---
language:
  - en
pretty_name: "ClimX: extreme-aware climate model emulation"
tags:
  - climate
  - earth-system-model
  - machine-learning
  - emulation
  - extremes
  - netcdf
license: mit
task_categories:
  - time-series-forecasting
  - other
---

<!-- NOTE: Math formatting convention: inline math uses \\( ... \\) and math blocks use $$ ... $$. -->

# ClimX: a challenge for extreme-aware climate model emulation

ClimX is a challenge about building **fast and accurate machine learning emulators** for the NorESM2-MM Earth System Model, with evaluation focused on **climate extremes** rather than mean climate alone.

## Dataset summary

This dataset contains the **full-resolution** ClimX data in **NetCDF-4** format (targets + forcings, depending on split) with a native grid of \\(192 \times 288\\) (about \\(1^\circ\\)) resolution. It also contains the **lite-resolution** version, with a native grid of \\(12 \times 18\\) (about \\(16^\circ\\)) resolution:

- **Lite-resolution**: <1GB, \\(16\times\\) spatially coarsened, meant for rapid prototyping.
- **Full-resolution**: ~200GB, full-resolution data for large-scale training.

## What you will do (high level)

You train an emulator that predicts **daily** 2D fields for 7 surface variables:

- `tas`, `tasmax`, `tasmin`
- `pr`, `huss`, `psl`, `sfcWind`

However, the **benchmark targets are 15 extreme indices** derived from daily temperature and precipitation (ETCCDI-style indices). The daily fields are an **intermediate output** your emulator produces (useful for diagnostics and for computing the indices).

Conceptually:

$$
x_t = g(f_t, f_{t-1}, \dots, f_{t-\alpha}, x_{t-1}, x_{t-2}, \dots, x_{t-\beta})
$$

where \\(f_t\\) are forcings (greenhouse gases + aerosols) and \\(x_t\\) is the climate state.

## Dataset structure

### Spatial and temporal shape

Full-resolution daily fields:

- **Historical**: `lat: 192, lon: 288, time: 60224`
- **Projections**: `lat: 192, lon: 288, time: 31389`

### Splits and scenarios (official challenge setup)

Training uses historical + several SSP scenarios; testing is on the held-out **SSP2-4.5** scenario:

- **Train**: historical (1850–2014) + `ssp126`, `ssp370`, `ssp585` (2015–2100)
- **Test (held-out)**: `ssp245` (2015–2100)

To avoid leakage, **targets for `ssp245` are withheld** in the official evaluation; only the **forcings** are provided for that scenario. The full outputs will be released after the competition.

## Evaluation metric

The primary leaderboard metric is the region-wise **normalized Nash–Sutcliffe efficiency (nNSE)**, averaged over 15 climate extreme indices.

For each index \\(v\\), grid cell \\((i,j)\\), a validity mask \\(\mathcal{V}\\) excludes cells with negligible temporal variability. Cell-level \\(R^2\\) and nNSE are:

$$
R^2_{ij} = 1 - \frac{\mathrm{MSE}_{ij}}{\mathrm{Var}_t(gt_{ij})}, \qquad \mathrm{nNSE}_{ij} = \frac{R^2_{ij}}{2 - R^2_{ij}}
$$

For each AR6 land region \\(k\\), the area-weighted regional score is:

$$
\mathrm{nNSE}_{kv} = \frac{\sum_{(i,j)\in k \cap \mathcal{V}} \cos\phi_i \, \mathrm{nNSE}_{ij}}{\sum_{(i,j)\in k \cap \mathcal{V}} \cos\phi_i}
$$

The final score averages uniformly over valid regions and indices:

$$
S = \frac{1}{|V|} \sum_{v \in V} \frac{1}{|K_v|} \sum_{k \in K_v} \mathrm{nNSE}_{kv}
$$

\\(S=1\\) is perfect agreement, \\(S=0\\) corresponds to a mean predictor, and \\(S \to -1\\) is pathological.

## How to load the data

This dataset is distributed as **NetCDF-4** files. There are two common ways to load it.

### Option 1 (recommended): clone the ClimX code and use the helper loader

The ClimX repository already includes a helper module (`src/data/climx_hf.py`) that allows you to download the dataset from Hugging Face and open it as three lazily-loaded “virtual” xarray datasets:

```bash
git clone https://github.com/IPL-UV/ClimX.git
cd ClimX
pip install -U "huggingface-hub" xarray netcdf4 dask
```

```python
from src.data.climx_hf import download_climx_from_hf, open_climx_virtual_datasets

# Download NetCDF artifacts from HF into a local cache directory.
root = download_climx_from_hf("/path/to/hf_cache", variant="full")

# Open as three virtual datasets (lazy / dask-friendly).
ds = open_climx_virtual_datasets(root, variant="full") # or "lite"

ds.hist # historical (targets + forcings)
ds.train # projections training scenarios (targets + forcings; excludes `ssp245` scenario)
ds.test_forcings # `ssp245` scenario forcings only (no targets)
```

### Option 2: download NetCDFs and open with xarray directly

You can also download files from Hugging Face and open them with **xarray**.

Example:

```python
from huggingface_hub import hf_hub_download
import xarray as xr

path = hf_hub_download(
    repo_id="isp-uv-es/ClimX",
    repo_type="dataset",
    filename="PATH/TO/A/FILE.nc",  # replace with an actual file in this dataset repo
)
ds = xr.open_dataset(path)
print(ds)
```

## Links

- [Kaggle competition](https://www.kaggle.com/competitions/climx)
- [Full dataset (this page)](https://huggingface.co/datasets/isp-uv-es/ClimX)
- [Public code repository (challenge materials)](https://github.com/IPL-UV/ClimX)
- [Website](https://ipl-uv.github.io/ClimX/)

## License and usage

The dataset is released under **MIT**. In addition, if you are participating in the ClimX competition, please follow the competition rules (notably: restrictions on external climate training data and redistribution of competition data).