Typical Meteorological Year (TMY)
A typical meteorological year (TMY) is a one-year, hour-by-hour weather data set assembled from the most representative months of a long record. It gives solar simulations a realistic year of irradiance, temperature and wind for a location without the extremes of any single real year.
A typical meteorological year (TMY) is a weather file that represents the long-term climate of one location in a single year of hourly data. It contains, for each of the 8,760 hours, the values a solar simulation needs: global, direct and diffuse irradiance, air temperature, wind speed and usually humidity.
It is stitched together from real months. A TMY is not an average. Averaging would smooth away clouds and sunny spells. Instead, each calendar month is chosen from the long record as the real month closest to that month's long-term statistics. A TMY might use a January from one year, a February from another and so on, which keeps realistic hour-to-hour variation.
It is the standard input for yield estimates. Because it represents typical conditions rather than one particular year, it is the right basis for a forecast that has to hold over the life of the system. Public sources such as the European Commission's PVGIS publish TMY data for most of the world.
It is the base for the uncertainty range. A TMY gives the expected year. How much real years scatter around it comes from the same long record, which is how probabilistic values like P90 are built.
Why it matters for solar installers
The weather data behind a yield estimate is the part of a proposal a customer never sees and a competitor can always question. SolarVis simulates every design hour by hour on PVGIS weather data, either a typical meteorological year or a specific historical year, and uses the TMY as the basis for P50, P90 and P99 analysis. See 3D design and simulation.
Common questions
- How is a typical meteorological year built?
- From a multi-year record of hourly weather, each calendar month is compared with the long-term statistics of that month, and the most typical real month is selected. Twelve such months, possibly from different years, are joined into one year of 8,760 hours.
- Why not simulate with last year's weather?
- A single real year may have been unusually sunny or dull, so a yield based on it could be too high or too low for the decades the system will run. A TMY represents the long-term typical conditions, which is what a 20-year savings forecast needs.
- Is a TMY a forecast?
- No. It describes typical past conditions, not the weather of any future year. Actual annual yield will scatter around the TMY result, and probabilistic analysis such as P90 expresses how far it is likely to scatter.