OSDF_Examples

About¶

Contains Jupyter notebook workflows which access climate data from various OSDF origins using PelicanFS. The examples include workflows where computations are executed on various platforms like NCAR’s Casper, TACC’s Stampede3 and cloud computing platforms like Jetstream2. If you use any workflow in this repository, please cite using

Example Workflows¶

1. Access CESM2 LENS data from the AWS opendata origin and the NCAR data origin and
   • a) Benchmark data access speeds for subsets of various sizes.
   • b) Bias-correct surface temperature using ERA5 reanalysis.
   • c) Compute surface ocean heat content.
   • d) Compute GMST anomaly and plot.
2. Access NOAA SONAR data from the AWS origin to plot echograms
3. Acesss Sentinel-2 satellite data from AWS to compute spectral change
4. Access NA-CORDEX data from NCAR’s Geoscience Data Exchange (GDEX) and make some diagnostic plots
5. Access CONUS 404 data from NCAR’s GDEX and make some diagnostic plots
6. Access data from the `SAAG’ dataset hosted on NCAR’s GDEX and make diagnostic plots
7. Benchmark data access speeds from the NCAR data origin using the DART reanalysis dataset and make diagnostic plots
8. Benchmark data access speeds from the NCAR’s data origin, when the data is accessed from the OSPool’s access point AP40
9. Run temperature bias-correction workflow on
   • a) Stampede3
   • b) NCAR’s Casper
10. Compute climatological average of daily temperature data using the geocat-comp package
11. Access CMIP6 zarr data from (> 27 Global Climate Models) AWS and
    • a) plot the evolution of Global Mean Surface Temperature with time. The model projections were also compared with the HadCRUT5 observational dataset.
    • b) compute Equilibrium Climate Sensitivity.

Machine learning workflows¶

1. Use logistic regression to predict Nino3.4 indices in advance. The training data are Sea Surface Temperature values and observed nino indices hosted on NCAR’s GDEX.

How to contribute¶

We welcome contributions from anyone who has an NCAR HPC account (Casper/Derecho access) and uses GDEX data! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b example/my-amazing-example)
3. Add your notebook/ workflow
4. Submit a pull request with a brief explanation of the changes

Support¶

Issues: Please report bugs and feature requests via GitHub Issues