International audience ; Local climate and environmental conditions can impact the performance of concentrator photovoltaic (CPV) systems. There is a lack of experimental performance analysis of CPV systems, especially in the region with high snowfall and very low temperature in winters. In this paper, we present first a CPV system performance in humid continental climate and identify snow and frost as sources of losses that are not considered in conventional predictive models. We propose then a method to account for the negative effect of snow and frost on the system, by adding monthly soiling factors in the predictive model. The monthly soiling factors are modeled based on average monthly snow fall and ambient temperature. Applying this method, decrease in Root Mean Square Error (RMSE) between predicted and actual energy production from 24.51 to 5.07 % validates our model in humid continental climate for CPV systems.