search

pySTEPS / pysteps

Python framework for short-term ensemble prediction systems.
https://pysteps.github.io/
BSD 3-Clause "New" or "Revised" License
441 stars 160 forks source link

Some problems about 'Verification' #366

Closed HY-10 closed 1 month ago

HY-10 commented 2 months ago

Can all models (e.g., sprog, anvil, linda, etc.) be run for comparison using the following metrics code (ROC, reliability diagram, FSS, etc.)? I have the following problems when running, and I hope to get your help.

image image image image

dnerini commented 1 month ago

dear @HY-10, ROC curves, rank histograms, and reliability diagrams are suitable for ensemble forecasts (e.g. steps, linda), while the FSS is typically used for deterministic forecasts (e.g. sprog, anvil). I hope this helps, feel free to reopen this issue if you have more questions.