Abstract
In recent decades, automated sensors for sunshine duration (SD) measurements have been introduced in meteorological networks, thereby replacing traditional instruments, most prominently the Campbell-Stokes (CS) sunshine recorder. Parallel records of automated and traditional SD recording systems are rare. Nevertheless, such records are important to understand the differences/similarities in SD totals obtained with different instruments and how changes in monitoring device type affect the homogeneity of SD records. This study investigates the differences/similarities in parallel SD records obtained with a CS and two automated SD sensors between 2007 and 2016 at the Kanzelhöhe Observatory, Austria. Comparing individual records of daily SD totals, we find differences of both positive and negative sign, with smallest differences between the automated sensors. The larger differences between CS-derived SD totals and those from automated sensors can be attributed (largely) to the higher sensitivity threshold of the CS instrument. Correspondingly, the closest agreement among all sensors is found during summer, the time of year when sensitivity thresholds are least critical. Furthermore, we investigate the performance of various models to create the so-called sensor-type-equivalent (STE) SD records. Our analysis shows that regression models including all available data on daily (or monthly) time scale perform better than simple three- (or four-) point regression models. Despite general good performance, none of the considered regression models (of linear or quadratic form) emerges as the “optimal” model. Although STEs prove useful for relating SD records of individual sensors on daily/monthly time scales, this does not ensure that STE (or joint) records can be used for trend analysis.
Highlights
Sunshine duration (SD) is an essential meteorological variable which has been routinely measured in meteorological practice since over 160 years (Sanchez-Lorenzo et al 2013)
The instrument itself was for the first time introduced as the BCampbell-Stokes recorder^ by Ellis (1888), and named following Curtis (1898) Bas the ‘Campbell Stokes’ by which justice is done to the two men whose joint creation it is.^ It is known within the community under this name ever since
The present study aims on closing this gap by evaluating 9 years of parallel measurements of sunshine duration by a CS and two automated systems (Kipp&Zonen CSD2 and Haenni Solar 111b) performed at the Kanzelhöhe Observatory for Solar and Environmental Research of the University of Graz, Austria (KSO, 46° 40′ 39′′ N, 13° 54′ 06′′ E, 1540 m.a.s.l.)
Summary
Sunshine duration (SD) is an essential meteorological variable which has been routinely measured in meteorological practice since over 160 years (Sanchez-Lorenzo et al 2013). Campbell developed the first recorder in 1853 and the systems description was published in Campbell’s own presentation in the British Meteorological Societies Report of the Council in 1857 (Curtis 1898). Stanhill (2003) and Sanchez-Lorenzo et al (2013) provide a comprehensive overview and history of what is known as the Campbell-Stokes sunshine recorder (hereinafter referred to as CS) and outline that many of the subsequent adjustments and improvements of the original design occurring between 1853 and 1880 have been described by Campbell himself in his BThermograph^ (Campbell 1883). There is ongoing debate about the individual contributions of Campbell and Stokes to what became known as the CS (see Sanchez-Lorenzo et al 2013; Stanhill 2003). For a brief description of the CS instrument, we refer the reader to section 2.1 of this manuscript, and for further details to Curtis (1898), Stanhill (2003), and Sanchez-Lorenzo et al (2013)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
