Study improves estimation of snow losses in solar power in Nordic conditions

As solar power becomes more widespread in snowy regions, it is essential to accurately model how snow accumulation on photovoltaic panels reduces electricity generation, i.e. causes snow losses. A recent study compared several models used to estimate snow losses on photovoltaic systems in Nordic environments.

The results show that the widely used Marion model with daily snow depth input tends to overestimate snow losses. In contrast, models incorporating hourly snow data significantly improve accuracy. In the best cases, model errors were reduced by approximately 30% compared to the original model.

Wind effects need to be considered in models

The study also demonstrates that solar power production can be modelled reliably using open meteorological datasets, such as ERA5 reanalysis data, without site-specific irradiance or temperature measurements. This enables solar energy assessments in data-scarce regions.

In addition, the results indicate that wind can prevent snow accumulation on photovoltaic panels. This may lead to systematic errors in models based solely on snowfall or snow depth data, particularly in situations where snow depth does not change significantly during snowfall events. The findings highlight the importance of including wind effects in future model development.

The analysis is based on a four-year solar power production dataset from the Finnish Meteorological Institute’s monitoring station in Sodankylä. The study was conducted in collaboration between the Finnish Meteorological Institute and LUT University.

The results support the expansion of solar energy in northern regions by improving the reliability of production estimates.

Further information:

Researcher Juha Karhu, Finnish Meteorological Institute, tel. +358 50 359 2183, JuhaA.Karhu@fmi.fi

Researcher Nashmin Hosseinpour, Aalto University, tel. +358 50 305 5512, Nashmin.Hosseinpour@aalto.fi

The research was published on 24 March 2026 in the IET Renewable Power Generation journal and is openly accessible.

Reference: Hosseinpour, N., Karhu, J. A., Lindfors, A. V. & Lassila, J. (2026). Comparative study of snow loss models for PV systems in Nordic conditions: Townsend and Marion-based approaches. IET Renewable Power Generation, 20:e70235. https://doi.org/10.1049/rpg2.70235