climpred: verification of weather and climate forecasts


Documentation Status JOSS paper DOI


CI Requirements Status codecov


Conda Version pypi Version




gitter chat GitHub contributors Conda downloads


climpred gallery climpred workshop climpred cloud demo


We are actively looking for new contributors for climpred! Riley moved to McKinsey’s Climate Analytics team. Aaron is finishing his PhD, but will stay in academia. We especially hope for python enthusiasts from seasonal, subseasonal or weather prediction community. In our past coding journey, collaborative coding, feedbacking issues and pull requests advanced our code and thinking about forecast verification more than we could have ever expected. Aaron can provide guidance on implementing new features into climpred. Feel free to implement your own new feature or take a look at the good first issue tag in the issues. Please reach out to us via gitter.

Version 2.1.3 Release

March 23th, 2021

After our JOSS paper is published, the climatology reference forecast was introduced and persistence forecasts are extended for all metrics (before probabilistic persistence skill was not calculated) (GH#566). Furthermore, any errors in or does not call silently anymore, but raises warnings instead (GH#552).


You can install the latest release of climpred using pip or conda:

pip install climpred
conda install -c conda-forge climpred

You can also install the bleeding edge (pre-release versions) by cloning this repository and running pip install . --upgrade in the main directory or

pip install git+

Getting Started

User Guide

Help & Reference