Extreme Weather Bench

Builds on the successful work of WeatherBench and introduces a set of high-impact weather events, spanning across multiple spatial and temporal scales and different parts of the weather spectrum.
https://github.com/brightbandtech/extremeweatherbench

Category: Atmosphere
Sub Category: Meteorological Observation and Forecast

Keywords

benchmarking meteorology

Last synced: about 13 hours ago
JSON representation

Repository metadata

Benchmarking of machine learning and numerical weather prediction (MLWP & NWP) models, with a focus on extreme and discrete meteorological cases and event types.

• Host: GitHub
• URL: https://github.com/brightbandtech/extremeweatherbench
• Owner: brightbandtech
• License: mit
• Created: 2024-08-15T15:33:50.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2025-04-22T00:27:45.000Z (6 days ago)
• Last Synced: 2025-04-24T22:36:33.790Z (3 days ago)
• Topics: benchmarking, meteorology
• Language: Python
• Homepage: https://www.brightband.com/benchmarks
• Size: 148 MB
• Stars: 46
• Watchers: 5
• Forks: 2
• Open Issues: 12
• Releases: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

README.md

pip install git+https://github.com/brightbandtech/ExtremeWeatherBench.git

from extremeweatherbench import config, events, evaluate
import pickle 

# Select model
model = 'FOUR_v200_GFS'

# Set up path to directory of file - zarr, json, or parquet
forecast_dir = f'assets/data/forecasts/{model}_combined_all.parq'

# Choose the event types you want to include
event_list = [events.HeatWave,
              events.Freeze]

# Set up configuration object that includes events and the forecast directory
heatwave_configuration = config.Config(
    event_types=event_list,
    forecast_dir=forecast_dir,
    )

# Use ForecastSchemaConfig to map forecast variable names to CF convention-based names used in EWB
# the sample forecast kerchunk references to the CIRA MLWP archive are the default configuration
default_forecast_config = config.ForecastSchemaConfig()

# Run the evaluate script which outputs a dict of event results with associated metrics and variables
cases = evaluate.evaluate(eval_config=heatwave_configuration, forecast_schema_config=default_forecast_config)

# Save the results to a pickle file
with open(f'cases_{model}.pkl', 'wb') as f:
    pickle.dump(cases, f)

Owner metadata

• Name: Brightband
• Login: brightbandtech
• Email:
• Kind: organization
• Description: Brightband is making weather and climate predictable for all, to help humanity adapt to increasingly extreme weather.
• Website: brightband.com
• Location: United States of America
• Twitter: brightbandtech
• Company:
• Icon url: https://avatars.githubusercontent.com/u/159490539?v=4
• Repositories: 1
• Last ynced at: 2024-09-18T00:32:39.957Z
• Profile URL: https://github.com/brightbandtech

GitHub Events

Total

• Create event: 36
• Release event: 1
• Issues event: 50
• Watch event: 39
• Delete event: 30
• Issue comment event: 65
• Public event: 1
• Push event: 308
• Pull request review event: 53
• Pull request review comment event: 100
• Pull request event: 63
• Fork event: 2

Last Year

• Create event: 36
• Release event: 1
• Issues event: 50
• Watch event: 39
• Delete event: 30
• Issue comment event: 65
• Public event: 1
• Push event: 308
• Pull request review event: 53
• Pull request review comment event: 100
• Pull request event: 63
• Fork event: 2

Committers metadata

Last synced: 6 days ago

Committer domains:

• danielrothenberg.com: 1
• ou.edu: 1

Issue and Pull Request metadata

Last synced: 1 day ago

Dependencies

Score: 5.159055299214529