Data

Tools

Indexes

Applications

Experiments

Open Sustainable Technology

Graph Weather

Data-driven approach for forecasting global weather using graph neural network.
https://github.com/openclimatefix/graph_weather

Category: Atmosphere
Sub Category: Meteorological Observation and Forecast

Keywords

forecasting-models graph-neural-networks pytorch weather

Keywords from Contributors

nowcasting pvoutput solar eumetsat gan nowcasting-models nowcasting-precipitation pytorch-implementation pytorch-lightning

Last synced: about 23 hours ago
JSON representation

Repository metadata

Graph-based weather forecasting models. Originally, PyTorch implementation of Ryan Keisler's 2022 "Forecasting Global Weather with Graph Neural Networks" paper (https://arxiv.org/abs/2202.07575)

README.md

Graph Weather

All Contributors

This repo implements graph neural networks for weather forecasting, originally an implementation of
the Graph Weather paper (https://arxiv.org/pdf/2202.07575.pdf) in PyTorch. Additionally, multiple other
models have now been added, as well as general models for assimilation and forecasting.

The models implemented include:

DeepMind's Functional Generative Network (FGN) for probablistic ensemble forecasting

DeepMind's GenCast for graph diffusion-based forecasting

WindBorne's WeatherMesh-3 for highly efficient forecasting with Neighborhood Attention

Microsoft's Aurora forecasting model

And FengWu-GHR forecasting, using LoRA to correct for per-forecast step errors

The components of these models should be fairly modular and be able to be swapped around to experiment with graph-based weather forecasting.

Installation

This library can be installed through

pip install graph-weather

Alternatively, you can install the latest version from the repository easily with pixi:

pixi install # `-e cuda` for GPU support, `-e cpu` for CPU-only

Example Usage

The models generate the graphs internally, so the only thing that needs to be passed to the model is the node features
in the same order as the lat_lons.

import torch
from graph_weather import GraphWeatherForecaster
from graph_weather.models.losses import NormalizedMSELoss

lat_lons = []
for lat in range(-90, 90, 1):
    for lon in range(0, 360, 1):
        lat_lons.append((lat, lon))
model = GraphWeatherForecaster(lat_lons)

# Generate 78 random features + 24 non-NWP features (i.e. landsea mask)
features = torch.randn((2, len(lat_lons), 102))

target = torch.randn((2, len(lat_lons), 78))
out = model(features)

criterion = NormalizedMSELoss(lat_lons=lat_lons, feature_variance=torch.randn((78,)))
loss = criterion(out, target)
loss.backward()

And for the assimilation model, which assumes each lat/lon point also has a height above ground, and each observation
is a single value + the relative time. The assimlation model also assumes the desired output grid is given to it as
well.

import torch
import numpy as np
from graph_weather import GraphWeatherAssimilator
from graph_weather.models.losses import NormalizedMSELoss

obs_lat_lons = []
for lat in range(-90, 90, 7):
    for lon in range(0, 180, 6):
        obs_lat_lons.append((lat, lon, np.random.random(1)))
    for lon in 360 * np.random.random(100):
        obs_lat_lons.append((lat, lon, np.random.random(1)))

output_lat_lons = []
for lat in range(-90, 90, 5):
    for lon in range(0, 360, 5):
        output_lat_lons.append((lat, lon))
model = GraphWeatherAssimilator(output_lat_lons=output_lat_lons, analysis_dim=24)

features = torch.randn((1, len(obs_lat_lons), 2))
lat_lon_heights = torch.tensor(obs_lat_lons)
out = model(features, lat_lon_heights)
assert not torch.isnan(out).all()
assert out.size() == (1, len(output_lat_lons), 24)

criterion = torch.nn.MSELoss()
loss = criterion(out, torch.randn((1, len(output_lat_lons), 24)))
loss.backward()

Pretrained Weights

Coming soon! We plan to train a model on GFS 0.25 degree operational forecasts, as well as MetOffice NWP forecasts.
We also plan trying out adaptive meshes, and predicting future satellite imagery as well.

Training Data

Training data will be available through HuggingFace Datasets for the GFS forecasts. The initial set of data is available for GFSv16 forecasts, raw observations, and FNL Analysis files from 2016 to 2022, and for ERA5 Reanlaysis. MetOffice NWP forecasts we cannot
redistribute, but can be accessed through CEDA.

Contributors ✨

Thanks goes to these wonderful people (emoji key):

This project follows the all-contributors specification. Contributions of any kind welcome!

Owner metadata

GitHub Events

Total
Last Year

Committers metadata

Last synced: 11 days ago

Total Commits: 282
Total Committers: 20
Avg Commits per committer: 14.1
Development Distribution Score (DDS): 0.528

Commits in past year: 79
Committers in past year: 11
Avg Commits per committer in past year: 7.182
Development Distribution Score (DDS) in past year: 0.506

Name Email Commits
BumpVersion Action b****n@g****s 133
Jacob Bieker j****b@b****h 83
allcontributors[bot] 4****]@u****m 22
pre-commit-ci[bot] 6****]@u****m 11
gbruno16 7****6@u****m 7
Lorenzo Breschi 5****d@u****m 4
Yuvraaj Narula 4****a@u****m 4
KAITE a****2@g****m 3
Aavash Subedi 8****i@u****m 2
Indranil Biswas i****7@g****m 2
Rahul Maurya 9****b@u****m 2
Ananya Kulkarni 1****k@u****m 1
Aniket Shaha 6****5@u****m 1
Francesco 3****a@u****m 1
Jacob Prince-Bieker j****b@j****i 1
Ramkumar R r****l@g****m 1
Vishal Gaur 3****t@u****m 1
Wendoom-dev i****a@g****m 1
assafshouval a****l@g****m 1
peterdudfield p****d@h****m 1

Committer domains:

Issue and Pull Request metadata

Last synced: 6 days ago

Total issues: 95
Total pull requests: 113
Average time to close issues: 4 months
Average time to close pull requests: 26 days
Total issue authors: 24
Total pull request authors: 24
Average comments per issue: 3.05
Average comments per pull request: 0.96
Merged pull request: 82
Bot issues: 1
Bot pull requests: 31

Past year issues: 21
Past year pull requests: 40
Past year average time to close issues: 26 days
Past year average time to close pull requests: about 1 month
Past year issue authors: 4
Past year pull request authors: 12
Past year average comments per issue: 3.57
Past year average comments per pull request: 1.78
Past year merged pull request: 19
Past year bot issues: 0
Past year bot pull requests: 4

More stats: https://issues.ecosyste.ms/repositories/lookup?url=https://github.com/openclimatefix/graph_weather

Top Issue Authors

  • jacobbieker (62)
  • paapu88 (4)
  • gbruno16 (3)
  • byphilipp (2)
  • vitusbenson (2)
  • JackKelly (2)
  • taeyoon91 (2)
  • rahul-maurya11b (2)
  • SauryChen (1)
  • Wonderdch (1)
  • allcontributors[bot] (1)
  • DarkSlice1 (1)
  • maziheng31 (1)
  • marvingabler (1)
  • ammoniaca (1)

Top Pull Request Authors

  • pre-commit-ci[bot] (19)
  • yuvraajnarula (17)
  • jacobbieker (14)
  • gbruno16 (14)
  • allcontributors[bot] (12)
  • rnwzd (8)
  • rahul-maurya11b (3)
  • aavashsubedi (3)
  • aniket2405 (2)
  • Wendoom-dev (2)
  • peterdudfield (2)
  • AswaniSahoo (2)
  • praj-tarun (2)
  • ACSE-vg822 (2)
  • 0xFrama (2)

Top Issue Labels

  • enhancement (64)
  • good first issue (24)
  • bug (14)
  • help wanted (4)
  • size-big (1)

Top Pull Request Labels

  • enhancement (5)
  • size-small (2)
  • documentation (1)
  • size-big (1)

Package metadata

proxy.golang.org: github.com/openclimatefix/graph_weather

pypi.org: graph-weather

Weather Forecasting with Graph Neural Networks

  • Homepage: https://github.com/openclimatefix/graph_weather
  • Documentation: https://graph-weather.readthedocs.io/
  • Licenses: MIT License
  • Latest release: 1.0.94 (published 11 months ago)
  • Last Synced: 2025-12-21T22:09:07.979Z (4 days ago)
  • Versions: 98
  • Dependent Packages: 0
  • Dependent Repositories: 1
  • Downloads: 270 Last month
  • Rankings:
    • Stargazers count: 6.814%
    • Dependent packages count: 7.306%
    • Forks count: 8.016%
    • Average: 14.844%
    • Dependent repos count: 22.077%
    • Downloads: 30.008%
  • Maintainers (2)

Dependencies

requirements.txt pypi
  • datasets *
  • einops *
  • h3 *
  • huggingface-hub *
  • torch *
  • torch-geometric *
.github/workflows/workflows.yaml actions
  • actions/checkout v2 composite
  • actions/setup-python v2 composite
  • codecov/codecov-action v2 composite
Dockerfile docker
  • ubuntu latest build
.github/workflows/release.yaml actions
setup.py pypi
environment.yml pypi
  • datasets *
  • einops *
  • fsspec *
  • huggingface-hub *
  • pysolar *
  • pytorch-lightning *
  • torch-geometric-temporal *

Score: 14.460275990941378