, and others.
- FairScale - FairScale is a PyTorch extension library for high performance and large scale training.
- Megatron-LM - Ongoing research training transformer models at scale.
- Colossal-AI - Making large AI models cheaper, faster, and more accessible.
- BMTrain - Efficient Training for Big Models.
- Mesh Tensorflow - Mesh TensorFlow: Model Parallelism Made Easier.
- maxtext - A simple, performant and scalable Jax LLM!
- Alpa - Alpa is a system for training and serving large-scale neural networks.
- GPT-NeoX - An implementation of model parallel autoregressive transformers on GPUs, based on the DeepSpeed library.
- Fluid-Earth - An interactive web application that allows you to visualize current and past conditions of Earth's atmosphere and oceans.
- TODO - A curated (still actively updated) list of practical guide resources of earth-related research.
This is an active repository and your contributions are always welcome!
I will keep some pull requests open if I'm not sure if they are awesome for ai4earth, you could vote for them by adding 👍 to them.
If you have any question about this opinionated list, do not hesitate to contact me hantao10200@gmail.com.
Footnotes
-
This is not legal advice. Please contact the original authors of the models for more information. ↩