Homogenization of incoming solar radiation measurements over Poland with satellite and climate reanalysis data

Well-maintained and regularly calibrated measuring instruments provide the most accurate solar radiation data. This extremely valuable research material makes it possible, among others, to analyse variability in solar radiation over the long term and its dependence on other atmospheric state elements such as cloud cover and atmospheric aerosol concentration. Unfortunately, ground-based measurements of solar radiation are often subject to various errors which are very difficult to detect. This is why quality control procedures and homogenisation of data are essential and should be performed prior to further analyses. This paper presents a method for quality control and homogenization of solar radiation data, which builds on the bias-based quality control (BQC) method (Urraca et al., 2017), and is tailored specially for detecting single erroneous daily values, and very long periods of small errors. The method was tested for 16 ground-based stations located in Poland for the period 1991–2015. In comparison with the number of errors detected by the BQC method, the number of detected errors increased significantly: 130 to 2890 more erroneous days were detected at each station. Consequently, the number of inhomogeneous data sets was reduced from 8 to 3 stations. The values on the days considered as erroneous were replaced with debiased values originating from the Surface Solar Radiation Data Set – Heliosat, Edition 2 (SARAH-2). The presented methodology can be also of use in any other places, especially those with many single erroneous days and no metadata publicly available.