Solar photovoltaic (PV) systems are frequently installed in climates with significant snowfall. To better understand the effects of snowfall on the performance of PV systems, a multi-angle, multi-technology PV system was commissioned and monitored over two winters. A novel methodology was introduced and validated with this system, which allows for the determination of snowfall losses from time-series performance data with correlated meteorological observations down to a 5-min resolution. In addition, a new method for determining the probability distribution of snow deposition on a module from image data was developed. It was found that the losses due to snowfall are dependent on the angle and technology being considered and the effects of increased albedo in the surroundings of a PV system can increase expected yields, particularly in the case of high tilt angle systems. Existing methods for predicting losses due to snowfall were investigated, and were found to provide overly conservative estimates of snow losses. Overall the results show that the proper assessment of snow related losses can help improve system performance and maintenance. It is concluded that proper characterization of the snowfall effect on PV system performance can influence better systems optimization for climates experiencing snowfall. (C) 2013 Elsevier Ltd. All rights reserved.