A machine-learning reference dataset for SO2 plumes observed by TROPOMI: uncertainties and emission estimates

This dataset relates to the SO2 plume detection dataset decribed in this paper (currenly in the discussion phase): https://egusphere.copernicus.org/preprints/2026/egusphere-2025-5900/#discussion File Description: Each CSV file contains all plumes of sulfur dioxide detected in TROPOMI level-2 data for each year using the machine learning algirothm described in the associated paper. The variables included in each file are: Variable Description Units or Format datetime The date and time of the beginning of the TROPOMI swath (same as the date and time provided in the TROPOMI level-2 files YYYY-MM-DD HH:MM:SS species The species of the plumes found (in this case, always SO2) string ('SO2') max_latitude The latitude of the maximum SO2 detection within the plume boundary degrees north max_longitude The longitude of the maximum SO2 detection within the plume boundary degrees east max_latitude_index The latitude array index relating to the assocaited TROPOMI file of the maximum SO2 detection within the plume boundary integer max_longitude_index The longitude array index relating to the assocaited TROPOMI file of the maximum SO2 detection within the plume boundary integer bbox_top The array index relating to the assocaited TROPOMI file of the top of the plume bounding box integer bbox_bottom The array index relating to the assocaited TROPOMI file of the bottom of the plume bounding box integer bbox_left The array index relating to the assocaited TROPOMI file of the left of the plume bounding box integer bbox_right The array index relating to the assocaited TROPOMI file of the right of the plume bounding box integer plume_max_val The maximum observation SO2 total vertical column within the plume boundary mol/m2 plume_mass Estimated mass of the detected plume kg plume_poly_50 The coordinates of the plume boundary degrees (array in json format) plume_poly_50_index The index of the coordinates of the plume boundary relating to the assocaited TROPOMI file integer array in json format ellipse_data Information to recreate the fitted ellipse around the plume. Information given refers to the bounding box of the plume (xc,yc),(d1,d2),angle file_wind Flag to indicate if wind data is provided in the level-2 TROPOMI file 0/1 False/True median_windspeed The median windspeed within the plume bounding box m/s min_windspeed The minimum windspeed within the plume bounding box m/s max_windspeed The maximum windspeed within the plume bounding box m/s std_windspeed The standard deviation of the windspeed within the plume bounding box m/s stderr_windspeed The standard error of the windspeed within the plume bounding box m/s plume_wd Wind direction within the plume degrees plume_length Estimated length of the primary axis of the detected plume meters plume_emission_estimate Estimated emission flux from the plume (based on the calculations presented in the associated paper) kg/hr bbox_pixel_areas Array of area of each pixel within the bounding box float array (m2) bbox_wind_v Array of the V wind vectors at for each pixel within the bounding box float array (m/s) bbox_wind_u Array of the U wind vectors at for each pixel within the bounding box float array (m/s) orig_file Name of the level-2 TROPOMI file used to detect the SO2 plume string