Photovoltaics and snow: An update from two winters of measurements in the SIERRA

With snowy locations becoming common for large photovoltaic (PV) installations, analytical models are now needed to estimate the impact of snow on energy production. A generalized monthly snow loss model is introduced here. The model was calibrated using ongoing measurements that began in December 2009 for three panel orientations at a BEW test station in Truckee, CA. Supplemental data for a fourth orientation from a nearby municipal PV system were also used to calibrate the model. Overall, the energy prediction error for the four panel orientations has been just 2% RMS on an annual rolling-average basis. Short-term errors are higher, a consequence of the variable nature of snowfall timing, quantity, and quality, and its complex interaction with temperature, wind, humidity, and ground interference. Despite these limitations, good quality, unbiased monthly loss estimates can now be used as inputs to the simulation programs PV investors rely on for decision-making. Loss profiles for three sample cities with prominent PV markets are provided. The samples have not been validated, but are included to show what the model predicts for other climates.