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

An updated version (version 2) of this SyCLoPS dataset will be launched by the end of September, 2024

This dataset was generated by the first version of the System for Classification of Low-Pressure Systems (SyCLoPS). This archive contains the classified global low-pressure system (LPS) dataset ("SyCLoPS_classified.parquet") from SyCLoPS based on ERA5, as well as other files required to reproduce the results using ERA5 or any other large-scale datasets. The classified LPS dataset contains all detected global LPS tracks from 1979-2022, and each LPS node in the LPS tracks is given an LPS label based on SyCLoPS's classification.

SyCLoPS is an all-in-one detection and classification framework that combines multiple sole-purpose low-pressure system (LPS) detectors. It is developed atop the TempestExtremes (TE; Ullrich & Zarzycki, 2017; Ullrich et al., 2021) software package with a standalone Python classifier ("SyCLoPS_classifier.py"). It is tuned and subsequently applied to the ERA5 reanalysis. The classifier assigns 16 different types of LPS node labels and 4 types of high-impact LPS track labels following a single intuitive workflow, as illustrated in "SyCLoPS_workflow.png." No topographical, latitudinal, or temporal restrictions need to be applied in order to use this framework. Details of SyCLoPS are described in the paper titled The System for Classification of Low-Pressure Systems (SyCLoPS): An All-in-One Objective Framework for Large-scale Datasets. The preprint version of the paper is available here: https://doi.org/10.22541/essoar.171319442.20959523/v1

The following script files are provided:

1. The classifier: “SyCLoPS_classifier.py”
2. Blob-node pairing and LOWSIZE calculator: “LOWSIZE_pair_cal.py”
3. Example uses of the classified catalog: “SyCLoPS_examples.py”
4. The TempestExtremes command lines: "TE_commands.sh“

The following data are provided:

1. The input LPS catalog: "SyCloPS_input.paqruet"
2. The output classified LPS catalog: "SyCLoPS_classified.parquet"
3. The column documentation for the two catalogs: "SyCLoPS_headers.png"
4. The additional track information file: "Additional_Track_Info.csv"
5. The labled size blobs of each year: "size_blobs_1979_2022.tar.gz"
6. The labled precipitation blobs of each year: "preci_blobs_1979_2022.tar.gz"

The size and precipitation blob tag file each has 44 compressed nc files for each year. Decompress each nc file before using them to optimize computation performance. Blobs are tagged with five different labels (ID numbers from 1-5). Blobs associated (paired) with 1 = TC nodes in TC tracks; 2 = TD and TLO (TLO(ML), TD(MD), TLO, and TD) nodes in MS tracks; 3 = STLC nodes in STLC tracks; 4 = PL nodes in PL tracks; and 5 = other LPS nodes. Users may alter this ID system according to alternate definitions. See details in the supporting information Text S6 of the SyCLoPS paper.

The TE software is available at: https://github.com/ClimateGlobalChange/tempestextremes, and its documentation can be found here: https://climate.ucdavis.edu/tempestextremes.php 

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