Scalable radar-driven approach with compact gradient-boosting models for gap filling in high-resolution precipitation measurements

Gauge-centered QPECommon data This dataset supports the study "Scalable radar-driven approach with compact gradient-boostingmodels for gap filling in high-resolution precipitation measurements". It provides an organisation of the radar based precipitation estimates and gauge based precipitation measurements, as provided by the German Weather Service (DWD). The data is compiled iso that precipitation records from the gauge stations are combined with all time-synchronous QPECommon data available within a 5 kilometer range around the gauge. The data is readily readable and interpretable by the suggested model. included files For all of the DWD`s rain gauge stations, one csv file is provided the file contains the precipitation meassurements at the station the file contains all synchronous QPECommon evaluations for the 10km x 10km area surrounding the station. data structure range: 2021-01-01 : 2024-12-31 precipitation sample rate: 10 minutes QPE sampling rate: 5 minutes (separated into QPE0 and QPE5, sampled at 10 minutes rate) file structure file is namend id_ID.csv (for ID=310, the name would be id_310.csv) rain gauge readings are stored to column ID_RR in id_ID.csv 10-minutes-aligned QPECommon estimates at coordinates x and y are stored to column 9x0000yQPE0 5-minutes-off QPECommon estimates at coordinates x and y are stored to column 9x0000yQPE5 purpose This dataset provides a properly formatted training data base, in order to fit the suggested imputation model to the german weather services rain gauge network.