Monsoon low-pressure system (LPS) tracks in ERA5 over India (1979-2019) with added environmental variables

Derived from the LPS v3.0 dataset (https://doi.org/10.5281/zenodo.7568990). Filtered to monsoon LPSs (majority of track lifetime between June and September), with genesis over the Bay of Bengal and making landfall over India. Temporal resolution also reduced from hourly to six-hourly. This dataset accompanies the paper "Using interpretable gradient-boosted decision-tree ensembles to uncover novel dynamical relationships governing monsoon low-pressure systems" (DOI to follow).

Aside from the core variables described in the main LPS dataset (linked above), this version includes a large number of environmental variables, listed below. All are computed from ERA5 unless otherwise stated, "mean" means that the variable is computed as an average within 400 km of the LPS centre, "mcz" means that the variable is computed as an average in the box [75-85°E, 18.5-27°N].
mean_u200: 200 hPa zonal wind (m s^-1)
mean_u850: 850 hPa zonal wind (m s^-1)
mean_skt: surface temperature (K)
mean_land_frac: fraction of area within 400 km that is over land
mcz_tcwv: mean total column water vapour over monsoon trough (kg m^-2)
vortex_depth: mean_vort_500 x mean_vort_700/mean_vort_850²
over_land: flag for LPS centre (Boolean)
dvo850_dt: rate of change of mean_vort_850 (10^-5 s^-1 day^-1) 
acc_land_time: accumulated time where over_land = True (hours)
total_land_time: final value of acc_land_time} for a given LPS (hours)
qshear_850: meridional shear of 850 hPa specific humidity over India (m³ m^-3 (°)^-1)
ushear_850: meridional shear of 850 hPa zonal wind over India (m s^-1 (°)^-1)
mean_cape: CAPE (J kg^-1)
mcz_cape: mean CAPE over the monsoon trough (J kg^-1)
mean_dthetae_dp_900_750: d(theta_e)/dp between 900 and 750 hPa, a measurement of atmospheric stability (K hPa^-1)
mean_dthetae_dp_750_500: d(theta_e)/dp between 750 and 500 hPa, a measurement of atmospheric stability (K hPa^-1)
mean_land_skt: land surface temperature (K; NaN over ocean)
mean_sst: sea surface temperature (K; NaN over land)
mean_swvl1: soil moisture in the top layer (m³ m^-3; <7 cm; NaN over ocean)
mean_swvl2: soil moisture in the second layer (m³ m^-3; 7-28 cm; NaN over ocean)
mean_swvl1_grad: mean absolute horizontal gradient of mean_swvl1 (m³ m^-4)
mean_swvl2_grad: mean absolute horizontal gradient of mean_swvl2 (m³ m^-4)
olr_90: 90th percentile of negative OLR (i.e. ~90th percentile of cloud top height) (W m^-2)
olr_75: 75th percentile of negative OLR (W m^-2)
olr_50: 50th percentile of negative OLR (W m^-2)
qshear_850_background: qshear\_850 averaged over the previous ten days (m³ m^-3 (°)^-1)
ushear_850_background: ushear\_850 averaged over the previous ten days (m³ s^-1 (°)^-1)
mean_q_850: 850 hPa specific humidity (m³ m^-3)
orography_height: elevation of land surface under LPS centre (m)
peak_vorticity: largest value of mean_vort_850} attained by a given LPS (10^-5 s^-1)
reached_peak: False if peak\_vorticity has not been reached yet, else True
mean_prcp_400: mean precipitation rate within 400 km of the LPS centre over the next six hours (kg m^-2 s^-1)
mean_prcp_800: mean precipitation rate within 800 km of the LPS centre over the next six hours (kg m^-2 s^-1)
max_prcp_400: maximum precipitation rate within 400 km of the LPS centre over the next six hours (kg m^-2 s^-1)
max_prcp_800: maximum precipitation rate within 800 km of the LPS centre over the next six hours (kg m^-2 s^-1)
mean_vimfd_400: vertically integrated moisture flux convergence (kg m^-2 s^-1)
mean_v200: 200 hPa meridional wind speed (m s^-1)
mean_v500: 500 hPa meridional wind speed (m s^-1)
mean_v850: 850 hPa meridional wind speed (m s^-1)
mean_u500: 500 hPa zonal wind speed (m s^-1)
zonal_speed: zonal (x) component of LPS propagation velocity (m s^-1)
merid_speed: meridional (y) component of LPS propagation velocity (m s^-1)
mean_prcp_imerg: as mean_prcp_400 but computed using IMERG data, rather than ERA5 (kg m^-2 hr^-1)

qshear_850, ushear_850 and their backgrounds are averaged over 5° longitude either side of the LPS centre, with the gradient computed between 10°N and 27°N, reflecting the moisture and zonal wind gradients across the monsoon region.