---
license: mit
---

# Samudra 🌊
This repository contains the data used to train and test the Samudra Ocean Emulator described in the paper ["Samudra: An AI Global Ocean Emulator for Climate"](https://arxiv.org/abs/2412.03795). Samudra is a skillful global emulator of the ocean component of a state-of-the-art climate model. We emulate key ocean variables, sea surface height, horizontal velocities, temperature, and salinity, across their full depth. Refer the code [here](https://github.com/m2lines/Samudra.git)

<p align="center">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/66843421443492c24ce5a19a/jP-SZ0oW9hVOWNZKMouj0.gif" >
</p>

The data is also publicly hosted as uncompressed zarrs on our pod:

```python
import xarray as xr
data = xr.open_dataset("https://nyu1.osn.mghpcc.org/m2lines-pubs/Samudra/OM4", engine='zarr', chunks={})
```


### Citing
If you use this data in your research, please consider citing the following paper:
```
@article{dheeshjith2024samudra,
  title={Samudra: An AI Global Ocean Emulator for Climate},
  author={Dheeshjith, Surya and Subel, Adam and Adcroft, Alistair and Busecke, Julius and Fernandez-Granda, Carlos and Gupta, Shubham and Zanna, Laure},
  journal={arXiv preprint arXiv:2412.03795},
  year={2024}
}
```