Examples#
This is a collection of examples in Earth2Studio that demonstrate various functionality and commonly used workflows.
Running Examples
Earth2Studio examples can be downloaded as a notebook or runnable Python script. Each requires installation of different optional dependency groups or additional packages for the specific models used or post-processing steps. Use uv to auto install dependencies on execution:
uv run <example_script>.py
If you are using a container or other type of environment, then pip installing will likely be needed. Look for the uv inline metadata blocks of the form:
# /// script
# dependencies = [
# "earth2studio @ git+https://github.com/NVIDIA/earth2studio.git",
# "cartopy",
# ]
# ///
Pip install these packages then execute the example with:
python <example_script>.py
Getting Started#
Introductory examples demonstrating the core inference workflows in Earth2Studio.
Medium Range#
Examples focused on medium-range global weather forecasting, ensemble methods, and post-processing workflows.
Downscaling#
Examples demonstrating generative and ensemble downscaling workflows for producing high-resolution outputs from global model forecasts.
Nowcasting#
Examples demonstrating high-resolution nowcasting and short-range forecasting with models such as StormCast and StormScope.
Data Assimilation#
Examples demonstrating data assimilation workflows in Earth2Studio.
Subseasonal to Seasonal#
Examples demonstrating subseasonal to seasonal (s2s) inference workflows with typical prediction timescales of two weeks to two years.
Miscellaneous#
Miscellaneous examples covering IO performance, custom data sources, and other utility workflows.
Extending Earth2Studio#
Examples focused on demonstrating how to extend Earth2Studio with user-defined components.
