---
license: cc-by-4.0
#User-Defined Tags
tags:
  - CMEMS
  - Marine forecast
  - Reanalysis
  - Physical Ocean
  - Global Ocean
language:
  - en
  - zh
---
<p align="center">
  <strong>
    <span style="font-size: 30px;">CMEMS</span>
  </strong>
</p>


## Dataset Description

CMEMS is an HDF5 gridded dataset for global ocean forecasting tasks. By integrating satellite and in situ observations with numerical ocean models, it provides real-time analyses, forecasts, and historical reconstructions for a variety of variables.

## Supported Tasks

This standardized data repository contains 7 annual files covering 1993-1999. Each file provides ocean variable fields, global means, and global standard deviations, and can be used for XiHe model training, validation, testing, inference input, and evaluation.

## Dataset Format and Structure

Each annual file is in HDF5 format and follows the path pattern `data/<year>.h5`. Each file contains:

| HDF5 Path | shape | dtype | Description |
|---|---|---|---|
| `fields` | `[3, 96, 2041, 4320]` | `float32` | 3 time samples, 96 variables, and a global grid |
| `global_means` | `[1, 96, 1, 1]` | `float32` | Global means aligned with the variable order |
| `global_stds` | `[1, 96, 1, 1]` | `float32` | Global standard deviations aligned with the variable order |

`fields.attrs["variables"]` contains the names of 96 variables, and `fields.attrs["time_step"]` is 24. The detailed schema is provided in `metadata/schema.yaml`.

## How to Use the Dataset

This dataset is compatible with the `OneScience-Sugon/XiHe` model.


Download the dataset:

```bash
hf download --dataset OneScience-Sugon/CMEMS --local-dir ./CMEMS
```


## Official OneScience Information

| Platform | OneScience Main Repository | Skills Repository |
|---|---|---|
| Gitee  | https://gitee.com/onescience-ai/onescience | https://gitee.com/onescience-ai/oneskills |
| GitHub  | https://github.com/onescience-ai/OneScience | https://github.com/onescience-ai/oneskills |

## Limitations and License

CMEMS follows an open data policy consistent with CC-BY-4.0 principles (official license agreement), and the dataset is licensed under CC-BY-4.0.