Key points are not available for this paper at this time.
Abstract Current interest into past climate change and its potential role for changes in the environment call for spatially distributed climate datasets of high temporal resolution and extending over several decades. To foster such research, we present a new gridded dataset of daily minimum and maximum temperature covering Austria at 1-km resolution and extending back till 1961 at daily time resolution. To account for the complex and highly variable thermal distributions in this high-mountain region, we adapt and employ a recently published interpolation method that estimates nonlinear temperature profiles with altitude and accounts for the non-Euclidean spatial representativity of station measurements. The spatial analysis builds upon 150 station series in and around Austria (homogenised where available), all of which extend over or were gap-filled to cover the entire study period. The restriction to (almost) complete records shall avoid long-term inconsistencies from changes in the station network. Systematic leave-one-out cross-validation reveals interpolation errors (mean absolute error) of about 1 °C. Errors are relatively larger for minimum compared to maximum temperatures, for the interior of the Alps compared to the flatland and for winter compared to summer. Visual comparisons suggest that valley-scale inversions and föhn are more realistically captured in the new compared to existing datasets. The usefulness of the presented dataset (SPARTACUS) is illustrated in preliminary analyses of long-term trends in climate impact indices. These reveal spatially variable and eventually considerable changes in the thermal climate in Austria.
Hiebl et al. (Fri,) studied this question.