Fast variability of solar radiation is the main cause of fluctuating photovoltaic power production. Shadows caused by moving clouds are the main reason of such variability. Irradiance transitions caused by edges of cloud shadows can be very steep and large and might lead to failures in maximum power point tracking causing extra losses. Further, fast fluctuations of the power fed to the electric grid can cause power balance and quality problems for the grid. This paper presents a method to identify shading periods caused by moving clouds in measured irradiance data. A total of around 12,000 shading periods were identified in a measured data of 15 months around midsummer in 2011-2014 and their shading strength, duration, time of occurrence etc. were analysed. It was found that the duration of shading periods varies a lot from about four seconds up to almost 1.5 h with an average duration of around 60 s. Furthermore, the Linear Cloud Edge method was used to determine the velocity of shadows and their speed, direction of movement, length etc. were analysed. The determination of velocity was conducted by two different ways based directly on measured irradiance values and on the curve fits of a mathematical model of irradiance transitions. The use of curve fits mitigates the effects of irregularities present on shadow edges. The speed of shadows varies greatly with an average value of around 13 m/s. (C) 2016 Elsevier Ltd. All rights reserved.