Key points are not available for this paper at this time.
Photovoltaic (PV) installations on rooftops in urban environments have a significant potential to reduce human environmental impact. However, the quality of remote sensing data available for solar potential estimates in different regions varies and regional authorities or policy makers need to know if the available data are suitable for solar potential estimates and/or whether to invest into expensive data collection. Published studies often mention the importance of identification of small disturbing structures on the roof (e.g. chimneys), which is however questionable due to the fact that estimated energy yields are much more dependent on the variability of annual solar irradiation. We used two different models: (i) Photovoltaic Geographical Information System based on coarse data and (ii) ArcGIS Area Solar Radiation tool based on digital surface models at very high resolutions (≤1 m) acquired from UAV photogrammetry. We compared solar irradiation estimates with ground-truth data from a PV system installed on the roof. We show that the effect of adopted model and resolution on estimated irradiation is negligible in comparison with the year-to-year variation of meteorological conditions. We suggest that accurate predictions can be made with relatively coarse building data (e.g. simple roof shapes that do not include dormers and chimneys).
Moudrý et al. (Tue,) studied this question.