Changelog

Temperature

• TX and TN are made available for the first time in seNorge
• The procedure to compute the sub-regional pseudo-background fields has been revised (with respect to seNorge2). The calculation of the regional pseudo-background field is now based on much more sub-regional fields than before. At the same time, the temperature spatial trends depend only on elevation (the dependence from the geographical coordinates has been removed). As a result, the pseudo-background field is a more accurate representation of the regional temperature field.
• The optimal interpolation (OI) procedure is performed on a gridpoint-by-gridpoint basis so to allow for the use of a variable horizontal de-correlation length in the definition of the background error correlation matrix. As a result, the effective resolution of the analysis temperature field varies throughout the domain: in data-dense area the effective resolution is higher and more details are represented, in data-sparse areas the effective resolution is coarser.
• The geographical parameters used in the OI are: geographical coordinates, elevation a.m.s.l. and land area fraction.

Precipitation

• The identification of events is no longer needed. The distinction between precipitation and non-precipitation regions is done within the iterative OI
• The precipitation observations are adjusted for the wind-induced undercatch (correction similar to Wolff, M. A., Isaksen, K., Petersen-Øverleir, A., Ødemark, K., Reitan, T., and Brækkan, R.: Derivation of a new continuous adjustment function for correcting wind-induced loss of solid precipitation: results of a Norwegian field study, Hydrol. Earth Syst. Sci., 19, 951-967, https://doi.org/10.5194/hess-19-951-2015, 2015.)
• The iterative OI scheme has been modified to use the relative anomalies: (observed value) / (long-term monthly average obtained from a dataset of high-resolution hindcast precipitation fields)
• The choice of the horizontal scales used by the iterative OI scheme has been automatized and made dependent on the station distribution. The iterative OI has been forced towards finer spatial scales.
• The iterative OI inner loop for the optimization of the vertical decorrelation length scale of the background error covariance matrix has been removed.