# microsoft/aurora

Implementation of the Aurora model for Earth system forecasting

Repository: https://github.com/microsoft/aurora
Canonical: https://ross.abutalabs.com/products/microsoft-aurora
Homepage: https://microsoft.github.io/aurora
Language: Python
License: NOASSERTION
License Family: other
Topics: atmospheric-chemistry, aurora-model, deep-learning, foundation-models, ocean-waves, tropical-cyclone-tracking, weather-prediction
Last push: 2026-08-19T20:14:02+00:00

## Health v2 (maintenance only)
Score: 89/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 96, longevity 55
- inputs: {"age_days": 779, "days_push": 14, "days_rel": 29, "gap_med": 26, "n_releases_24m": 14}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1010, forks 166 (observed 2026-08-28T04:03:12.928222+00:00)

## What it is
Aurora is Microsoft's implementation of a deep learning foundation model for Earth system forecasting, predicting atmospheric variables like temperature, plus specialised versions for air pollution, ocean waves, and tropical cyclone tracking. It is a Python library pretrained on large datasets that can be fine-tuned for specialised atmospheric forecasting tasks with little data.

## Use cases
- forecast weather variables like temperature with a machine learning model
- predict atmospheric pollution such as nitrogen dioxide
- forecast ocean wave direction and height
- track tropical cyclone paths
- fine-tune a weather foundation model on small specialised datasets
- run Aurora on ERA5 reanalysis data
- research deep learning for Earth system modelling

## When to choose
- you need state-of-the-art ML-based atmospheric or ocean forecasting
- you want to fine-tune a pretrained Earth system foundation model on limited data
- you are doing research on weather, waves, or cyclone prediction with deep learning

## When to avoid
- you need operational deterministic NWP rather than an experimental ML model
- you have no GPU resources for inference
- commercial use without licensing arrangements with Microsoft
- you need a general-purpose climate simulation package rather than forecasting

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, simulation
- domain: weather, machine-learning, artificial-intelligence, data-science
- platform: python
- tags: foundation-model, earth-system-forecasting, atmospheric-chemistry, ocean-waves, tropical-cyclone-tracking, weather-prediction, era5, gpu

## Member repositories
- microsoft/aurora (main) score 89

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:12.928222+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:57.505608+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/aurora (fetched 2026-08-28T04:03:12.928222+00:00, sha 9e698f1e55c6)
  - homepage: https://microsoft.github.io/aurora (fetched 2026-08-29T13:11:54.734815+00:00, sha 9de206263ecd)
- Data as of 2026-08-30T08:39:29.467469+00:00.