A new surface solar radiation database of 74 daily series is set up for the Piedmont region (northwest Italy) for the 1990-2016 period. All the series are subjected to a detailed quality control, homogenization and gap-filling procedure and are transformed into relative annual/seasonal anomaly series. Finally, a gridded version (0.5°x0.5°) of the database is generated. The resulting series show an increasing tendency of about +2.5% per decade at annual scale, with strongest trend in autumn (+4% per decade). The only exception is winter, showing a negative but not significant trend.Considering the plain and mountain mean series, the trends are more intense for low than for high elevations with a negative vertical gradient of about -0.03% per decade per 100m at annual scale and values up to -0.07% per decade per 100m in spring.Focusing on clear days only (selected by CMSAF-COMET satellite data over the 1991-2015 period), trend significance strongly increases and both low and high elevation records exhibit a positive trend in all seasons. However, the trends result slightly lower than for all-sky days (with the only exception of winter). The differences observed under clear-sky conditions between low and high elevations are more pronounced in winter, where the trend shows a negative vertical gradient of about -0.1% per decade every 100m.Overall, this paper shows how a high station density allows, performing a more detailed quality control thanks to the higher performances in detecting the inhomogeneities with higher data availability and capturing regional peculiarities otherwise impossible to observe.

No abstract

No abstract