Irradiance losses:
Snow & soiling
Modelling recommendations
In PV simulation software, monthly snow and soiling loss values are typically needed as input.
- Generalized loss values - snow: For roof systems, generalized snow loss data from NS 3031:2025 can be used as indicative estimates. The loss tables are also found in [1].
- Generalized loss values - soiling: We recommend a monthly soiling loss of 1% [1], in line with NS 3031:2025. For locations or periods with prominent soiling sources combined with limited precipitation, a higher soiling loss should be considered.
- Modelling monthly snow losses: Because snow loss values vary significantly between systems, using generalized values is a simplification. The most accurate snow loss estimates are achieved when modelled for the specific system configuration and location. We recommend using the model developed by Marion et al. [2], together with long time series of historical weather data, to capture the expected variability, as described in [1] (see also the code example in the report attachment). For roof‑mounted systems, we recommend accounting for slower snow clearing during periods with thick snow layers [1,3], whereas for ground‑mounted systems we recommend using the original snow‑clearing coefficient [4]. For snow input-data, SeNorge should be used for locations in Norway (data can also be collected through API), as open satellite‑based snow data sources typically overestimate the number of snowfall events [4].
Background: State of the art & knowledge gaps
Snow losses can vary substantially between locations and from year to year, as they are influenced by both system design and local snow and weather conditions. For roof mounted systems, annual losses in the range of 0-15% and monthly losses in the range of 0-100% have been documented in Norway [5]. For ground mounted systems, snow losses in Norway are less documented, but annual losses up to 10 % have been observed in Norwegian mountain climate [4]. Within FME Solar/SuSolTech, we have validated and adapted the model developed by Marion et al. [2] for roof- and ground-mounted systems, achieving good agreement at daily and monthly time scales [3,4]. Future work in FME Solar WP2 aims to extend this model to improve accuracy and enable loss modelling at higher temporal resolution by more comprehensively accounting for different module technologies and system configurations, as well as a wider range of snow and weather conditions.
Soiling losses (e.g. due to dust or dirt) are typically low in Norway because of frequent precipitation, but can in some cases be higher due to a combination of significant soiling sources (e.g. agricultural activity [6], pollen) and extended periods with little rainfall. Another notable soiling source is bird droppings, which can cause substantial losses and are not easily removed by rain. There is a need for more data on soiling losses and cleaning effects under Nordic conditions to better quantify these impacts.
ReFERENCES
[1] M.B. Øgaard, C. Seiffert, Innstrålingstap i solcelleanlegg: Snø og støv, IFE Report, 2024.
[2] B. Marion, R. Schaefer, H. Caine, G. Sanchez, Measured and modeled photovoltaic system energy losses from snow for Colorado and Wisconsin locations, Solar Energy, 2013.
[3] M.B. Øgaard, I. Frimannslund, H. N. Riise and J. Selj, Snow loss modeling for roof mounted photovoltaic systems: Improving the Marion snow loss model, IEEE Journal of Photovoltaics, 2022.
[4] M.B. Øgaard, E.W. Eriksen, D. Dahlioui, S. Rønneberg, T. Müller, H.N. Riise, Analyzing and modeling snow loss and snow loss accumulation in ground-mounted photovoltaic systems, EPJ Photovoltaics, 2026.
[5] M.B. Øgaard, H.N. Riise, J. Selj, Estimation of snow loss for photovoltaic plants in Norway, EUPVSEC proceedings, 2021.
[6] H.N. Riise, M.B. Øgaard, T.U. Nærland, Soiling and snow impact on a PV plant at a farm in Norway, EUPVSEC proceedings, 2021.
Last update: 27.4.2026