Platform data

1

Search results for: "keyword:"XGBoost""

Overview

1-2 of 2 datasets.

(Click for the detailed view)

» Training datasets for visibility forecasting based on station, Meteosat Second Generation (MSG) and COSMO-REA6 at three DWD station locations created by: Kai Richter (last update: 2025-07-01)

» Forecasted visibility values of detected fog events between 11/2009 and 02/2016 at DWD stations 1639, 198 and 2044 as output of XGBoost modeling created by: Kai Richter (last update: 2025-06-30)

Detailed view

Training datasets for visibility forecasting based on station, Meteosat Second Generation (MSG) and COSMO-REA6 at three DWD station locations

view metadata

Dataset ID:	587
Last updated:	2025-07-01
Dataset creators:	Kai Richter
Contact:	Maik Dobbermann Kai Richter
Temporal coverage:	2009-11-15 00:00:00 - 2016-02-26 23:00:00 - time zone: Europe/Berlin
Geographic coverage:	Europe
Abstract:	https://vhrz669.hrz.uni-marburg.de/lcrs/publicatio... Abstract: https://vhrz669.hrz.uni-marburg.de/lcrs/publications.do?citid=580
Additional info:	For DWD stations 1639, 198 and 2044, four differen... Additional info: For DWD stations 1639, 198 and 2044, four different combinations of parameters have been compiled, including "Station", "MSG+COSMO", "Station+MSG+COSMO", and "Visibility-Only". This adds up to 12 distinct training datasets, that are spanned by the included entitys. Value progressions were smoothed with Gaussian weighting (Moving average window = 4; Sigma = 1). For all datasets, attribute "vis_delay" is the target variable. All other variables are predictor variables, except attribute "TIMESTAMP".
Keywords:	\| fog forecasting \| fog horizontal visibility \| XGBoost \|
Intellectual rights:	lcrs.de

Entity name:	model_data_final_198_vars_msg_timeperiod_km_clipped_trend1h_smoothed
Entity type:	dataTable
Attributes:	Mapping date (timestamp) (Unitless) Visibility (vis) [Kilometer] IR 3.9 µm channel (IR_039_msg) [Degree Kelvin] IR 8.7 µm channel (IR_087_msg) [Degree Kelvin] IR 9.7 µm channel (IR_097_msg) [Degree Kelvin] IR 10.8 µm channel (IR_108_msg) [Degree Kelvin] IR 12.0 µm channel (IR_120_msg) [Degree Kelvin] WV 6.2 µm channel (WV_062_msg) [Degree Kelvin] WV 7.3 µm channel (WV_073_msg) [Degree Kelvin] Channel Difference IR 3.9 - IR 10.8 (Diff_IR_039-IR_108) [Degree Kelvin] Channel Difference IR 8.7 - IR 10.8 (Diff_IR_087-IR_108) [Degree Kelvin] Channel Difference IR 10.8 - IR 12.0 (Diff_IR_108-IR_120) [Degree Kelvin] Channel Difference IR 3.9 - WV 7.3 (Diff_IR_039-WV_073) [Degree Kelvin] Total cloud cover (CLCT_csm) [Percent] Height of the planetary boundary layer (H_PBL_csm) [Meter] Pressure at mean sea level (PMSL_csm) [Pascal] 2m specific humidity (QV_2M_csm) [Kilogram per kilogram] 2m relative humidity (RELHUM_2M_csm) [Percent] Net short wave radiation flux at the surface (SOBS_RAD_csm) [Watt per square meter] 2m temperature (T_2M_csm) [Degree Kelvin] Soil temperature (T_SOIL_csm) [Degree Kelvin] Net long wave radiation flux at the surface (THBS_Rad_csm) [Watt per square meter] 10m zonal wind speed (U_10M_csm) [Meter per second] Dewpoint depression (dewpoint_dep_csm) [Degrees Celsius] IR 3.9 µm channel trend (IR_039_msg_trend_1h) [Degree Kelvin] IR 8.7 µm channel trend (IR_087_msg_trend_1h) [Degree Kelvin] WV 6.2 µm channel trend (WV_062_msg_trend_1h) [Degree Kelvin] Visibility trend (V_VV_dwd_trend_1h) [Kilometer] 2m temperature trend (T_2M_csm_trend_1h) [Degree Kelvin] 2m specific humidity trend (RELHUM_2M_csm_trend_1h) [Percent] Dewpoint depression trend (dewpoint_dep_csm_trend_1h) [Degree] Visibility delay (vis_delay) [Kilometer]

» back to overview

Forecasted visibility values of detected fog events between 11/2009 and 02/2016 at DWD stations 1639, 198 and 2044 as output of XGBoost modeling

view metadata

Dataset ID:	586
Last updated:	2025-06-30
Dataset creators:	Kai Richter
Contact:	Maik Dobbermann Kai Richter
Temporal coverage:	2014-10-01 00:00:00 - 2016-02-26 00:00:00 - time zone: Europe/Berlin
Geographic coverage:	Europe
Abstract:	https://vhrz669.hrz.uni-marburg.de/lcrs/publicatio... Abstract: https://vhrz669.hrz.uni-marburg.de/lcrs/publications.do?citid=580
Additional info:	Each entity represents the model output (test data... Additional info: Each entity represents the model output (test data) for one model variant (Station, MSG+COSMO, Station+MSG+COSMO, Visibility-Only) of one DWD station (198 - Artern; 1639 - Gießen/Wettenberg; 2044 - Harzgerode). Each entity includes csv files of both hourly resoluted and Gaussian-interpolated minute-resoluted visibility value series. Each csv file represents one fog event. Attribute(s) Attribute Name -- Attribute Label Unit Attribute definition "TIMESTAMP" -- Measurement Timestamp Datetime Timestamp of measurement "Truth" -- Original visibility values km Original visibility values in 60 min "Predicted" -- Forecasted visibility values km Forecasted visibility values in 60 min predicted by XGBoost model "vis" -- Visibility values km Visibility values at the current timestamp
Keywords:	\| fog forecasting \| fog horizontal visibility \| XGBoost \|
Intellectual rights:	lcrs.de

Entity name:	modeloutput_198_msg_cosmo
Entity info:	Model output for visibility nowcasting of MSG+COSMO variant, DWD station 198 (Artern)
Entity type:	otherEntity

Entity name:	modeloutput_198_station_msg_cosmo
Entity info:	Model output for visibility nowcasting of Station+MSG+COSMO variant, DWD station 198 (Artern)
Entity type:	otherEntity

» back to overview

1

« Home To the top »

Data Services
List datasets

Quick search

Publications:
Datasets:

Latest datasets

» Training datasets for visibility forecasting based on station, Meteosat Second Generation (MSG) and COSMO-REA6 at three DWD station locations created by: Kai Richter

Latest publications

» Machine Learning-supported visibility forecasting by combining station, Meteosat and reanalysis data

System message

It is strongly recommended that you activate JavaScript in your browser to properly use this site.