Herbie is a python package that downloads recent and archived numerical weather prediction (NWP) model output from different cloud archive sources. NWP data is usually in GRIB2 format and can be read with xarray+cfgrib. Much of this data is made available through the NOAA Big Data Program which has made weather data more accessible than ever before.
Herbie helps you discover and download data from:
- High Resolution Rapid Refresh (HRRR) | HRRR-Alaska
- Rapid Refresh (RAP)
- Global Forecast System (GFS)
- Global Ensemble Forecast System (GEFS) β¨ new in Herbie 0.0.10
- ECMWF Open Data Forecast Products β¨ new in Herbie 0.0.8
- National Blend of Models (NBM)
- Rapid Refresh Forecast System - Prototype (RRFS)
π Herbie Documentation
Requires cURL and Python 3.8+ with requests, numpy, pandas, xarray, and cfgrib. Optional packages are matplotlib, cartopy, and Carpenter Workshop.
pip install herbie-data
or
pip install git+https://github.com/blaylockbk/Herbie.git
or, create the provided conda environment.
- Search different data sources for model output.
- Download full GRIB2 files
- Download subset GRIB2 files (by grib field)
- Read data with xarray
- Read index file with Pandas
- Plot data with Cartopy (very early development)
graph TD;
d1[(HRRR)] -.-> H
d2[(RAP)] -.-> H
d3[(GFS)] -.-> H
d33[(GEFS)] -.-> H
d4[(ECMWF)] -.-> H
d5[(NBM)] -.-> H
d6[(RRFS)] -.-> H
H((Herbie))
H --- .download
H --- .xarray
H --- .read_idx
style H fill:#d8c89d,stroke:#0c3576,stroke-width:4px,color:#000000
from herbie.archive import Herbie
# Herbie object for the HRRR model 6-hr surface forecast product
H = Herbie(
'2021-01-01 12:00',
model='hrrr',
product='sfc',
fxx=6
)
# Download the full GRIB2 file
H.download()
# Download a subset, like all fields at 500 mb
H.download(":500 mb")
# Read subset with xarray, like 2-m temperature.
H.xarray("TMP:2 m")
Herbie downloads model data from the following sources, but can be extended to include others:
- NOMADS
- Big Data Program Partners (AWS, Google, Azure)
- ECMWF Open Data Azure storage
- University of Utah CHPC Pando archive
During my PhD at the University of Utah, I created, at the time, the only publicly-accessible archive of HRRR data. Over 1,000 research scientists and professionals used that archive.
β I wrote a paper about archiving HRRR data in the cloud
Blaylock B., J. Horel and S. Liston, 2017: Cloud Archiving and Data Mining of High Resolution Rapid Refresh Model Output. Computers and Geosciences. 109, 43-50. https://doi.org/10.1016/j.cageo.2017.08.005
In the later half of 2020, the HRRR data was made available through the NOAA Big Data Program. Herbie organizes and expands my original download scripts into a more coherent package with the extended ability to download more than just the HRRR and RAP model data and from many different archive sources.
I originally released this package under the name βHRRR-Bβ because it only dealt with the HRRR data set and the βBβ was for my first name initial. Since then, I have added the ability to download RAP, GFS, NBM, RRFS, and others with potentially more models in the future. Thus, this package was rename Herbie, named after one of my favorite childhood movie characters.
π¨π»βπ» Contributing Guidelines
π¬ GitHub Discussions
π GitHub Issues
π Personal Webpage
π University of Utah HRRR archive
Thanks for using Herbie, and Happy Racing ππ
- Brian
P.S. If you like Herbie, check out my other repos:
- π GOES-2-go: A python package to download GOES-East/West data and make RGB composites.
- π‘ SynopticPy: A python package to download mesonet data from the Synoptic API.
- π¨ Carpenter Workshop: A python package with various tools I made that are useful (like easy funxtions to build Cartopy maps).
- π¬ Bubble Print: A silly little python package that gives your print statement's personality.
- π MET Syntax: An extension for Visual Studio Code that gives syntax highlighting for Model Evaluation Tools (MET) configuration files.
As an alternative you can use rclone to download files from AWS or GCP. I quite like rclone. Here is a short rclone tutorial