Data reconstruction and homogenization for reducing uncertainties in high-resolution climate analysis in Alpine regions

Analysis of climatic series needs pre-processing to attain spatial- and time-consistent homogeneity. The latter, in high-resolution investigations, can rely on the strong correlations among series, which in turn requires a strict fulfilment of the quality standard in terms of completeness. Fifty-nine daily precipitation and temperature series of 50 years from Trentino, northern Italy, were pre-processed for climatic analysis. This study describes: (1) the preliminary gap-filling protocol for daily series, based on geostatistical correlations on both horizontal and vertical domains; (2) an algorithm to reduce inhomogeneity owing to the systematic snowfall underestimation of rain gauges; and (3) the processing protocol to take into account any source of undocumented inhomogeneity in series. This was performed by application of the t test and F-test of R code RHtestV2. This pre-processing shows straightforward results; correction of snowfall measurements re-evaluates attribution of patterns of altitudinal trends in time trends; homogenization increases the strength of the climatic signal and reduces the scattering of time trends, assessed over a few decades, of a factor of 2