Solar TMY Data 2026: Design Guide

A 5 MW project in Rajasthan was modeled with two different TMY files. The first came from a free global database with 0.5° spatial resolution. The second came from a satellite-derived source with 250 m resolution and local ground-station calibration. The difference in predicted annual yield was 6.7%. At a PPA price of $0.04 per kWh, that gap translated to more than $500,000 in cumulative revenue over the project life.

Solar TMY data is a Typical Meteorological Year file: a synthetic year of hourly weather values built from 10–30 years of historical records. It is the standard input for solar energy yield simulations and the largest single source of uncertainty in most production estimates.

The TMY file you choose shapes every downstream design decision. It affects module count, inverter sizing, tilt optimization, financial projections, and lender confidence. Yet many design workflows treat weather data as a default setting rather than an active choice. This guide explains how TMY data is constructed, what separates a screening-grade file from a bankable one, and how to integrate TMY data into a professional weather data integration workflow in 2026.

In this guide:

• How TMY files are constructed from decades of weather data
• The key parameters inside a solar TMY file and why each matters
• A source-by-source comparison of free and commercial TMY datasets in 2026
• When to use TMY, full time series, or long-term averages
• How P50 and P90 TMY datasets fit into project finance
• Accuracy limits, common mistakes, and microclimate traps
• A step-by-step TMY workflow from site location to yield estimate
• How SurgePV connects TMY data to design, simulation, and proposals

How Solar TMY Data Is Built

A TMY file is not an average year. It is a collage. Engineers select the most representative January, February, March, and so on, from a multi-year historical record. The result is a single synthetic year that preserves realistic day-to-day and hour-to-hour variability while matching long-term climatology.

The Sandia Method Behind TMY2 and TMY3

The classic TMY construction method was formalized by Wilcox and Marion at NREL in 2008. It underpins the TMY2 and TMY3 datasets widely used in North America. The process works as follows:

1. Collect 10–30 years of hourly data for solar irradiance, temperature, dew point, and wind speed.
2. For each calendar month across all years, compute daily means, maximums, and minimums.
3. Build cumulative distribution functions (CDFs) for each parameter using all years combined.
4. For each candidate year and month, build a separate CDF and compare it to the long-term CDF.
5. Calculate the Finkelstein-Schafer statistic, which measures how closely the candidate month’s distribution matches the long-term distribution.
6. Rank candidate months by weighted sum of these statistics and select the most typical month.
7. Concatenate the 12 selected months, smoothing the transitions at month boundaries.

The weighting scheme matters. Older TMY methods were designed for building energy simulations, so they gave equal weight to temperature, humidity, and solar radiation. For PV design, solar radiation should dominate the selection. Modern providers such as Solargis and Solcast use PV-specific weighting schemes that prioritize GHI, DNI, and DHI.

ISO 15927-4 and Regional Variants

Outside North America, the ISO 15927-4 standard governs TMY construction. It uses air temperature, relative humidity, and solar radiation as the core variables. The European Commission’s PVGIS tool and many national meteorological services follow methods aligned with this standard.

The differences between methods are not academic. A TMY built with building-energy weighting may select a month that is typical in temperature but slightly atypical in irradiance. For a PV project, that small irradiance deviation can shift annual yield by 2–4%.

Why 8,760 Hours?

Most TMY files contain 8,760 hourly values, one for each hour of a non-leap year. Some providers now offer sub-hourly TMY at 15- or 30-minute resolution. Sub-hourly data matters for modeling inverter clipping, battery dispatch, and rapid cloud transients. For standard rooftop and commercial design, hourly data remains the norm.

What a Solar TMY File Contains

Not all TMY files contain the same parameters. Understanding which fields your simulation tool actually uses helps you choose the right source and avoid missing inputs.

GHI, DNI, and DHI are the core trio. Every credible PV simulator needs them. The relationship GHI = DNI × cos(zenith) + DHI holds for each timestep. If a TMY file provides GHI and DNI but not DHI, the simulator must estimate DHI using a decomposition model, adding another layer of uncertainty. See the SurgePV glossary entries for GHI, DNI, and DHI for deeper definitions.

Temperature and wind speed drive module temperature. A typical silicon module loses 0.3–0.5% of output for every degree Celsius above 25°C. In hot climates, an accurate ambient temperature record is as important as an accurate irradiance record.

Albedo is often missing from basic TMY files. For bifacial projects, this is a serious gap. Ground reflectivity can range from 0.15 for dark soil to 0.65 for fresh snow. Bifacial gain varies almost linearly with albedo in many designs.

Free vs. Commercial TMY Sources in 2026

The right TMY source depends on project scale, geography, budget, and bankability requirements. Here is how the major providers compare.

NREL NSRDB

The National Solar Radiation Database is the default free source for the United States and nearby territories. The modern Physical Solar Model (PSM) uses satellite imagery and atmospheric physics to produce gridded irradiance at 4 km spatial resolution and half-hourly temporal resolution. NSRDB TMY files are available through the NREL PVWatts API and can be downloaded directly for any location.

A 2018 American Meteorological Society study found NSRDB (1998–2015) GHI biases within ±5% and DNI biases within ±10% when compared to high-quality ground stations. That makes NSRDB suitable for most US residential, commercial, and early-stage utility work.

PVGIS

The European Commission’s Photovoltaic Geographical Information System is the free standard for Europe, Africa, and parts of Asia. PVGIS 5 uses the SARAH-3 satellite-derived dataset at 250 m resolution. It is widely used for EU residential and commercial proposals and is accepted by many European banks for smaller projects.

PVGIS also provides TMY, monthly average, and daily radiation data. The TMY option is best for simulation tools limited to 8,760 hourly values.

Solargis and Solcast

Solargis and Solcast are the leading commercial providers of bankable solar resource data. Both offer global coverage, sub-hourly resolution, probabilistic P50/P90 datasets, and extensive independent validation.

Solcast reports a 207-site DNV validation of its TMY methodology. Solargis publishes validation against ground stations across multiple continents. For utility-scale projects seeking non-recourse financing, these datasets are often mandatory.

NASA POWER

NASA POWER provides free global data at 0.5° resolution, roughly 55 km at the equator. It is useful for early screening in regions without better coverage, but the coarse resolution can miss local microclimates. Use NASA POWER for feasibility-level estimates, not for final design or financing.

TMY vs. Time Series vs. Long-Term Average

These three data products are often confused. Choosing the wrong one is a common source of error in solar analysis.

TMY data is a single synthetic year. It is designed for quick energy simulations, lender-grade yield assessments, and design optimization. It captures typical diurnal and seasonal patterns without requiring the user to process decades of data.

Time-series data preserves the full historical record. It is essential for understanding interannual variability, detecting trends, correlating production with energy prices, and training operational forecasting models. If you need to know how often a site experiences an unusually cloudy year, use time series, not TMY.

Long-term averages (LTA) reduce everything to smooth monthly or daily mean curves. They are useful for site screening and high-level resource mapping, but they erase real weather variability. Simulating a PV system with LTA data can underestimate clipping losses, miss extreme soiling events, and produce unrealistic production profiles.

The choice depends on the question:

P50, P90, and Probabilistic TMY

A single TMY file gives one answer. Real projects need a range. That is where probabilistic datasets come in.

P50: The Median Year

P50 TMY represents the median expected conditions. In any given year, there is a 50% probability that actual production will exceed the P50 estimate and a 50% probability it will fall below. P50 is the standard headline number for energy yield reports and customer proposals.

P90: The Conservative Case

P90 TMY represents a year exceeded with 90% probability. It is the stress case used by lenders to size debt. If a project can service its debt at the P90 production level, the lender has a high degree of confidence in repayment.

The gap between P50 and P90 depends on local climate variability. In stable desert climates, the gap may be 5%. In monsoon-affected or maritime climates, it can exceed 12%. Solargis and Solcast construct P90 TMY by selecting months that align with the probabilistic target rather than the median.

Beyond P90

Some assessments use P75, P95, or P99. P99 is an extreme downside case. Research by kWh Analytics found that roughly 1 in 7 US solar projects underperform their P99 estimate, suggesting that uncertainty is often underestimated in commercial yield assessments.

For project finance, the standard practice is to present P50, P75, and P90. The P90 number is what most non-recourse lenders use for debt sizing.

Accuracy, Uncertainty, and Common Traps

TMY data is a model of reality, not reality itself. Understanding its limits prevents costly design mistakes. For a broader comparison of how simulation tools perform against real output, see our solar simulation accuracy guide.

Typical Uncertainty Ranges

Independent validation studies report the following typical uncertainty ranges for satellite-derived irradiance:

• Annual GHI: ±3–5% in most climates
• Annual DNI: ±8–12%
• Annual POA irradiance: ±4–7% after transposition
• Annual energy yield from solar resource alone: ±5–8%

These figures assume a well-maintained ground station or high-quality satellite dataset. Free, coarse-resolution products can double these errors in difficult climates.

The Microclimate Problem

A TMY file represents the average conditions within its spatial grid cell. A 4 km grid cell may include mountains, coastline, urban heat islands, and agricultural land. A site 2 km inland from a foggy coast can have meaningfully different irradiance than the cell average.

For high-value projects, compare the TMY file against a short-term ground measurement campaign. Even 6–12 months of on-site pyranometer data can reduce solar resource uncertainty by 30–50% when used for site adaptation.

Soiling Is Not in the TMY File

TMY files contain weather. They do not contain dust, pollen, bird droppings, or snow accumulation. Soiling losses must be added separately based on local experience, satellite-derived soiling models, or measured data. In dusty climates, soiling can reduce annual yield by 5–10% if modules are not cleaned regularly.

Climate Trend Risk

TMY datasets are built from historical records. They do not automatically account for shifting climate patterns. Rising temperatures, changing cloud regimes, and increasing aerosol loading can make a TMY based on 1990–2010 data less representative of 2026–2050 conditions. Some providers now offer trend-adjusted or recent-period TMY products to address this.

What Most Designers Get Wrong

The most common TMY mistakes are:

• Using a single global database for all projects regardless of region
• Ignoring spatial resolution and assuming the nearest grid point is representative
• Confusing GHI with plane-of-array irradiance
• Forgetting that DNI uncertainty is much higher than GHI uncertainty
• Applying P50 production estimates in financial models without a downside case

TMY Data in the Solar Design Workflow

A professional workflow treats TMY selection as a deliberate step, not a default. Here is a practical sequence. For quick standalone checks, the SurgePV irradiance estimator and solar power calculator can help validate assumptions before building a full model.

Step 1: Define the Project Requirements

Determine the required accuracy and bankability level. A residential proposal may need only a free TMY source. A utility-scale project seeking non-recourse debt needs a bankable commercial dataset and often a ground-measurement campaign.

Step 2: Select the TMY Source

Choose a source based on geography and project scale:

• US residential and commercial: NSRDB via PVWatts or SurgePV
• EU residential and commercial: PVGIS SARAH-3
• Global commercial and utility: Solargis, Solcast, or Meteonorm
• Early global screening: NASA POWER or PVGIS

Step 3: Validate the File Against Local Knowledge

Check the annual GHI and monthly profile against known local conditions. If the file shows a cloudy July for a site famous for summer drought, investigate. Cross-check with a second independent source when possible.

Step 4: Run the Energy Simulation

Import the TMY into solar design software. The simulator will transpose GHI/DNI/DHI to plane-of-array irradiance, apply temperature and electrical models, and calculate hourly AC output. Make sure the tool models shading accurately; a shadow analysis step is essential for anything other than open-field design.

Step 5: Apply Probabilistic Analysis

For financed projects, generate P50 and P90 yield estimates. Use the P90 number for lender discussions and the P50 number for customer proposals. Document the source, vintage, and uncertainty assumptions.

Step 6: Update After Commissioning

Once the system is operating, compare actual production to the TMY-based forecast. Update soiling, availability, and degradation assumptions for future projects. This closes the loop between design assumption and field reality.

How SurgePV Automates TMY-Based Design

Manual TMY workflows are slow and error-prone. SurgePV pulls satellite-derived weather data automatically for any project location and connects it directly to layout, simulation, and proposal generation.

When you enter an address in SurgePV, the platform selects the appropriate TMY source for the region. It then runs the full calculation chain: transposition to plane-of-array irradiance, module temperature modeling, string and inverter sizing, shading losses, and AC output aggregation. The result is an hourly production profile that feeds into the generation and financial tool and the solar proposal engine.

For designers who want to compare scenarios quickly, SurgePV supports multiple weather datasets and tilt/azimuth options in the same project. Clara AI accelerates the layout and system sizing process, so a complete TMY-based design can move from site address to customer-ready proposal in minutes rather than hours.

The platform also helps avoid common TMY mistakes. It flags when the selected data source has coarse spatial resolution for the site, validates shading geometry against the production profile, and supports P50/P90 reporting for commercial projects.

Key Takeaways for 2026

Solar TMY data is the foundation of every credible energy yield estimate. The file you choose affects production forecasts, financial returns, and lender confidence.

• Match the TMY source to the project. Free sources like NSRDB and PVGIS are sufficient for most distributed solar. Commercial datasets like Solargis and Solcast are the standard for bankable utility-scale work.
• Understand the difference between TMY, time series, and long-term averages. Use the right data type for the question you are trying to answer.
• Always present P50 and P90 for financed projects. The P90 case is what protects lenders and investors from downside weather years.
• Validate against local knowledge and, for high-value projects, ground measurements. Satellite data is excellent but not infallible in complex terrain or microclimates.
• Close the loop after commissioning. Compare actual production to the TMY forecast and update your assumptions for the next project.

For teams designing at scale, the next step is to integrate TMY data into a unified design-to-proposal workflow rather than treating it as a standalone file exchange.

Frequently Asked Questions

What is solar TMY data?

Solar TMY data is a Typical Meteorological Year file: a synthetic year of hourly weather values built from 10–30 years of historical records. It represents long-term typical conditions at a specific location and is the standard input for solar energy yield simulations.

What parameters are included in a TMY file?

A solar TMY file typically includes global horizontal irradiance (GHI), direct normal irradiance (DNI), diffuse horizontal irradiance (DHI), ambient temperature, wind speed, relative humidity, atmospheric pressure, and sometimes albedo or aerosol data. These parameters feed the transposition, temperature, and electrical models in PV simulation software.

What is the difference between TMY and time-series data?

TMY data condenses multi-year records into one representative year of 8,760 hourly values, which is ideal for standard energy yield simulations. Time-series data preserves the full historical record year by year and is used for interannual variability analysis, trend detection, and operational forecasting. Long-term averages smooth everything into monthly means and are only suitable for early screening.

What are P50 and P90 TMY datasets?

P50 TMY represents the median or typical year. P90 TMY represents a conservative year exceeded with 90% probability, used by lenders and investors to size debt against downside production scenarios. The gap between P50 and P90 annual yield typically ranges from 5% to 12%, depending on local climate variability.

Which TMY data source should I use for solar design?

For US projects, NREL NSRDB TMY3 or PSM TMY is the free standard. For European projects, PVGIS SARAH-3 or Solargis is common. For global commercial and utility-scale work, Solargis, Meteonorm, or Solcast provide bankable-grade satellite-derived TMY. NASA POWER is useful for free global screening but has coarser spatial resolution.

How accurate is TMY data?

Validation studies show satellite-derived TMY GHI within ±3–5% of high-quality ground measurements in most climates, with larger errors in cloudy coastal or high-aerosol regions. DNI uncertainty is typically higher, often ±8–12%. Annual energy yield uncertainty from the solar resource alone usually ranges from 5% to 8% for bankable datasets.

Can I use TMY data for operational solar forecasting?

No. TMY data represents long-term typical conditions, not actual weather on a given day. Operational forecasting requires real-time or near-real-time satellite, sky-camera, or numerical weather prediction inputs. TMY is used for pre-construction design and energy yield assessment, not for live dispatch.

How does climate change affect TMY data?

TMY datasets built from older records may become less representative as irradiance, temperature, and precipitation patterns shift. Using data with recent-period coverage or trend-corrected TMY products helps reduce this drift. For projects financed over 25 years, climate-adjusted resource assumptions are increasingly important.