In short, Stockholm
- Winter share
- 6% Of the year's solar power comes in December–February: 465 kWh with 20 panels (8.4 kW).
- Daylight in December
- 6.1 h The noon sun stands 7.2° high at the December solstice.
- Snow load
- 2.0 kN/m² Boverket's EKS. The installer sizes the fixings for it.
- Snow loss
- 44–239 kWh/yr Solar power the snow can take in a year, 0.6–3.0% of the year's. Not in the yearly kWh.
- Pollen loss
- 0.5–4.0% Of the solar power in February–May, until the rain washes the pollen off. That is 16–124 kWh a year.
- Cold gain in May
- +2.0% More power from the same sun than on a warm July day. Not in the yearly kWh.
The seasons
In Stockholm winter gives 6% of the year's solar. Spring and autumn bring 53% of the savings.
Summer brings 43% of the solar power and the reference house sells 45% of it at the spot price (0,79 kr on average). In winter the panels cover 6% of the house's power; every kWh you do not buy is worth 1,76 kr.
| Season | Solar | Used at home | Sold | Saved |
|---|---|---|---|---|
| Winter | 465 kWh | 465 kWh | 0 kWh | 994 kr |
| Spring and autumn | 4,087 kWh | 2,686 kWh | 1,401 kWh | 4 981 kr |
| Summer | 3,432 kWh | 1,874 kWh | 1,558 kWh | 3 480 kr |
Where the solar power goes, kWh per season.
Solar 465 kWh · sold 0 kWh · bought 7,165 kWh · saved 994 kr
Solar 4,087 kWh · sold 1,401 kWh · bought 6,772 kWh · saved 4 981 kr
Solar 3,432 kWh · sold 1,558 kWh · bought 1,143 kWh · saved 3 480 kr
Snow
Snow takes part of the year's solar power. A snowy winter takes more.
Snow on the panels blocks the light. In Stockholm the day is 6.1 hours long in December, and winter gives 6% of the year's solar power. The ground snow load is 2.0 kN/m² (Boverket's EKS); the installer sizes the fixings for it. Do not go up on the roof to clear it.
We use the snow days at the SMHI station Stockholm-Observatoriekullen: days with snow on the ground, and how much solar power the roof makes on those days. On most snow days the snow slides off or melts in the sun, so only part of that power is lost. The loss is not in the yearly kWh of the other tables. The roof slopes 27°: steeper roofs clear faster, flat roofs lose more, and a single winter can be much worse.
Snow days per month, SMHI Stockholm-Observatoriekullen 1991–2020
Cold
Cold cells give more power from the same sun.
A solar panel loses 0.29% of its power for each degree the cell is warmer than 25 °C. The panel's datasheet gives this; the 3 datasheets we read (see Table and sources) give 0.24 to 0.29%/°C, and older panels lose more, so the gain is larger. On a sunny day the cell gets warmer than the air. In Stockholm the cell is on average 28.8 °C in May and 35.6 °C in July when the sun shines on the roof. The gain is not in the yearly kWh of the other tables: our calculation only takes off for heat above 25 °C.
Extra power against July heat per month, kWh
Pollen
Pollen puts a thin film on the panels in spring.
Tree pollen (alder, hazel, birch) flies in February–May at the Pollenrapporten station Stockholm. Pollen can take 0.5–4.0% of the solar power in those months. No study of pollen on Swedish roofs exists, so the range is our judgement: the low figure is a European model for the Nordic countries, the high one a measurement with much pollen in Belgium. Rain washes some off, a hose or a soft brush the rest. The loss is not in the yearly kWh of the other tables.
Pollen loss per month, kWh
Winter per city
Winter in all 20 cities, north to south.
| City | Winter share | Day in December | Snow load | Snow loss kWh/yr | Cold gain May % |
|---|---|---|---|---|---|
| Luleå | 3% | 3.1 h | 3.0 kN/m² | 116–632 | +2.7 |
| Umeå | 3% | 4.3 h | 3.0 kN/m² | 110–598 | +2.4 |
| Sundsvall | 5% | 5.0 h | 3.5 kN/m² | 86–468 | +2.2 |
| Gävle | 5% | 5.6 h | 2.5 kN/m² | 69–374 | +2.0 |
| Uppsala | 5% | 5.9 h | 2.0 kN/m² | 46–252 | +1.9 |
| Västerås | 5% | 6.0 h | 2.0 kN/m² | 39–213 | +1.8 |
| Karlstad | 6% | 6.1 h | 2.5 kN/m² | 44–242 | +2.1 |
| Eskilstuna | 6% | 6.1 h | 2.5 kN/m² | 47–256 | +1.8 |
| Stockholm | 6% | 6.1 h | 2.0 kN/m² | 44–239 | +2.0 |
| Örebro | 6% | 6.1 h | 2.5 kN/m² | 44–241 | +1.8 |
| Norrköping | 6% | 6.3 h | 2.0 kN/m² | 41–224 | +1.7 |
| Linköping | 6% | 6.3 h | 2.0 kN/m² | 35–192 | +1.6 |
| Jönköping | 6% | 6.5 h | 2.5 kN/m² | 53–289 | +1.4 |
| Borås | 6% | 6.5 h | 2.0 kN/m² | 36–198 | +1.3 |
| Göteborg | 7% | 6.5 h | 1.5 kN/m² | 26–140 | +1.3 |
| Växjö | 6% | 6.7 h | 2.0 kN/m² | 35–189 | +1.4 |
| Halmstad | 7% | 6.8 h | 1.5 kN/m² | 17–92 | +1.3 |
| Helsingborg | 7% | 6.9 h | 1.0 kN/m² | 19–106 | +1.4 |
| Lund | 7% | 7.0 h | 1.5 kN/m² | 18–99 | +1.5 |
| Malmö | 7% | 7.0 h | 1.0 kN/m² | 19–102 | +1.6 |
Questions and answers
Do I have to clear snow off the panels?
No. In Stockholm the snow can take 44–239 kWh a year, 0.6–3.0% of the year's solar power; it slides off or melts. Do not go up on a snowy roof. The installer sizes the fixings for the snow load, 2.0 kN/m² in Stockholm.
How much do the panels give in December?
Little. In Stockholm, 20 panels facing south give 465 kWh from December to February, 6% of the year. The day is 6.1 hours long in December.
Do the panels work better in the cold?
Yes. A cold panel gives more power from the same sun than a warm one. In Stockholm the same sun gives about 2.0% more power in May than on a warm July day. The gain is not in the yearly kWh of our tables.
The data
The data behind every number on the page, per month or year. Click a dotted number to open it.
Solar and use per month
The house uses the most power in winter, when the sun gives the least. Summer gives a surplus that is sold.
| Month | Solar | Used | Sold | Bought | Saved |
|---|---|---|---|---|---|
| Jan | 119 | 119 | 0 | 2,482 | 268 kr |
| Feb | 265 | 265 | 0 | 1,938 | 602 kr |
| Mar | 662 | 630 | 32 | 1,174 | 875 kr |
| Apr | 987 | 455 | 532 | 348 | 780 kr |
| May | 1,186 | 415 | 771 | 187 | 1 070 kr |
| Jun | 1,239 | 562 | 677 | 241 | 1 291 kr |
| Jul | 1,209 | 656 | 553 | 350 | 1 030 kr |
| Aug | 984 | 656 | 329 | 552 | 1 160 kr |
| Sep | 711 | 645 | 66 | 966 | 1 179 kr |
| Oct | 396 | 396 | 0 | 1,620 | 732 kr |
| Nov | 146 | 146 | 0 | 2,477 | 345 kr |
| Dec | 82 | 82 | 0 | 2,745 | 124 kr |
| Year | 7,986 | 5,027 | 2,960 | 15,080 | 9 456 kr |
Roof facing south, 20,000 kWh a year, someone home in the day. kWh and kr per month.
Source: priv/data/city_solar/<city>.json · scenarios[id=S-20-b0-nocar-20-home].months, seasons
Snow days per month, SMHI Stockholm-Observatoriekullen 1991–2020
Days with at least one centimetre of snow on the ground at Stockholm-Observatoriekullen, and the solar power the snow can take from the reference roof. The loss is not in the yearly kWh of the other tables.
| Month | Snow days | Snow depth, cm | Solar, kWh | Loss, kWh (rough split) |
|---|---|---|---|---|
| Oct | 0.3 | 0.0 | 396 | 0.3–1.5 |
| Nov | 4.6 | 1.3 | 146 | 1.6–8.5 |
| Dec | 11.4 | 4.1 | 82 | 2.1–11.5 |
| Jan | 19.3 | 6.8 | 119 | 5.2–28.2 |
| Feb | 18.9 | 8.7 | 265 | 12.5–68.0 |
| Mar | 13.0 | 5.9 | 662 | 19.4–105.5 |
| Apr | 1.3 | 0.1 | 987 | 3.0–16.3 |
| Winter | 44–239 |
SMHI station Stockholm-Observatoriekullen (1.6 km from the centre), winters 1991-2020, 28 years with data. The loss = the share of the month's solar made on snow days × a factor 0.07–0.38, calibrated on measured snow losses in northern Sweden. The roof slopes 27°; steeper roofs clear faster, flatter roofs lose more. A single winter can be much worse. The factor is calibrated on the whole winter; the split per month is rough.
Source: SMHI open data, snow depth at 07, corrected archive (https://opendata-download-metobs.smhi.se/api/version/1.0/parameter/8/station/98210/period/corrected-archive/data.csv) · van Noord M., Landelius T., Andersson S. (2021). Snow-Induced PV Loss Modeling Using Production-Data Inferred PV System Models. Energies 14(6):1574 (https://doi.org/10.3390/en14061574) · Marion B. et al. (2013). Measured and modeled photovoltaic system energy losses from snow for Colorado and Wisconsin locations. Solar Energy 97:112-121 (https://doi.org/10.1016/j.solener.2013.07.029)
Extra power against July heat per month, kWh
How much more power the panels give from the same sun than on a warm July day, because the cells are colder. The gain is not in the yearly kWh of the other tables.
| Month | Air, °C | Cell, °C | Against STC, % | Against July, % | Extra against July heat, kWh |
|---|---|---|---|---|---|
| Jan | -0.9 | 3.2 | 6.3 | 9.7 | 10.5 |
| Feb | -0.9 | 6.9 | 5.3 | 8.6 | 21.0 |
| Mar | 1.2 | 14.8 | 3.0 | 6.2 | 38.8 |
| Apr | 5.5 | 22.8 | 0.6 | 3.8 | 36.4 |
| May | 10.5 | 28.8 | -1.1 | 2.0 | 23.8 |
| Jun | 15.0 | 34.0 | -2.6 | 0.5 | 5.8 |
| Jul | 17.4 | 35.6 | -3.1 | 0.0 | 0.0 |
| Aug | 16.4 | 32.8 | -2.3 | 0.8 | 8.1 |
| Sep | 12.8 | 26.9 | -0.5 | 2.6 | 18.1 |
| Oct | 7.9 | 17.7 | 2.1 | 5.4 | 20.1 |
| Nov | 3.8 | 9.2 | 4.6 | 7.9 | 10.7 |
| Dec | 0.7 | 3.9 | 6.1 | 9.5 | 7.1 |
Cell temperature with the Faiman model for a panel close to the roof (U0 25, U1 0), weighted by the sunlight on the roof (27°, south). The air is the daily mean temperature (PVGIS), a little below the daytime one; so the cells come out a little too cold in every month. Temperature coefficient -0.29%/°C, the median of 3 datasheets (-0.29 to -0.24%/°C).
Source: Faiman D. (2008). Assessing the outdoor operating temperature of photovoltaic modules. Progress in Photovoltaics 16(4):307-315 (https://doi.org/10.1002/pip.813) · PVGIS 5.3, the daily mean air temperature (T2m) in the city's weather (https://re.jrc.ec.europa.eu/api/v5_3/) · LONGi Hi-MO 6 Explorer LR5-54HTH 420-440M (https://static.longi.com/L_Gi_LE_PM_T_PMD_059_F130_LR_5_54_HTH_420_440_M_30_30_and_15_Frame_Explorer_V19_4448f2931d.pdf) · Jinko Solar Tiger Neo JKM420-440N-54HL4R-B (https://www.jinkosolar.com/uploads/JKM420-440N-54HL4R-B-F1.3-EN.pdf) · REC Alpha Pure-R 400-430W (https://www.recgroup.com/sites/default/files/2025-11/DS_Alpha_Pure-R_IEC_EN_AUCEC_101125_0.pdf)
Pollen loss per month, kWh
The months with tree pollen (alder, hazel, birch) at the Pollenrapporten station Stockholm, and the solar power a layer of pollen can take. The loss is not in the yearly kWh of the other tables.
| Month | Solar, kWh | Loss, kWh |
|---|---|---|
| Feb | 265 | 1.3–10.6 |
| Mar | 662 | 3.3–26.5 |
| Apr | 987 | 4.9–39.5 |
| May | 1,186 | 5.9–47.4 |
| The pollen months | 3,100 | 16–124 |
Station Stockholm (4.1 km away), seasons 2024, 2025, 2026. Loss 0.5–4.0% of the solar power in those months. No study of pollen on Swedish roofs exists: the low figure is a European model for the Nordic countries, the high one a measurement with much pollen in Belgium. Rain washes some off; a hose or a soft brush the rest. Grass pollen in summer is not included.
Source: Pollenrapporten, open API (Swedish Museum of Natural History, Palynological Laboratory) (https://api.pollenrapporten.se/v1/pollen-count?region_id=2a2a2a2a-2a2a-4a2a-aa2a-2a2a2a303a32&start_date=2024-01-01&end_date=2026-11-10) · Fernández Solas Á. et al. (2025). Photovoltaic soiling loss in Europe: Geographical distribution and cleaning recommendations. Renewable Energy 239, 122086 (https://doi.org/10.1016/j.renene.2024.122086) · Appels et al. (2013), Solar Energy 96, as cited by Fernández Solas et al. (2025): 3–4% loss with high pollen in Belgium (https://doi.org/10.1016/j.solener.2013.07.017)