TMY — Typical Meteorological Year
A Typical Meteorological Year (TMY) is a dataset of hourly meteorological data representing a "typical" or average year at a specific location. It's constructed by selecting the most statistically representative month from a multi-year historical dataset, then joining these months into a single year-long dataset of 8,760 hourly values.
TMY data is the standard input format for energy simulation tools like PVsyst, EnergyPlus, and SAM (System Advisor Model). Using TMY data allows simulations to represent typical long-term conditions rather than any specific historical year.
How TMY is Created
The most common TMY methodology (Sandia/NREL TMY3) evaluates each month of the historical record using a weighted combination of meteorological parameters including:
- Global and direct solar irradiance (highest weight)
- Dry-bulb temperature
- Dew point temperature
- Wind speed
The month that best matches the long-term statistical distribution for all parameters is selected. The 12 selected months (not necessarily from the same year) are spliced together with smoothing at the boundaries.
TMY vs. Long-Term Averages
SolarScope uses long-term monthly averages (20+ years of NASA POWER data) rather than TMY datasets directly. For preliminary analysis, monthly averages provide equivalent accuracy to TMY for estimating annual production. For detailed hourly modeling, TMY datasets provide more realistic hour-by-hour variation.
NREL's NSRDB provides TMY3 data files for the US that can be used directly in PVsyst or other detailed modeling tools for sites advancing beyond pre-feasibility.
P50 and P90 Production
TMY represents approximately P50 production — the production level exceeded 50% of the time (median). For project financing, lenders often require P90 production estimates — the level exceeded 90% of the time — which accounts for year-to-year weather variability. P90 is typically 5–10% lower than P50 for solar projects.