Datasets for the automatic identification of Foreshock Transients

These files contain observational data from the Cluster 1 spacecraft used for foreshock transient (FT) identification.

The initial training datasets are:

`train_ft_init.pkl`: 83 foreshock transient events from 2003–2009, selected from the Wang et al. (2013) catalog.

`train_sw_init.pkl`: 350 solar wind intervals from 2003–2009, selected from the Michotte de Welle et al. (2024) catalog.

The final training datasets are:

`train_ft_final.pkl`

`train_sw_final.pkl`

These final datasets include the initial training events together with additional events selected during the validation process.

The detected Foreshock transients from 2001 to 2011are:

`ft_2001_2011_updated.pkl`The catalog 'ft_2001_2011_btot_updated.csv' lists the start and end times of the detected FT events by the finalized model from 2001 to 2011. It also includes the 83 selected events from train_ft_init.pkl. Events marked as True in the "in_original_catalog" column correspond to these selected events from the original training set. The "recalled" column marks 1 if an event detected by the model is recalled from the original catalog.

Each .pkl file contains the following columns:

`Time`: timestamp in YYYY-MM-DD hh:mm:ss format

Magnetic field components, `Bx`, `By`, and `Bz`, in nT

Bulk velocity components, `Vx`,`Vy`,`Vz`, in km/s

Ion number density, `n`, in cm^-3

Ion temperature, `T`, in eV

Ion thermal pressure, `P`, in km^2 s^-2 cm^-3

Ion energy spectrogram with 31 energy channels, in keV cm^-2 s^-1 sr^-1 keV^-1