Daylight solar radiation AMY data derived from satellite remote sensing: Validation against ground measurements and comparison with TMYs

The standardised weather files commonly used for building simulation are compiled from many years of data. Particular to a specific location, these standardised weather files are generally known as Typical Meteorological Years (TMYs). In contrast, Actual Meteorological Years (AMYs) comprise data for a specific site over a defined period of an actual calendar year. The Copernicus Atmosphere Monitoring Service (CAMS) provides freely available satellite-derived radiation data covering Europe, Africa, the Middle East and parts of South America. CAMS data were used as the basis for solar radiation AMYs. For three locations in Europe, multi-year AMYs are used to test the suitability of TMY files as a reliable representation of prevailing sun and sky conditions. Examples are given for London (Gatwick), Rome (Fiumicino) and Stockholm (Arlanda), where, for all three locations, a full decade of AMY data at both 15 min and 1 h time-steps are evaluated alongside four contending standardised TMY files. For all three locations, the decade of AMY data proved to be surprisingly homogeneous, whereas the four TMYs were at variance with each other, and markedly dissimilar to the AMYs. Consequently, the authors propose a reconsideration of the use of TMYs for compliance purposes in particular, and building simulation in general. Given the unexpected findings, and their potentially far-reaching implications, the weather file evaluation is preceded by a detailed validation of CAMS-derived illuminance data against ground measurements taken in the UK. The results of the validation revealed remarkably good agreement between the CAMS-derived and ground measured illuminance data. Practical applications This paper provides compelling evidence that the methods currently used to select solar radiation data for TMYs result in standardised weather files that do not faithfully represent actually occurring conditions over a recent decade. A more reliable method for the evaluation of ‘typical’ annual profiles of solar radiation is described. The findings have relevance for the selection and curation of solar radiation data for all building simulation applications. In addition to supporting the basis of the TMY evaluation, the validation of CAMS-derived illuminance data revealed that CAMS more generally can serve as a valuable – and freely-available – daylight resource for a variety of practical applications. These include the in-situ validation of CBDM metrics and the generation of boundary daylight conditions for light-dosimetry field studies. Or indeed any application where reliable recent data on daylight/solar parameters for specific locations and at high temporal resolution are needed.