Key points are not available for this paper at this time.
This article describes a two‐step spatial interpolation method for 2 m temperature. A scale‐separation approach based on statistical interpolation (specifically a modification of classical optimal interpolation (OI)) has been implemented. The method presented makes use of in situ observations only and has been used to build the seNorge2 dataset of hourly and daily averaged temperature over the Norwegian mainland. The main original contribution is the blending of several temperature fields, each representing the OI background in a subregion, into a single deterministic regional (i.e. domain‐wide) temperature background, which may cover a wide geographic area. As required by climate and hydrological applications, temperature analyses are produced with the highest effective resolution compatible with the observational network. A spatial consistency test has been included, so that the method is robust from a statistical point of view. The quality of the long‐term seNorge2 dataset has been evaluated, both by summary statistics and by analysing challenging case studies. The blending of subregional background fields does not introduce discontinuities in the temperature analysis. The effects of station density variations in time and space have also been addressed. The evaluation shows that the predicted temperature fields are indeed unbiased estimates of the true state, except for extremely low temperatures, where a warm bias is present. For both hourly and daily averaged temperature, the precision of the estimates at grid points varies between 0.8 and 2.4 °C.
Lussana et al. (Fri,) studied this question.