openclimatefix / openclimatefix/metnet

[Paper] MetNet Global

Open
#74 9 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
305
Forks
64
Avg merge
5d 4h
Merged PRs (30d)
1

Description

Arxiv/Blog/Paper Link

https://arxiv.org/pdf/2510.13050

Detailed Description

This is the global MetNet paper, using satellite imagery for 0.5km and 15 minutely precipitation forecasts globally, up to 12 hours.

There are a few interesting points in it. The main targets are CORRA precipitation from GPM, with auxiliary targets being precipitation radar outputs from Japan, Europe, and the US, as well as GPM IMERG Final precipitation estimates.

They use 7 geostationary satellites, the 3 EUMETSAT ones + GK-2A, Himawari9, GOES-19/GOES-18, and blend them into a single mosaic using satpy.

Similar to MetNet-2 on the lead time conditioning, as well as some other architectural tricks to save memory (cropping the feature maps for example as you go through the residual blocks)

Image Image Image

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading the linked arXiv paper, then inspect the repository's current MetNet and MetNet-2 implementation. The issue does not name files, tests, a requested entry point, or acceptance criteria, so the intended change and definition of done need to be clarified before work begins.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.