# NOAA RRFS Ensemble Forecast System (REFS) Ensemble Means, Spreads and Probabilities, dynamical.org Icechunk Zarr

## Facts

- Provider: NOAA NWS NCEP (model); dynamical.org (Icechunk Zarr archive)
- License: CC-BY-4.0 (https://stac.dynamical.org/noaa-refs-forecast-hourly-virtual/collection.json)
- Stored at: Amazon S3, s3://dynamical-noaa-refs (us-west-2)
- Access: public, readable without an account
- Category: Climate & Weather
- Source: https://registry.opendata.aws/dynamical-noaa-refs/
- Page: https://scigantic.com/archive/a66ca9d2-fdc9-5fab-bff1-69b4b3b5a118
- Keywords: Climate & Weather, AWS Open Data

## About

Ensemble products from the NOAA NCEP RRFS Ensemble Forecast System (REFS), which combines the current and 6-hour-old cycles of the deterministic Rapid Refresh Forecast System (RRFS) and its ensemble (RRFS ENS) with two HRRR members into means, spreads and threshold probabilities on a 3 km grid over the continental United States. Reformatted by dynamical.org as two Icechunk Zarr stores. noaa-refs-forecast-hourly-virtual has 243 variables at hourly steps from 1 to 60 hours with a statistic dimension (mean, standard_deviation): CAPE and CIN, dew point and geopotential height at standard pressure levels, radar reflectivity, fog liquid water, cloud base and ceiling, and ensemble agreement scale (EAS) probabilities of 1-hour precipitation and snowfall above set thresholds. noaa-refs-forecast-3-hourly-virtual has 96 variables at 3-hour steps from 3 to 60 hours: EAS probabilities of 3, 6, 12 and 24 hour precipitation and snowfall above thresholds, localized and mean probability-matched mean precipitation, and neighborhood probabilities of precipitation exceeding 1 to 100 year average recurrence intervals. Both are initialized every 6 hours from 2026-08-13 to the present. REFS is pre-operational, so expect changes. Open with dynamical_catalog.open("noaa-refs-forecast-hourly-virtual") (needs gribberish) or icechunk plus xarray. Licensed CC BY 4.0 per dynamical.org; credit NOAA NCEP and dynamical.org.

## Load the data

```python
import boto3
from botocore import UNSIGNED
from botocore.client import Config
s3 = boto3.client('s3', region_name='us-west-2', config=Config(signature_version=UNSIGNED))
s3.list_objects_v2(Bucket='dynamical-noaa-refs', Prefix='', Delimiter='/')
```

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

## Data structure

Example file: `noaa-refs-forecast-3-hourly-virtual/v0.1.0.icechunk/chunks/00AW2KNYZ9H0EFWMWA90`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/a66ca9d2-fdc9-5fab-bff1-69b4b3b5a118
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/a66ca9d2-fdc9-5fab-bff1-69b4b3b5a118
- Over MCP: `get_schema_card` with id `a66ca9d2-fdc9-5fab-bff1-69b4b3b5a118` at https://api.scigantic.com/mcp

## Related topics

- [Climate & Weather datasets](https://scigantic.com/topics/climate-and-weather)
