How much power does a solar panel produce?

A solar panel's label says something like "100 W" or "400 W". That number is real, but it was measured in a laboratory under bright, cool, perfect light. What you actually want to know is how much energy the panel gives you in a day, in your place, in the months you need it. This guide walks from the label to that answer, with real figures for Denver in the US and London in the UK.

The short version

• The watt rating is measured at Standard Test Conditions: sunlight of 1,000 W/m² and cells at 25 °C (77 °F). Real days are almost always weaker.
• Energy per day ≈ panel watts × peak sun hours × what is left after losses. That is a rough rule, not a guarantee.
• In PVGIS, a 100 W panel facing south at a 30° tilt gives about 446 Wh a day over the year in Denver and about 277 Wh in London.
• Winter is the hard part: in December the same panel gives about 331 Wh a day in Denver but only about 109 Wh in London.
• Heat lowers output, by roughly 0.3 % to 0.5 % for every degree Celsius the cells run above 25 °C.
• Averages hide bad runs. In the London record, about one December day in five gave less than 30 Wh from that 100 W panel.
• That is why our quiz simulates your system day by day over many years of real weather instead of using one average.

What the watt rating on a panel means

The rating is the panel's peak power: what it delivers at Standard Test Conditions, or STC. The European Commission's Joint Research Centre, which runs the PVGIS solar calculator, defines these as sunlight of 1,000 watts per square metre on the panel, a module temperature of 25 °C (77 °F) and a standard light spectrum (PVGIS FAQ and user manual). A square metre is about 10.8 square feet.

Think of the rating as a car's top speed: useful for comparing panels, but it does not tell you how far you will travel in a day.

From watts to watt-hours

A watt is how fast energy flows right now. A watt-hour (Wh) is an amount of energy: one watt for one hour. In the water comparison from our glossary, watts are how fast the tap runs and watt-hours are the water in the bucket. A panel's watts change with every cloud; its watt-hours over a day are what you size a system with. They depend on how much sunlight lands on the panel and how much of it the system turns into usable electricity.

Irradiance and peak sun hours

Irradiance is the strength of the sunlight hitting the panel at one moment, in watts per square metre (W/m²). It rises from zero at dawn, peaks around midday and drops under clouds.

Adding up a whole day of irradiance gives the day's irradiation, in kilowatt-hours per square metre (kWh/m²). Because the lab rating uses 1,000 W/m², which is 1 kW/m², you can read that daily total as peak sun hours: the number of hours of full lab-strength sunshine that would deliver the same energy. A day with 3.4 kWh/m² on the panel is 3.4 peak sun hours, even if the sun was up for 15 hours.

PVGIS gives these figures for a surface facing south at a 30° tilt (PVGIS PVcalc, checked 2026-10-20):

• Denver, Colorado: about 5.7 peak sun hours a day averaged over the year, 7.0 in June and 3.9 in December (satellite data for 2005 to 2015).
• London: about 3.4 a day over the year, 5.3 in June and 1.3 in December (satellite data for 2005 to 2023).

London's December sunlight is about a quarter of its June sunlight (our calculation: 1.29 ÷ 5.28); Denver's is a little over half (3.89 ÷ 7.00).

The rough rule, and why it is only rough

A quick estimate many people use:

Energy per day (Wh) ≈ panel watts × peak sun hours × (1 − losses)

For a 100 W panel in London in December, with about 19 % of losses (the total PVGIS reports for this case, explained below): 100 × 1.29 × 0.81 ≈ 104 Wh a day (our calculation). PVGIS's own monthly figure for this case is 109 Wh a day, close but not the same, because PVGIS calculates hour by hour and the losses change through the day and the year.

The rule is fine for a first feel. It is not fine for sizing a system you depend on, for reasons the rest of this guide explains.

What reduces the output

Heat

Solar cells lose power as they get hotter. For crystalline silicon panels, the most common kind, power falls by about 0.3 % to 0.5 % for every 1 °C above 25 °C (Sandia National Laboratories; PVEducation gives 0.4 % to 0.5 % for silicon). This figure is the panel's temperature coefficient, printed on its datasheet.

Panels in sunshine run much hotter than the air. PVGIS estimates the panel temperature from air temperature, sunlight and wind, using a model by Faiman, and it allows for panels mounted with air behind them running cooler than panels built into a roof (PVGIS methods page).

For our example, PVGIS reports that heat and weak light together cost about 6.6 % of a year's output in sunny, hot-summer Denver and about 4.8 % in London (PVGIS PVcalc, checked 2026-10-20). Cool, bright days suit panels, because cooler cells lose less.

Weak light and reflection

Panel efficiency stays nearly constant from about 400 W/m² to 1,000 W/m² but drops in weaker light (PVGIS methods page), so a grey winter day is a double loss. Light hitting the glass at a shallow angle also partly bounces off: about 2.6 % of the year in Denver and 3.2 % in London for our example (PVGIS).

System losses

The rest of the system takes its share too. PVGIS suggests a default of 14 % for these "system losses" in its grid-connected calculator, covering cables, inverters, dirt or snow and module ageing (PVGIS user manual). NREL's PVWatts calculator also uses 14 % by default and lists the categories one by one (PVWatts Version 5 Manual, NREL).

PVGIS multiplies the losses rather than adding them (PVGIS FAQ). In total it reports about 21.8 % of losses for our Denver example and 19.3 % for London, where the cooler climate and a small gain from the light's colour spectrum help (PVGIS PVcalc, checked 2026-10-20).

Off-grid systems have extra steps: the charge controller, the battery (some energy is lost charging and discharging it) and, for household appliances, the inverter. Our engine counts each of these separately, as described on the methodology page.

Tilt and direction

In the Northern Hemisphere panels collect the most facing roughly south. A steep tilt catches the low winter sun better; a flatter one does well in summer.

Here is the same 1 kW of panels in London, per day, from PVGIS (PVcalc, checked 2026-10-20; divide by ten for a 100 W panel):

• Lying flat: about 0.46 kWh a day in December, 4.19 kWh in June, 845 kWh over the year.
• 30° tilt, facing south: about 1.09 kWh in December, 4.16 kWh in June, 1,010 kWh over the year.
• 60° tilt, facing south: about 1.42 kWh in December, 3.44 kWh in June, 977 kWh over the year.
• 30° tilt, facing west: about 0.49 kWh in December, 3.88 kWh in June, 800 kWh over the year.

So in London a steep, south-facing panel gives roughly three times the December energy of a flat one (our calculation: 1.42 ÷ 0.46), while losing about a sixth of its June output. If you need power mainly in winter, tilt matters more than almost anything else you can change for free. Our quiz offers tilt choices for this reason.

Worked example: one panel in Denver and in London

Take one 100 W panel, crystalline silicon, mounted on a frame with air behind it, tilted at 30° and facing south. Figures are from PVGIS with its default 14 % system loss; the per-100 W numbers are our calculation (PVGIS's 1 kW results divided by ten).

Denver, Colorado (PVGIS-NSRDB satellite data, 2005 to 2015):

• Average over the year: about 446 Wh a day, about 163 kWh a year.
• June: about 520 Wh a day.
• December: about 331 Wh a day.
• The yearly total varies from year to year with a standard deviation of about 3.4 kWh for this panel, around 2 %.

London (PVGIS-SARAH3 satellite data, 2005 to 2023):

• Average over the year: about 277 Wh a day, about 101 kWh a year.
• June: about 416 Wh a day.
• December: about 109 Wh a day.
• Year-to-year standard deviation: about 4.1 kWh, around 4 %.

What does 109 Wh mean? It would run a 10 W LED light for about 11 hours, and not much more (our arithmetic: 109 ÷ 10). If your cabin needs 500 Wh a day in December, in London you would need roughly 460 W of panels by this average alone, before battery and controller losses (our calculation: 500 ÷ 109 × 100). In Denver the same need takes roughly 150 W.

Why winter is different

In winter the days are shorter, the sun is lower, so its light crosses more air and hits the panel at a shallower angle, and in many places there is more cloud. Denver is about 12 degrees of latitude further south than London, so its winter sun stays higher.

Snow can also cover panels completely for days. PVGIS's 14 % includes an allowance for dirt or snow, but in snowy places a winter day under snow can produce close to nothing. Our engine adds a separate snow estimate for each location from snow climate records (see the methodology page).

For off-grid systems in the UK, the honest answer is often that covering December with panels alone takes a very large system. A hybrid, with a generator or grid charging for the darkest weeks, can make more sense. Our quiz says so when it finds that.

Why an average is not enough

A monthly average hides runs of dull days. A battery can carry you through one grey day, but a week of them drains it.

We downloaded PVGIS's hour-by-hour output for our London example and added it up day by day (our calculation from the PVGIS hourly series, checked 2026-10-20). Over the 589 December days from 2005 to 2023:

• The average December day gave about 105 Wh from the 100 W panel.
• The best December day gave about 282 Wh; the worst gave almost nothing.
• About 22 % of December days, roughly one in five, gave less than 30 Wh.
• Whole Decembers differed a lot: the dullest (2023) gave about 1.9 kWh for the month from the 100 W panel, the brightest (2014) about 4.4 kWh, more than twice as much.

In Denver, by the same calculation over 342 December days from 2005 to 2015, only about 1.5 % of December days fell below 30 Wh.

A system sized for London's December average would lean on its battery for many days. Whether the battery lasts depends on the order of good and bad days, which an average cannot show.

How our quiz uses real weather instead of an average

Our engine does not multiply your panels by one average. It works like this:

1. It takes the PVGIS hourly sunlight, air temperature and wind for the roughly 5 km (3 mile) grid square around your location, for every year in the record (in the UK 2005 to 2023; in most of the US 2005 to 2015).
2. For each day it calculates what your panels would give, with the panel temperature and your tilt and direction.
3. It runs your system day by day through all those years: the sun charges the battery, your appliances drain it, and it records every day the battery ran out. It also stops a lithium iron phosphate battery from charging on days below its minimum charging temperature.
4. For each battery size it finds the smallest panel power that gets through a target share of the days in your season without a shortfall: 90 % for Minimum, 97 % for Medium and 99.5 % for With margin. It then picks the battery size beyond which a bigger battery barely reduces the panels needed.
5. It checks that the parts fit together, including that the panels' voltage on the coldest morning stays under the controller's limit.

Each result tells you how many days in a typical year you might still run short, and which month drives the size. The methodology page gives the details and sources, and our engine's yearly output was checked against PVGIS's own calculator on 22 test points in the US and the UK.

What to check before you buy

1. Write down your daily energy need in watt-hours for the months you will use the system.
2. Use your location's figures for those months, not the yearly average, at the tilt and direction you can actually achieve.
3. Allow for losses: about 14 % on the panel side, plus controller, battery and inverter losses off-grid.
4. Plan for runs of bad days with a battery, a backup source, or both.
5. In a cold climate, check the panels' cold-weather voltage against the controller.

What can go wrong

• Sized for summer. A system that works in June can fall far short in December, especially in the UK.
• Sized from the label. Expecting 100 W × 8 hours of daylight = 800 Wh from a 100 W panel. Even Denver's June gives about 520 Wh in PVGIS.
• Flat mounting for winter use. In London a flat panel gave less than half the December energy of a 30° one.
• Shade. Even shade on part of a panel can cut its output sharply; a tree or building that clears the panel in summer can shade it when the winter sun is low.
• Snow and dirt. A snowed-over panel gives close to nothing until it clears.

Questions people ask

Will a 400 W panel give me 400 W? Only in bright, cool, lab-like conditions. On most days the output is lower, and the daily energy is what counts.

Do panels work on cloudy days? Yes, but much less. The London December figures above show how little a dull winter day can give.

Is PVGIS reliable? It is a free tool from the European Commission's Joint Research Centre based on satellite data. Like any model it gives estimates; our engine's yearly results stayed within about 1.3 % of PVGIS's own calculator on our 22 test points.

Next step

Want these numbers for your own place, season and appliances? Size my system: tell us where and how you will use solar, and get a checked parts list where every number is explained. Terms you are unsure of are in our glossary, and the battery side is covered in how big a solar battery you need.

Our results and guides are estimates and do not replace advice from a qualified electrician.