Google WeatherNext 3 forecast, 48 hour, 0.05 degree, virtual
| ID |
google-weathernext3-forecast-48-hour-0-05-degree-virtual
|
| Spatial domain | Global |
| Spatial resolution | 0.05 degrees (~5km) |
| Time domain | Forecasts initialized 2026-01-01 00:00:00 UTC to Present |
| Time resolution | Forecasts initialized every 1 hours |
| Forecast domain | Forecast lead time 1-48 hours ahead |
| Forecast resolution | Hourly |
STAC (browse) · validation report
WeatherNext 3 is Google's global probabilistic weather forecasting model. It combines atmospheric analyses with satellite observations and produces forecasts trained against gridded fields and weather station observations.
This data product provides 48-hour forecasts from the station-trained temperature and dew point head, evaluated on a global grid. These fields are distinct from the gridded fields in the 0.1 degree products. Google's precomputed ensemble mean and percentiles are selected along the statistic dimension. The spatial dimensions y and x hold latitude and 0-360 degree longitude.
Related Datasets
Examples
|
WeatherNext 3 forecasts, virtual (Github) |
|
WeatherNext 3 forecasts, virtual (Colab) |
import dynamical_catalog # dynamical-catalog>=1.0.0
ds = dynamical_catalog.open("google-weathernext3-forecast-48-hour-0-05-degree-virtual", chunks=None)
ds["temperature_2m"].sel(init_time="2026-07-01T00", lead_time="24h", statistic="mean", y=slice(25, 55), x=slice(230, 300))import icechunk
import pystac
import xarray as xr
catalog = pystac.Catalog.from_file("https://stac.dynamical.org/catalog.json")
collection = catalog.get_child("google-weathernext3-forecast-48-hour-0-05-degree-virtual")
asset = collection.assets["icechunk-https"]
authorize = icechunk.containers_credentials({"https://wn.dynamical.org/chunks/": icechunk.Credentials.HttpAccess()})
repo = icechunk.Repository.open(
icechunk.http_storage(asset.href),
authorize_virtual_chunk_access=authorize,
)
session = repo.readonly_session("main")
ds = xr.open_zarr(session.store, chunks=None)
ds["temperature_2m"].sel(init_time="2026-07-01T00", lead_time="24h", statistic="mean", y=slice(25, 55), x=slice(230, 300))Explore
Dimensions
| min | max | units | |
|---|---|---|---|
| init_time | 2026-01-01T00:00:00Z | Present | seconds since 1970-01-01 |
| lead_time | 3600 | 172800 | seconds |
| statistic | Present | 1 | |
| x | 0 | 359.95 | degree_east |
| y | -90 | 90 | degree_north |
Variables
Access a variable by its name (the bold identifier, e.g. ds["dew_point_temperature_2m"]).
Dimensions: init_time × statistic × lead_time × y × x
| variable | units |
|---|---|
dew_point_temperature_2m
2 metre dewpoint temperature (2d)WeatherNext 3's station-trained head, defined everywhere.
|
degree_Celsius |
temperature_2m
2 metre temperature (2t)WeatherNext 3's station-trained head, defined everywhere.
|
degree_Celsius |
Details
License
Dataset licensed under GDM Real-Time Weather Forecasting Experimental Data Terms of Use for Real-Time Data & Forecasts and CC BY 4.0 for Historical Data.
Attribution and citation
Google WeatherNext 3 forecast data processed by dynamical.org from Google's WeatherNext 3 experimental data. Copyright 2024-6 Google LLC. © 2024-6 Google LLC, whose machine learning models were used to create the experimental data made available under the following licence terms https://storage.googleapis.com/weathernext-public/terms-of-use.pdf. This data is intended for experimental modelling only and is not intended, validated, or approved for real world use.
Or Google WeatherNext 3 from dynamical.org.
Source
This data product is built from Google's WeatherNext 3 statistics archive. Google supplies the mean and the 10th, 25th, 50th, 75th and 90th percentiles of its ensemble; these are exposed as mean, p10, p25, p50, p75 and p90 on the statistic dimension. Individual ensemble members are not included. Google's attributions and acknowledgements list third-party materials whose terms may also apply.
Modifications. dynamical.org restructures the source into an Icechunk virtual data product with dynamical.org's variable and dimension names and converts units as data is read. Temperatures are converted from kelvin to degrees Celsius.
Data availability
This data product starts with the 2026-01-01T00 UTC initialization and is updated every hour. Forecasts include steps whose valid time is in the future. Forecasts are initialized every hour; the six-hourly source forecasts are restricted to their first 48 hours. Recent initializations may be incomplete while source data arrives; unpublished steps read as NaN.
Chunks
This dataset is stored in Zarr format, which splits each variable into a grid of chunks — the smallest unit read from storage. When possible, aligning your reads with this dataset's chunk grid can significantly improve data access speed.
The element count and coordinate span of this dataset:
| dimension | chunk |
|---|---|
| init_time | 1 (1 hour) |
| statistic | 1 |
| lead_time | 1 (1 hour) |
| y | 3601 (180.05°) |
| x | 7200 (360°) |
| uncompressed | 98.9 MiB |
Validation report
Review the validation report for variable availability, missing data, known quirks, fill values, value distributions, and sample plots.