The role of temporal resolution of meteorological inputs from reanalysis data in estimating air humidity for modelling applications

Reliable values of relative humidity are basic inputs for modelling in many disciplines and in the most disparate scientific fields. Unfortunately, humidity variables remain less focused than other meteorological parameters and generally suffer from considerable uncertainty mainly due to the fact that their observations are not widely available, fostering the use of the observations of other meteorological quantities to estimate them. The aim of this work is to assess the loss in daily maximum and minimum relative humidity accuracy when sampling interval becomes coarser than one hour. For this purpose, meteorological data from the ERA5 dataset, the most advanced reanalysis product released by the European Centre for Medium-Range Weather Forecasts (ECMWF), are used. Among the many advantages of ERA5 over the previous release ERA-Interim are the finer temporal resolution and data archived at the hourly time step. Near-surface relative humidity is derived using 1- and 3-hourly reanalysis data of 2-m temperature, 2-m dew-point temperature and surface pressure. Deviations from the actual values, as obtained from reference measures acquired at 15 minute intervals, are evaluated. Results show that the biases of the ERA5-based values are consistently reduced compared to its predecessor and that the performance of the calculated 1-hourly time resolution relative humidity data is almost equivalent to using observations. Reducing the sampling interval from three to one hour provides a significant improvement in data quality. The results indicate significant increases in errors in the estimates when the temporal resolution of the meteorological inputs becomes coarser than one hour, exceeding also the numerical and approximation errors due to simplifying assumptions in the theoretical and empirical formulas used. The positive impact of improving temporal resolution from ERA-Interim to ERA5 reanalysis is also quantified by the number of the correct relative humidity extremes increasing by 50%.