Source Cooperative is a Radiant Earth project

Jacob Prince-Bieker

Machine Learning Researcher, focused on climate, weather, and renewable energy

Products

AOML Ocean GTS Observations
This repo is all available ocean observations from AOML in the flattened GTS stream, converted to Icechunk as one large store, from 2012-01-01 to 2025-06-30.
Published on 7 Jul 2025Public
Global Hydrology Estimate and Enterprise Rain Rate
This contains data from the Global Hydrology Estimate product and Enterprise Rain Rate product from NOAA in Icechunk format from. https://registry.opendata.aws/noaa-ghe The Enterprise Rain Rate is the replacement for the GHE.
Published on 28 May 2025Public
Geostationary Full-Disk Imagery
This repo contains full-disk geostationary imagery from GOES/Himawari/G2-KA in icechunk format, converted from the public AWS archives. The data is processed by Satpy, and is the L1b data. The datasets are very incomplete, but the goal is to slowly add more historical imagery as time allows, as well as near-real time.
Published on 10 Jul 2025Public
GEOS Atmosphere Composition Analysis
GEOS-CF High temporal resolution analysis, in Icechunk. 15-minutely, 0.25 degree resolution products. This is the analysis product, available from 2018-Present, with a 1 day delay. More information is available: https://gmao.gsfc.nasa.gov/weather_prediction/GEOS-CF/.
Published on 29 Jun 2025Public
Global Mosaic of Geostationary Images
This is the Global Mosaic of Geostationary Images, converted to Zarr v3. It is a mosaic of multiple geostationary weather satellites, on an hourly cadence. This is a unified version of the Hugging Face dataset [here](https://huggingface.co/datasets/jacobbieker/global-mosaic-of-geostationary-images) with some extensions.
Published on 20 Nov 2024Public
GPM IMERG Early and Final Half-Hourly Precipitation Datasets
This contains two Zarrs, comprising of the GPM IMERG Half Hourly Early, and Final precipitation products. The time frame is from 2000-01-01 to present. The dataest is in the process of being generated, so might be incomplete.
Published on 30 Nov 2024Public
MRMS Rainfall Measurements
MRMS Rainfall Measurements, converted to Zarr format. All are the hourly accumulations. A few different options are included, including the Gauge measurement, MultiSensor MultiPass measurement, and the derived precipitation rate.
Published on 21 May 2024Public
Cloud-Optimized NREL National Solar Radiation Database
This contains international data listed here (https://nsrdb.nrel.gov/data-sets/international-data) in Icechunk format. The first location available is the India data, as it is the smallest. Plans include reprocessing the full set of hourly/minutely data available here: https://data.openei.org/s3_viewer?bucket=nrel-pds-nsrdb The data has been chunked to optimize reading the full spatial area at single timesteps. If you want to read a single point across time, it might be easier to use the NSRDB's API to access the individual points, or rechunk the data here.
Published on 30 May 2025Public
Polar-Orbiting Satellite Imagery
This contains data from the JPSS and MetOp satellites in Icechunk format. The data has been converted with Satpy, and is updated as time allows. The archives are very incomplete, but the goal is to continually expand them.
Published on 10 Jul 2025Public
International Precipitation Radar
This contains various sets of precipitation radar estimates from different countries, initially the UK and Finland. The data has been optimized with icechunk. The data stored in each store is limited to what is outputted by the national weather services. In Finland's case, this data contains the hourly 1-hour, 12-hour, and 24-hour precipitation accumulations. For the UK, this contains the composite rain rate radar output. For US MRMS outputs, there is a different collection containing that output.
Published on 1 Jun 2025Public
SILAM Global Dust Model Forecasts
Finnish SILAM Global Dust Model Forecasts, in Zarr format, at a 0.1 degree resolution. Additionally includes the FMI aerosol AWS forecast in Icechunk.
Published on 19 Apr 2024Public
WeatherReal-filtered NOAA Integrated Surface Dataset
This is the ISD data, filtered as in the WeatherReal dataset paper. They have been converted to Zarr.
Published on 5 Jun 2025Public