Data

Tools

Indexes

Applications

Experiments

Open Source Science

chaosbench

https://github.com/leap-stc/chaosbench

Science Score: 23.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ✓
    Academic publication links
    Links to: arxiv.org
  • ○
    Committers with academic emails
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (10.4%) to scientific vocabulary
Last synced: 11 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: leap-stc
  • License: gpl-3.0
  • Language: HTML
  • Default Branch: main
  • Size: 66.7 MB
Statistics
  • Stars: 42
  • Watchers: 1
  • Forks: 6
  • Open Issues: 2
  • Releases: 0
Created almost 3 years ago · Last pushed over 1 year ago
Metadata Files
Readme License

README.md

ChaosBench: A Multi-Channel, Physics-Based Benchmark for Subseasonal-to-Seasonal Climate Prediction

NeurIPS 2024 Oral

Homepage arXiv Huggingface Dataset License Badge


ChaosBench is a benchmark project to improve and extend the predictability range of deep weather emulators to the subseasonal-to-seasonal (S2S) range. Predictability at this scale is more challenging due to its: (1) double sensitivities to intial condition (in weather-scale) and boundary condition (in climate-scale), (2) butterfly effect, and our (3) inherent lack of understanding of physical processes operating at this scale. Thus, given the high socioeconomic stakes for accurate, reliable, and stable S2S forecasts (e.g., for disaster/extremes preparedness), this benchmark is timely for DL-accelerated solutions.

💡 View our interactive leaderboard here

Features

Overview of ChaosBench

1️⃣ Diverse Observations. Spanning over 45 years (1979-), we include ERA5/LRA5/ORAS5 reanalysis for a fully-coupled Earth system emulation (atmosphere-terrestrial-sea-ice)

2️⃣ Diverse Baselines. Wide selection of physics-based forecasts from leading national weather agencies in Europe, the UK, America, and Asia

3️⃣ Differentiable Physics Metrics. In addition to deterministic and probabilistic metrics, we introduce two differentiable physics-based metrics to minimize the decay of power spectra at long forecasting horizon (reduce blurriness)

4️⃣ Large-Scale Benchmarking. Systematic large-scale evaluation for state-of-the-art ML-based weather emulators like ViT/ClimaX, PanguWeather, GraphCast, and FourcastNetV2

Getting Started

Build Your Own Model

Benchmarking

Citation

If you find any of the code and dataset useful, feel free to acknowledge our work through:

```bibtex @article{nathaniel2024chaosbench, title={Chaosbench: A multi-channel, physics-based benchmark for subseasonal-to-seasonal climate prediction}, author={Nathaniel, Juan and Qu, Yongquan and Nguyen, Tung and Yu, Sungduk and Busecke, Julius and Grover, Aditya and Gentine, Pierre}, journal={arXiv preprint arXiv:2402.00712}, year={2024} }

Owner

  • Name: LEAP
  • Login: leap-stc
  • Kind: organization
  • Location: United States of America

Learning the Earth with AI and Physics

GitHub Events

Total
  • Issues event: 2
  • Watch event: 24
  • Issue comment event: 2
  • Push event: 27
  • Pull request event: 2
  • Fork event: 6
Last Year
  • Issues event: 2
  • Watch event: 24
  • Issue comment event: 2
  • Push event: 27
  • Pull request event: 2
  • Fork event: 6

Committers

Last synced: 11 months ago

All Time
  • Total Commits: 64
  • Total Committers: 3
  • Avg Commits per committer: 21.333
  • Development Distribution Score (DDS): 0.125
Past Year
  • Commits: 16
  • Committers: 1
  • Avg Commits per committer: 16.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
juannat7 n****n@g****m 56
Juan Nathaniel 4****7@u****m 7
YONGQUAN-QU 5****U@u****m 1

Issues and Pull Requests

Last synced: 11 months ago

All Time
  • Total issues: 2
  • Total pull requests: 2
  • Average time to close issues: N/A
  • Average time to close pull requests: 2 minutes
  • Total issue authors: 2
  • Total pull request authors: 2
  • Average comments per issue: 1.5
  • Average comments per pull request: 0.0
  • Merged pull requests: 1
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 2
  • Pull requests: 1
  • Average time to close issues: N/A
  • Average time to close pull requests: 3 minutes
  • Issue authors: 2
  • Pull request authors: 1
  • Average comments per issue: 1.5
  • Average comments per pull request: 0.0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
  • willzhang3 (1)
  • SauryChen (1)
Pull Request Authors
  • Cas-Dec (2)
  • yongquan-qu (1)
Top Labels
Issue Labels
Pull Request Labels

Dependencies

.github/workflows/build-book.yaml actions
  • CumulusDS/get-yaml-paths-action v1.0.2 composite
  • actions/cache v3 composite
  • actions/checkout v3 composite
  • actions/upload-artifact v4 composite
  • andstor/file-existence-action v2 composite
  • conda-incubator/setup-miniconda v2 composite
  • tj-actions/changed-files v36 composite
.github/workflows/deploy-book.yaml actions
  • actions/download-artifact v3 composite
  • dawidd6/action-download-artifact v2.27.0 composite
  • peaceiris/actions-gh-pages v3.9.3 composite
.github/workflows/publish-website.yml actions
requirements.txt pypi
  • PyYAML ==6.0.1
  • PyYAML ==6.0
  • cdsapi ==0.6.1
  • cfgrib ==0.9.10.4
  • ecmwf_api_client ==1.6.3
  • einops ==0.7.0
  • lightning ==2.0.9
  • matplotlib ==3.7.2
  • numpy ==1.25.2
  • pandas ==2.0.3
  • scipy ==1.12.0
  • seaborn ==0.13.2
  • timm ==0.9.7
  • torch ==2.0.1
  • torchist ==0.2.1
  • tqdm ==4.66.1
  • xarray ==2023.8.0
website/environment.yml pypi