Hi Everyone,
xCDAT (Xarray Climate Data Analysis Tools) v0.7.0 was released last week! We usually recommend using the newest version to get the latest enhancements and bug fixes.
Thank you to all the contributors who open or contribute to GitHub Issues, ask questions, and provide input! If you’re interested, we always welcome new contributors 
Please feel free to post questions or comments in the GitHub Discussions forum.
Changelog –
- This minor release includes enhancements to the performance of the Regrid2 API and fixes Regrid2 to align the behavior of how missing values are handled with CDAT.
- Full changelog
Resources –
Best,
Tom
2 Likes
Thanks for the efforts on this open package @tomvothecoder.
For those who have yet to install and try out xCDAT and never used CDAT, specifically those not inside the LLNL tent, can you outline what you see as the use cases xCDAT is aimed at? And possibly compare and contrast to other approaches like ESMvalTool, if you can? Thanks
Hi @Thomas_Moore – thanks for the question.
The xcdat developers have not have a lot of experience with ESMValTool, but this is my understanding doing a quick review of ESMValTool’s documentation: it does appear that some of the core functionality overlaps with xcdat. ESMValTool leverages iris, but xcdat has been intentionally designed to extend xarray and be interoperable with the growing xarray ecosystem (note that it does appear that ESMValTool has some experimental compatibility with xarray).
Part of the difference is also scope: ESMValTool seems to emphasize end-to-end data processing capabilities (finding, selecting, and fixing data, using basic climate utilities, and developing plots/metrics/diagnostics via recipes). xcdat on the other hand tries to maintain scope as a climate utility that does the basics well (e.g., regridding, temporal and spatial averaging, calculation of departures, etc.). With that said, xcdat is being adopted as a data processing engine within the PMP and E3SM diags to do basic climate data operations.
Within LLNL, xcdat (and xarray) is becoming a basic, default tool for routine climate research analysis (essentially to wrangle/load/regrid/fix data that goes into analyses). As you noted, this package follows from CDAT (see a basic xCDAT-CDAT mapping). Our goal is to continue to make xcdat a more robust tool and to add maintainable functionality/features that help with basic climate data processing and analysis.
1 Like
Thank you @pochedley for that detailed answer. And thank you, @tomvothecoder, and the others on the team for sharing this work openly.
I’m probably in the minority here but despite my level of interest and use of these kinds of workflows I struggle sometimes to see how the global efforts across this area all could fit together, where there’s overlap, synergy, dissonance, and duplication.
It’s hard to build such packages and harder to keep them up-to-date and growing. So kudos to the xCDAT team and LLNL for the investment in people, time, and effort.