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.