NeuralGCM-0.7 Deterministic 15-Day Forecasts for 2020 (WeatherBench 2)

Provider
Google Research (Kochkov et al., NeuralGCM)
License
CC-BY-SA-4.0 (source)
Format
Zarr (in sampled files)
Stored at
Google Cloud Storage, gs://weatherbench2/datasets/neuralgcm_deterministic (us)
Access
Free and public. No account with the data provider is needed.
DOI
10.1038/s41586-024-07744-y
Category
Climate & Weather
Source
https://doi.org/10.1038/s41586-024-07744-y

About NeuralGCM-0.7 Deterministic 15-Day Forecasts for 2020 (WeatherBench 2)

Medium-range weather forecasts from NeuralGCM-0.7, the deterministic 0.7 degree version of the hybrid model of Kochkov et al. (Nature, 2024) that couples a differentiable spectral dynamical core with learned physics. Forecasts for 2020 are stored in the WeatherBench 2 bucket for scoring against ERA5. Measured 2026-11-16: 2020-512x256.zarr has 797 initialisation times every 12 hours from 2019-12-15 to 2021-01-16, 31 lead times from 0 to 360 h (15 days, 12-hourly), 37 pressure levels from 1 to 1000 hPa on a 512x256 Gaussian-like grid, about 3.8 TB uncompressed. Variables: geopotential, temperature, specific humidity, specific cloud ice and liquid water content, u and v wind, wind speed and cumulative precipitation minus evaporation. Regridded copies on 240x121 and 64x32 equiangular conservative grids sit beside it. Open with xarray.open_zarr and token='anon'; the matching checkpoint is gs://neuralgcm/models/v1/deterministic_0_7_deg.pkl. Google states the NeuralGCM simulation datasets are released under a Creative Commons licence (the docs name CC-BY 4.0 and link CC BY-SA 4.0, so treat as CC BY-SA 4.0); cite Kochkov et al. 2024.

How to access NeuralGCM-0.7 Deterministic 15-Day Forecasts for 2020 (WeatherBench 2) in Python

import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('weatherbench2/datasets/neuralgcm_deterministic/', detail=False)[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

File types in sampled files: Zarr

Files

Example file: datasets/neuralgcm_deterministic/2020-240x121_equiangular_with_poles_conservative.zarr/P_minus_E_cumulative/.zarray

Analyze this dataset without downloading it

On Scigantic, NeuralGCM-0.7 Deterministic 15-Day Forecasts for 2020 (WeatherBench 2) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Climate & Weather datasets