Behind every solar PV proposal, payback calculation and promise made to a customer sits a single number: how much energy the system will produce in a year. One of the most widely used free tools for that estimate is PVGIS. In this article we explain what PVGIS is, what inputs it needs, what its outputs mean, how to use it step by step and where its limits are.
What is PVGIS?
PVGIS (Photovoltaic Geographical Information System) is a free web tool developed by the European Commission's Joint Research Centre (JRC). It calculates solar irradiation at a chosen point and the expected output of a photovoltaic system installed there.
The irradiation data behind PVGIS is largely derived from satellite imagery. Where satellite coverage is not available, climate reanalysis databases are used. The tool works from multi-year data to produce monthly and annual averages, so the result is a long-term average, not the measurement of a single year.
PVGIS includes several calculators: grid-connected PV output, tracking systems, off-grid systems, monthly irradiation data, daily profiles and typical meteorological year (TMY) data, among others. For most PV projects, the grid-connected calculator is the one you will use.
What inputs does PVGIS need?
Location
Pick a point on the map or enter latitude and longitude. The location is used for the irradiation data and also to calculate the horizon profile, meaning distant obstacles such as mountains and hills.
Radiation database
Depending on the location, one or more databases may be available. In most cases the default suggested by the tool is fine. Different databases can give results a few percent apart for the same point; that spread is part of the estimate's natural uncertainty.
PV technology
Options include crystalline silicon, CIS and CdTe. The large majority of today's projects use crystalline silicon modules.
Installed peak power (kWp)
This is the total peak power of the array: number of modules × module power. For example, 18 modules of 550 Wp give 9.9 kWp. Enter the modules' total DC power here, not the inverter rating.
System losses (%)
This covers cable losses, inverter losses, soiling, module mismatch and similar losses. The PVGIS default is 14%. Because the tool calculates some losses itself, such as temperature and angular reflection, this field should contain only system-related losses. Counting losses twice will understate production.
Mounting type, slope and azimuth
- Mounting type: choose "free-standing" (ventilated at the back, as on ground mounts or racked rooftop systems) or "building integrated" (no rear ventilation). Modules without rear ventilation run hotter, which lowers output.
- Slope: the module's angle from horizontal. For flush roof mounting, enter the roof pitch.
- Azimuth: in PVGIS, 0° is south, −90° is east and +90° is west. Do not confuse this with tools that use compass north as 0°; it is a common mistake.
- Optimisation options: for ground-mounted systems, the tool can also find the slope and azimuth that maximise annual yield.
What do the outputs mean?
| Output | Meaning |
|---|---|
| E_m | Average electricity production per month (kWh/month) |
| E_d | Average daily production for that month (kWh/day) |
| H(i)_m | Average monthly irradiation on the module plane (kWh/m²/month) |
| SD_m | Year-to-year standard deviation of monthly production, showing weather-driven variability |
| E_y | Total annual production (kWh/year) |
| Loss breakdown | Angle of incidence, spectral effects, temperature and low irradiance, and total loss |
Divide E_y by the installed power and you get specific yield (kWh/kWp), which is handy for comparing sites and designs. H(i)_m is the energy reaching the module plane; how much of it becomes electricity depends on system efficiency and losses.
SD_m matters when you talk to the customer: the PVGIS figure is an average, and some years will come in below it and some above. A year that falls slightly short of the estimate does not necessarily mean the design was wrong.
Step by step
- Open the PVGIS web tool and select the grid-connected PV calculator.
- Mark the site on the map or enter its coordinates.
- Make sure the calculated horizon option is on; this matters in hilly areas.
- Choose the radiation database and PV technology.
- Enter installed peak power in kWp.
- Enter system losses. If unsure, start with the default and note your assumption.
- Enter mounting type, slope and azimuth. For roofs with several orientations, calculate each face separately and add the results.
- Run the calculation and review the monthly table, annual production and loss breakdown. You can download the results as CSV or PDF.
Limits of PVGIS
- It does not know about near shading. The horizon calculation accounts for distant terrain such as mountains and hills. Nearby obstacles like chimneys, neighbouring buildings, trees or inter-row shading are not included; they require a separate shade analysis.
- The loss figure is your assumption. Whatever you enter in the system loss field shapes the result. Long cable runs, dusty sites or mismatched strings may need more than the default.
- It gives averages. Actual output in any single year can deviate from the average depending on the weather.
- It does not model inverter limits or string design. Clipping and operation outside the MPPT voltage window must be checked separately. See our article on stringing and MPPT.
- It may not fully capture local microclimate. Fog, snow cover or heavy dust may not show up fully in satellite data.
Combining PVGIS data with consumption
A production estimate alone does not answer the customer's real question, which is how much of their consumption the system will cover. To answer it, put the monthly E_m values next to the monthly consumption taken from the customer's bills.
The table below consists entirely of example values; it is not an actual PVGIS output. It assumes a 10 kWp system and a home whose consumption rises in summer because of air conditioning.
| Month | Production (kWh) | Consumption (kWh) | Difference (kWh) |
|---|---|---|---|
| January | 750 | 1,300 | −550 |
| February | 900 | 1,200 | −300 |
| March | 1,250 | 1,150 | 100 |
| April | 1,450 | 1,050 | 400 |
| May | 1,650 | 1,100 | 550 |
| June | 1,750 | 1,300 | 450 |
| July | 1,850 | 1,500 | 350 |
| August | 1,750 | 1,500 | 250 |
| September | 1,450 | 1,200 | 250 |
| October | 1,150 | 1,050 | 100 |
| November | 850 | 1,100 | −250 |
| December | 700 | 1,300 | −600 |
| Total | 15,500 | 14,750 | 750 |
On an annual basis, production appears to exceed consumption by 750 kWh. Month by month the picture changes: March to October shows a combined surplus of 2,450 kWh, while November to February shows a combined shortfall of 1,700 kWh. And even within a month, production happens during the day while part of the consumption happens in the evening. How surplus and shortfall are treated depends on current regulations and the distribution company's terms, so sizing a system on annual totals alone can be misleading. For the financial side, see our article on solar payback.
Common mistakes
- Entering azimuth as a compass bearing (north = 0°)
- Entering inverter power instead of installed module power
- Adding temperature loss to system losses and counting it twice
- Combining surfaces with different orientations in one calculation
- Ignoring near shading altogether
- Presenting average production as if it were guaranteed
Energy yield estimation with PVAGE
When you run a system analysis in PVAGE (see energy yield estimation), you select a province/district or enter coordinates and get monthly irradiation and production data straight from PVGIS, with no need to run the numbers on a separate site and copy them across. Once you choose a load profile (home, office, shop, industry, farm, irrigation) and monthly distribution (flat, summer- or winter-weighted, or all 12 months entered manually), you get a month-by-month production, consumption and difference table showing surplus and shortfall.
You can set the target capacity automatically or manually and see the MPPT limit check in the module-inverter matching. The effect of nearby obstacles can be assessed separately through per-panel shading loss in the roof design. Results can be saved as scenarios and presented to the customer as a PDF report.