File size: 1,574 Bytes
e8345f6
 
 
f02a1e4
 
 
e8345f6
495d193
 
 
 
 
e8345f6
495d193
 
085250d
 
 
 
 
495d193
 
43ed63d
 
 
495d193
 
43ed63d
 
ec15704
e8345f6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
---
license: apache-2.0
---

<img src="ACE-logo.png" alt="Logo for the ACE Project" style="width: auto; height: 50px;">

# ACE-climSST

Ai2 Climate Emulator (ACE) is a family of models designed to simulate atmospheric variability from the time scale of days to centuries.

**Disclaimer: ACE models are research tools and should not be used for operational climate predictions.**

ACE-climSST is the original ACE model, as described in [ACE: A fast, skillful learned global atmospheric model for climate prediction](https://arxiv.org/abs/2310.02074).
It is trained on output from the FV3GFS atmospheric model forced with annually-repeating climatological sea surface temperature.

Quick links:
- 📃 [Paper](https://arxiv.org/abs/2310.02074)
- 💻 [Code](https://github.com/ai2cm/ace)
- 💬 [Docs](https://ai2-climate-emulator.readthedocs.io/en/stable/)
- 📂 [All Models](https://huggingface.co/collections/allenai/ace-67327d822f0f0d8e0e5e6ca4)

Briefly, the strengths of ACE-climSST are:
- long-term stability,
- highly accurate time-mean climate compared to its target dataset,
- very fast inference compared to typical physics-based atmospheric models.

Some known weaknesses are:
- responses to El Niño-like sea surface temperature variability and long-term warming trends are not accurately captured,
- small but non-zero drifts in total dry air mass of the atmosphere.

Note the checkpoint provided here is the same as the one in [this Zenodo repository](https://zenodo.org/records/10791087), just with the optimizer state removed to decrease the checkpoint size.