GridPassport's city pages combine two different datasets: a weather-based estimate for a named future date and amulti-year monthly solar profile. Both use the same reference PV system so locations can be compared. Neither is a measurement or a production guarantee for a specific roof.
Method version 1.0 · Data snapshot updated August 30, 2026
Reference system
The reference profile is a comparison tool. The interactive forecast lets a reader change location, size, direction, tilt, shade and losses.
Tomorrow
The public forecast uses hourly weather inputs from Open-Meteo. GridPassport projects available sunlight onto the configured array, applies the stated system losses and returns energy in kWh. The strongest window is the local four-hour period with the highest expected production.
The city snapshot is prepared by the GridPassport API and read at build time. Opening a city page does not start 253 separate live forecasts.
Uncertainty
The p10-p90 interval is labelled “likely range”. It reflects model and weather uncertainty; it is not a promise that measured production will remain inside the interval.
Typical year
Monthly values come from the European Commission Joint Research Centre PVGIS 5.3 fixed-system calculation. They use a multi-year solar radiation and climate series to estimate a typical month and typical year at each location.
Monthly values must not be read as a forecast for a particular calendar year. The variability column describes the spread in the underlying annual series.
Publication rules
The local calendar date for which production is estimated. Because cities have different time zones, the date may not switch everywhere at the same moment.
When the displayed forecast snapshot was generated. This timestamp, not the forecast date, is used for page metadata and sitemap lastmod.
The latest scheduled snapshot is available. The backend normally refreshes the complete country-page dataset every three to six hours.
A refresh did not complete, so the last successful value is retained and labelled stale. The page never silently presents an old date as tomorrow.
Interpretation
Roof direction and tilt, local shade, snow, dirt, inverter clipping, panel temperature, equipment efficiency, outages and weather at the exact property can all move real output away from the reference estimate.
Country summaries
Country pages average the named city reference points. They compare local conditions for the same 10 kWp setup; they do not estimate the output of every roof or the electricity generated by the country.
Data reuse
GridPassport publishes its static Solar Forecast CSV and JSON datasets under theCreative Commons Attribution 4.0 International licence. You may copy, redistribute and adapt them, including commercially, provided that you give attribution and identify modifications.
This licence does not cover GridPassport software, API access, model implementation, trademarks, website text or third-party images unless explicitly stated.
Open-Meteo weather data is provided underCC BY 4.0. Reuse requires credit to Open-Meteo, a link to the licence and an indication that GridPassport transformed the input.
PVGIS information is supplied by the European Commission Joint Research Centre. Itspublished usage conditionsstate that it is free and has no restrictions on use. GridPassport nevertheless identifies JRC PVGIS in the required source notice so that the origin of the typical profiles remains clear.
Attribution must identify GridPassport, Open-Meteo and European Commission JRC PVGIS. A compact form is: “GridPassport Solar Dataset (CC BY 4.0), using Open-Meteo weather data (CC BY 4.0) transformed by GridPassport and typical PV profiles from European Commission JRC PVGIS.” Link to the dataset and licence where practical, and state if you modified the data.
No separate permission is required for uses allowed by CC BY 4.0. For supported API or production integration, contact hello@gridpassport.com.
Sources and access
Each city and country page offers static CSV and JSON files with the values, timestamps, configuration and source metadata used on that page.