then recorded in NetCDF format. After these controls, data have been manually checked and an extended procedure to  reconstruct the temperature time series for missing data has been applied. The spatiotemporal method used to reconstruct the data has been linear interpolation for 1-hour gaps, and the Empirical Orthogonal Function algorithm (EOF) for the 2-hour and longer gaps. The introduction of the complete and homogeneous data set of hourly reanalysis ERA5 (ECMWF), allowed filling the longest gaps with statistical and physical consistency. The final product of this work are continuous station time series of hourly temperatures that will be available to the public at the end of 2020 , while a daily version of the original time series is already available at the regional website (https://annali.regione.umbria.it/).
Keywords --- Agro-meteorology, Quality control, Missing data, ERA5, Empirical Orthogonal Function