The System for Classification of Low-Pressure Systems (SyCLoPS) Dataset (Based on ERA5)

A Newer Version (Ver 6) has been published.

This is the ERA5 System for Classification of Low-Pressure Systems (SyCLoPS) dataset version 5. Details of SyCLoPS algorithms are described in the paper titled The System for Classification of Low-Pressure Systems (SyCLoPS): An All-in-One Objective Framework for Large-scale Data sets published on J. Geophys. Res. Atm.: https://doi.org/10.1029/2024JD041287

We keep the most up-to-date version of the SyCLoPS user manual online at THIS PLACE. The latest version of the required software Tempestextremes (TE) can be attained from THIS GitHub PAGE.

Known issues

1. The master branch of TempestExtremes now lacks the full ability to calculate parameters with missing value (e.g., 1e20) data. Therefore, it is now not directly applicable to some model outputs and reanalyses that have missing values where the data level is below the surface. We are working on this issue and users can expect a newer TE version with fixes in the near future. For now, users can use a specific fork of TempestExtremes, which can be found here: https://github.com/yepkids/tempestextremes. This fork provides a temporary solution that adds missing value support for operators used by SyCLoPS. Note that this is not a stable release, and please report any problems with this fork to Yushan Han (yshhan@ucdavis.edu).
   
   

2. In SyCLoPS_classifier.py, the alternative criteria for the regional data mode have two typos in the inequality notations (this bug does not affect the default classification process). Please refer to the user manual for further details on how to fix this manually.
   
   
3.  Please note that some datasets may have different variable names and units than those used in SyCLoPS. Remember to change those names and convert units (e.g., from geopotential height to geopotential) in the codes. We'll update the classifier code to more automatically detect variable names and units in the next release.

Major updates and bug fixes in this version:

1. We have fixed typos in the TE command files and added more tips/comments on operations.
2. TE's StitchNodes now output a csv file, which is more feasible than the txt format previously used. The Python classifier can now process the TE's LPS track file with higher speed. One may rerun the StichNodes command in the latest "TE_commands.sh" to produce a csv LPS track file.
3. Removed the computation of the WS parameter in TE commands (made it optional) as it does not affect the classification results.
4. The Python classifier now supports parallel computation in Python version >= 3.10 with much lower memory requirements (this also works for lower version Python). It requires installation of the "multiprocess" package. 
5. The Python classifier now supports more custom options for execution, including options to support regional models, skip WS and IKE parameter operations, and specify time resolution.
6. The latest TE version and the Python classifier now support missing/fill values in the datasets (most climate model outputs use 1e20 as the default fill value). Fill values are ignored or integrated as needed throughout the process.
7. A slight change in the definition of extratropical transition (EXT) to be compatible with data of any time resolution. This results in a very small increase in the total number of EXT cases.

The following files are updated and reuploaded in this version:

1. The output classified LPS catalog: "SyCLoPS_classified.parquet" (See the SyCLoPS manual, Chapter 3  on how to load and use this catalog)
2. The shell script containing TE command lines: "TE_commands.sh"
3. The SyCLoPS classifier Python script: “SyCLoPS_classifier.py”
4. Additional TE commands for tagging (masking) blobs: "TE_optional.sh"

The following files are not changed and can be obtained from version 4:

1. The input LPS catalog: "SyCLoPS_input.parquet"
2. The labeled size blobs of each year: "size_blobs_1979_2022.tar.gz"
3. The labeled precipitation blobs of each year: "preci_blobs_1979_2022.tar.gz"
4. A sample TE input file that contains a list of required ERA5 variable filenames: "DetectNodes_inputfile_example.txt"
5. The optional blob tagging tool (used with "TE_optional.sh"): “Blob_idtag.py"

Please contact Yushan Han (yshhan@ucdavis.edu) if you have questions about the SyCLoPS framework. Please contact Paul Ullrich (paullrich@ucdavis.edu) if you have any questions about the TE software.

Acknowledgments: We thank Kian Huang at the University of Utah for providing feedback and insightful suggestions that helped develop this new dataset version.

See below for a table of atmospheric variables required for SyCLoPS and a flowchart of the classification process. See the SyCLoPS manual for more details.