Identified, tracked and categorized jets, jet properties and trained SOM

This zip file has two subfolders: Jets and SOM. The files are all results of the two methods presented in the paper. The .parquet files represent dataframes, to be opened with a dataframe compatible library like Pandas or Polars. The .nc files are NETCDF4 files to be opened with xarray or the netcdf package (latter untested). Jets/all_jets_one_df.parquet: All jets found in ERA5 six-hourly wind speed fields between 01.01.1959 and 31.12.2022 included. The dataframe has the following columns: [time, lon, lat, s, lev, u, v, theta, sigma, alignment, ds, jet ID, s_low, ratio, is_polar] Jets/all_jets_over_time.parquet: Result of the tracking algorithm. Links a flag, which is unique per tracked jet over time, to one or several pairs of a timestamp and a jet ID to find the instantaneous jet core in all_jets_one_df.parquet. Jets/flags.parquet: Result of the tracking algorithm. A forward map from a timestamp and a jet ID to a flag. Jets/props_as_df_raw.parquet: The jet properties for each jet in Jets/all_jets_one_df.parquet, except those extracted from the tracking results. Jets/props_as_df_uncat.parquet: The jet properties for each jet in Jets/all_jets_one_df.parquet . Jets/props_as_df.parquet: The jet properties for each timestamp and jet category. Jets/props_as_df_extras.parquet: The jet properties for each timestamp and jet category, plus the double jet index. Jets/all_props_over_time.parquet: The jet properties reordered in the same order as Jets/all_jets_over_time.parquet, sorted by flag first and timestamp second. Used to determine the COM speed of the jets. Jets/s_q.nc: Climatology of wind speed quantiles in the NAtl sector for quantiles ranging between 0.6 and 0.99. Used to determine thresholds. SOM/som_6_4_pbc_euclidean.npy: the weight matrix of each cluster in the SOM. This can be used to recreated the SOM python object. SOM/labels_som_6_4_pbc_euclidean.nc: The assignment of each timestamp in ERA5 (six hourly, 1959-2022, restricted to JJA) converted to a xarray.DataArray to add some useful metadata like a time coordinate. SOM/SOM/labels_som_6_4_pbc_euclidean.nc: The center of each cluster in the SOM, for the variables u, v, U and theta.