Skip to content

[Paper] AI-based Data Assimilation #173

@jacobbieker

Description

@jacobbieker

Arxiv/Blog/Paper Link

https://arxiv.org/pdf/2406.00390

Detailed Description

Changes the data assimilation approach to be self-supervised and more similar to the physics assimilation process. The loss is the 3D-Var equation, and all that is needed is a first guess state and observations. The first guess state is generally a previous forecast step, but can also be blank if none exist.

They do this for 2m temperature as a real-world case over Germany and get interesting and promising results.

Context

Data assimilation is a important process, this could be an interesting way of doing it.

Metadata

Metadata

Assignees

Labels

enhancementNew feature or request

Type

No type
No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions