Imagine a 6 MW rooftop array near Hamburg modeled with a single-axis tracker because the same hardware had delivered strong results in southern Spain. After the first full year, production comes in 11% below the consultant’s estimate. The problem is not shading, soiling, or inverter clipping. The site simply has a higher solar diffuse fraction than the model assumed. Northern Germany’s cloudy maritime climate sends roughly half its annual solar resource as scattered light, not direct beam. Trackers cannot steer scattered light.
This is a composite example, but the error pattern is real. Designers routinely import assumptions from one climate to another without checking the DHI/GHI split.
The solar diffuse fraction is the single best predictor of whether a site rewards beam-tracking technology or flat, diffuse-capturing geometry. It also drives tilt optimization, bifacial gain expectations, and shading-loss estimates. Yet many designers still treat it as a hidden output rather than an input they actively design around.
In this guide:
- What solar diffuse fraction means and how it differs from DHI
- The math: how to calculate diffuse fraction from GHI or measured DHI
- How clearness index, aerosols, and clouds control the fraction
- Diffuse fraction ranges by climate type, with real locations
- Why high diffuse fractions change tracker, tilt, and bifacial decisions
- The most common modeling mistakes and how to avoid them
- How SurgePV handles diffuse fraction in design and yield simulations
Related reading: Solar Irradiance GHI DNI DHI Guide 2026
Quick Answer
Solar diffuse fraction is the share of global horizontal irradiance that arrives as scattered sky light, calculated as DHI divided by GHI. It ranges from roughly 0.15 in deserts to more than 0.60 in cloudy maritime climates, and it directly affects tracker economics, optimal tilt, and shading-loss estimates.
What Solar Diffuse Fraction Actually Is
Solar diffuse fraction is a ratio, not an absolute irradiance value. The formula is simple:
Diffuse Fraction = DHI / GHIWhere:
- DHI = Diffuse Horizontal Irradiance, the scattered sunlight reaching a horizontal surface from the entire sky dome except the sun disk
- GHI = Global Horizontal Irradiance, the total sunlight on a horizontal surface
Because GHI = DNI × cos(θz) + DHI, diffuse fraction tells you how much of the total resource is not direct beam. A diffuse fraction of 0.30 means 30% of the energy on a horizontal plane comes from scattered light. The remaining 70% comes from the direct beam component projected onto that horizontal plane.
This distinction matters because direct and diffuse light behave differently on a solar panel. Direct beam arrives from a single direction and creates strong shadows. Diffuse light arrives from the whole sky dome and casts weak, soft shadows. A tracker maximizes direct beam capture. A fixed-tilt array with the right tilt captures diffuse light from the widest possible slice of sky.
The term is sometimes used interchangeably with DHI, but that is imprecise. DHI is a power density measured in W/m². Diffuse fraction is a dimensionless ratio. A site can have low DHI in absolute terms but a high diffuse fraction if direct beam is also weak. Think of DHI as the amount and diffuse fraction as the composition. For a deeper look at DHI itself, see our DHI glossary entry.
How Diffuse Fraction Connects to the Clearness Index
The clearness index, written as Kt or kt, is the ratio of GHI to extraterrestrial irradiance on a horizontal plane:
Kt = GHI / (Eext × cos(θz))Eext is the solar constant adjusted for the earth-sun distance, roughly 1,361 W/m². θz is the solar zenith angle. Kt ranges from near 0 during heavy overcast to about 0.75 under very clear skies. Values above 0.80 usually involve cloud enhancement effects.
Diffuse fraction and clearness index move in opposite directions. The lower the Kt, the cloudier the sky, and the higher the diffuse fraction. This relationship is the foundation of every empirical decomposition model. A scatter plot of diffuse fraction against clearness index slopes downward from top-left to bottom-right. At Kt below 0.2, diffuse fraction is often above 0.8. At Kt above 0.7, it typically falls below 0.25.
However, the relationship is not one-to-one. The same Kt can map to multiple diffuse fractions depending on solar elevation, aerosol optical depth, cloud type, and hour-to-hour variability. This scatter is why simple single-variable models work for annual estimates but struggle with hourly accuracy. It is also why modern solar design software uses either measured DHI or satellite-derived decomposition rather than a single universal curve.
How to Calculate Diffuse Fraction
There are three practical ways to obtain diffuse fraction for a project.
Measured DHI
The most accurate method is direct measurement. A pyranometer fitted with a shadow band or rotating shading disk blocks the direct beam and records only diffuse radiation. The diffuse fraction is then:
Diffuse Fraction = Measured DHI / Measured GHIResearch-grade stations belong to networks such as the Baseline Surface Radiation Network (BSRN). The drawback is coverage. There are fewer than 1,000 research-grade stations worldwide, and many project locations sit hundreds of kilometers from the nearest one. For utility-scale projects, developers sometimes install temporary monitoring stations for 6–12 months to reduce uncertainty.
Satellite-Derived DHI
Most commercial solar design tools use satellite-derived DHI from databases such as NREL’s National Solar Radiation Database (NSRDB), Meteonorm, or the EU Joint Research Centre’s PVGIS. These services decompose satellite cloud observations into GHI, DNI, and DHI using physical or empirical models. Spatial resolution is typically 4 km × 4 km or finer, and temporal resolution ranges from 30 minutes to one minute.
When the database provides DHI directly, diffuse fraction is just DHI divided by GHI. This is the standard approach for residential and commercial design work.
Decomposition Models from GHI
When only GHI is available, empirical models estimate diffuse fraction from Kt and other variables. The classic choices are:
| Model | Inputs | Best For |
|---|---|---|
| Erbs et al. (1982) | GHI, zenith angle, day of year | Hourly data; most widely used |
| Orgill and Hollands (1977) | Kt | Simple hourly estimates |
| Reindl et al. (1990) | Kt, solar altitude, temperature, humidity | Locations with meteorological data |
| Boland-Ridley-Lauret (BRL) | Kt, daily Kt, solar altitude, solar time, persistence | Cloudy, variable climates |
| Engerer2 (2015) | Kt, solar altitude, clearness variability | Cloud enhancement events |
The Erbs model remains the default in many tools because it is simple and well validated. For Kt ≤ 0.22, it sets diffuse fraction to 1 − 0.09Kt. For 0.22 < Kt ≤ 0.80, it uses a fourth-order polynomial. For Kt > 0.80, it fixes diffuse fraction at 0.165. The pvlib implementation of Erbs is a common reference.
For a worked example, consider a site where hourly GHI is 600 W/m², extraterrestrial horizontal irradiance is 1,000 W/m², and measured DHI is 240 W/m². The clearness index is 0.60. The diffuse fraction is 240 / 600 = 0.40. The Erbs model would estimate roughly 0.36 for that Kt, close enough for annual modeling but off by about 10% for that specific hour.
Diffuse Fraction by Climate Type
Climate is the strongest driver of long-term diffuse fraction. The table below shows typical annual ranges based on satellite-derived datasets and ground-station observations:
| Climate Type | Diffuse Fraction | DHI as % of GHI | Example Locations |
|---|---|---|---|
| Arid desert | 0.15–0.25 | 15–25% | Phoenix, Riyadh, Atacama |
| Mediterranean | 0.25–0.35 | 25–35% | Rome, Los Angeles, Cape Town |
| Subtropical humid | 0.30–0.40 | 30–40% | Houston, Mumbai, Brisbane |
| Temperate continental | 0.30–0.45 | 30–45% | Berlin, Chicago, Warsaw |
| Temperate oceanic | 0.45–0.60 | 45–60% | London, Amsterdam, Seattle |
| Tropical monsoon | 0.35–0.50 | 35–50% | Bangkok, Jakarta, Manila |
| Subarctic / coastal | 0.45–0.65 | 45–65% | Reykjavik, Anchorage, Bergen |
These ranges are annual averages. A single site can swing from 0.10 on a clear winter noon to 1.0 during a summer thunderstorm. What matters for design is the annual or monthly profile, not a snapshot.
A useful rule of thumb: if the annual diffuse fraction is above 0.45, direct-beam technologies lose much of their advantage. Concentrating photovoltaics and single-axis trackers fall into this group. If the fraction is below 0.25, beam-dominated designs usually win.
Why Diffuse Fraction Changes Design Decisions
Diffuse fraction is not an academic curiosity. It changes three major design choices.
Tracker vs. Fixed Tilt
Single-axis trackers increase annual yield by pointing modules toward the sun. Their gain comes mostly from direct beam capture. In a desert climate with a diffuse fraction of 0.20, a tracker can add 25–35% over an optimized fixed-tilt array. In a cloudy climate with a diffuse fraction of 0.55, the gain often drops to 10–15%. At that point, the tracker capital cost, O&M burden, and land requirement may no longer pay back.
A 2024 analysis of tracker net gain against diffuse fraction found a clear threshold. For climates with annual diffuse fraction above roughly 0.45, tracker installation becomes economically questionable unless land or labor costs are unusually low. Southern Europe can still justify trackers; northern Europe often cannot.
Optimal Tilt Angle
The optimal tilt angle for a fixed array balances direct beam geometry against diffuse sky capture. In beam-dominated climates, tilting toward the sun at roughly latitude minus 10–15° maximizes annual energy. In diffuse-dominated climates, lower tilts sometimes perform better because they present a larger aperture to the sky dome. Flat commercial roofs in London or Amsterdam often produce more per kWp at 10–15° tilt than at 35–40°.
The exact optimum depends on the diffuse fraction profile through the year. A site with cloudy summers and clear winters may favor a steeper tilt than the annual average suggests.
Bifacial Gain
Bifacial modules capture light reflected from the ground onto the rear side. Diffuse light contributes to both front and rear irradiance. In high-diffuse climates, the rear-side gain can be 8–15% with high albedo and open row spacing. In low-diffuse desert climates, the gain is often 5–10% unless the ground is highly reflective. Designers working in cloudy climates should not assume bifacial gain is small; the diffuse sky can raise rear-side irradiance more than direct beam alone would.
Worked Example: Two Identical Systems, Two Skies
Consider two 500 kW flat-roof commercial projects. Both use the same 580 W monocrystalline modules, the same string inverters, and the same row spacing. One is in Phoenix, Arizona. The other is in Amsterdam, Netherlands. The only deliberate difference is the mounting strategy.
Phoenix has an annual diffuse fraction of roughly 0.22. Amsterdam is near 0.52. Both sites receive similar annual GHI per square meter, but the composition of light is completely different.
| Design Parameter | Phoenix (Beam-Dominated) | Amsterdam (Diffuse-Dominated) |
|---|---|---|
| Annual diffuse fraction | 0.22 | 0.52 |
| Optimal fixed tilt | 28–33° | 10–15° |
| Tracker gain over fixed tilt | 26–30% | 8–12% |
| Shading loss from adjacent row | 2.5–3.5% | 1.5–2.5% |
| Recommended mounting | Tracker or steep fixed tilt | Low-tilt fixed |
| Bifacial rear gain (white roof) | 7–9% | 10–13% |
The Phoenix site should almost certainly use trackers if land and O&M budgets allow. The Amsterdam site should stay fixed-tilt at a low angle. A tracker in Amsterdam adds hardware cost and moving parts for a single-digit production gain. The wrong choice in either direction costs money for 25 years.
This example is intentionally simplified. Local soiling, wind loads, labor rates, and electricity tariffs also matter. But the diffuse fraction is the first filter. It tells you which family of designs is even worth simulating in detail.
Diffuse Fraction in Shading, Yield, and Performance Ratio
Shading and Yield Modeling
Shading losses depend heavily on whether the blocked light is direct or diffuse. When a row, chimney, or tree shades a module, it removes direct beam first. Diffuse light from the rest of the sky dome still reaches the module. In a high-diffuse climate, the same physical shade causes a smaller percentage loss than in a beam-dominated climate.
This is why shadow analysis software separates direct and diffuse shading. A simulation that treats all shading as 100% loss will overestimate losses in cloudy climates and underestimate them in deserts. The difference can be 2–5% of annual yield, which is often larger than the inverter efficiency assumption.
In yield modeling, the diffuse fraction feeds into transposition models such as Perez, Hay-Davies, or Reindl. These models convert horizontal DHI into plane-of-array diffuse irradiance. The Perez model breaks DHI into three components: isotropic background, circumsolar brightening near the sun, and horizon brightening. This anisotropic treatment is more accurate than assuming the sky is uniformly bright, especially for tilted arrays.
Performance Ratio Impact
Performance ratio (PR) is the ratio of actual AC energy output to the theoretical maximum based on installed capacity and in-plane irradiance. It captures every real-world loss: temperature, soiling, shading, inverter efficiency, wiring, mismatch, and downtime. Diffuse fraction affects PR indirectly but measurably.
In high-diffuse climates, module operating temperatures are often lower because diffuse light is less intense per unit area and cloud cover reduces peak irradiance. Lower temperatures raise PR. At the same time, high diffuse fractions increase the importance of diffuse transposition accuracy. If the model assumes an isotropic sky when the real sky has strong horizon or circumsolar brightening, plane-of-array irradiance can be off by 2–4%.
In low-diffuse desert climates, PR is usually dragged down by temperature losses and soiling. The direct beam is strong, but modules run hot and dust accumulates. The design response is different: prioritize low temperature coefficient, cleaning access, and tracker geometry that avoids self-shading at low sun angles.
For financiers and independent engineers, the performance ratio assumption is often the largest driver of long-term revenue. A PR that is 0.02 too optimistic on a 10 MW project can overstate 25-year revenue by several hundred thousand dollars. Getting the diffuse fraction right is one of the easiest ways to keep PR assumptions defensible.
Common Mistakes and Misconceptions
Mistake 1: Using GHI-only tools that hide the diffuse fraction.
Some simple calculators ask only for annual kWh/m² and a generic performance ratio. They never decompose GHI into DNI and DHI. That can work for rough residential estimates in sunny climates, but it fails for cloudy sites or any technology choice that depends on beam versus diffuse split. Always check whether your tool exposes DNI and DHI.
Mistake 2: Applying decomposition models to sub-hourly data.
Erbs and similar models were developed with hourly data. Using them on one-minute data can produce physically impossible results, including negative DNI estimates during cloud enhancement events. For high-resolution modeling, use models explicitly validated at one-minute resolution, such as Engerer2 or recent machine-learning approaches.
Mistake 3: Ignoring diurnal and seasonal variation.
Annual diffuse fraction is useful, but monthly values can differ by 20 percentage points. A site may be beam-dominated in winter and diffuse-dominated in monsoon summer. Seasonal storage or time-of-use value calculations need monthly or hourly decomposition, not an annual average.
Mistake 4: Assuming diffuse fraction is constant over time.
Aerosol reductions, cloud-cover shifts, and urbanization change diffuse fraction. A study of China from 1961 to 2016 found that decreasing aerosol optical depth reduced diffuse fraction in some regions while total radiation changed little. Using a 20-year-old TMY file may misrepresent current conditions. For critical projects, compare recent satellite data against the TMY source period.
Practical Design Checklist
Use this checklist when diffuse fraction could affect a project:
- Pull monthly DHI and GHI, not just annual GHI. Look for seasonal swings.
- Calculate monthly diffuse fraction to see whether the site is consistently cloudy or only seasonally so.
- Compare tracker gain against fixed-tilt gain in your simulation. Do not assume tracker numbers from a different climate.
- Check shading separately for direct and diffuse components.
- Validate decomposition against nearby ground stations if the project size justifies it.
- Use recent weather data, especially in regions with changing aerosol or cloud patterns.
- Document the DHI source in the technical report for financiers and independent engineers.
How SurgePV Models Diffuse Fraction
SurgePV’s solar design software pulls location-specific irradiance data from high-resolution satellite databases. The platform decomposes GHI into DNI and DHI automatically, then applies anisotropic transposition models to calculate plane-of-array irradiance for the actual module tilt and orientation.
Key capabilities useful for diffuse-fraction-sensitive designs:
- Location-specific TMY data: DNI, DHI, and GHI are extracted for the project coordinate, not a regional average. This matters near coastlines, mountains, and urban heat islands.
- Bifacial and rear-side gain modeling: The software accounts for ground-reflected and diffuse rear irradiance based on albedo and row spacing.
- Shading analysis: Direct and diffuse shading are treated separately, so high-diffuse climates do not get penalized with unrealistic loss assumptions.
- Tracker economics: Side-by-side fixed-tilt and tracker simulations let you see the actual gain for the site’s diffuse fraction before committing hardware.
- Generation and financial tool: The generation and financial tool turns the hourly irradiance decomposition into cash-flow projections, payback, and performance ratio reporting.
For quick checks, the irradiance estimator gives diffuse fraction values by location, and the solar panel sizer helps match system capacity to available roof area and resource. When the design is ready, solar proposal software turns the yield numbers into client-ready documents. Those documents show the technology choice and the production forecast side by side. For larger pipelines, Clara AI accelerates preliminary designs by automating layout and irradiance decomposition across multiple sites.
Model Diffuse Fraction Accurately for Any Site
SurgePV automates GHI-to-DHI decomposition, transposition, and shading analysis so your production forecasts match the real sky conditions.
Book a DemoNo commitment required · 20 minutes · Live project walkthrough
Diffuse fraction is not a footnote in a weather file. It is a design input that changes technology choice, tilt angle, shading loss, and financial return. Three actions will put it to work:
- Pull the DHI/GHI split for every project before you choose trackers, tilt, or module type. Use the monthly profile, not just the annual average.
- Validate your decomposition model against local data or satellite-derived DHI. Erbs is a safe default, but cloudy or polluted sites may need Reindl, BRL, or measured data.
- Run fixed-tilt and tracker simulations side by side in solar design software that separates direct and diffuse shading. The diffuse fraction will tell you which option is actually worth paying for.
Sources
- NREL National Solar Radiation Database (NSRDB) — Satellite-derived GHI, DNI, and DHI for the Americas
- Best Practices Handbook for the Collection and Use of Solar Resource Data, 4th Edition — NREL and IEA-PVPS, 2024
- pvlib Erbs model documentation — Reference implementation of the Erbs decomposition
- Effects of Topography and Geography on Solar Diffuse Fraction Modeling in Taiwan — Atmosphere, 2024
- Re-modeling the solar diffuse fraction in Taiwan on basis of a typical-meteorological-year data — Energy, 2025
Frequently Asked Questions
What is solar diffuse fraction?
Solar diffuse fraction is the ratio of diffuse horizontal irradiance (DHI) to global horizontal irradiance (GHI). It shows what share of sunlight arrives as scattered sky light rather than a direct beam. Values range from about 0.15 in deserts to more than 0.60 in cloudy maritime climates.
How do you calculate diffuse fraction?
You calculate diffuse fraction as DHI divided by GHI. When only GHI is measured, empirical models such as Erbs, Reindl, or Boland-Ridley-Lauret estimate it from the clearness index, solar elevation angle, and sometimes temperature or humidity. Most solar design software runs this decomposition automatically using satellite weather data.
Why does diffuse fraction matter for solar design?
Diffuse fraction determines how much energy reaches panels from the sky dome instead of the sun disk. High diffuse fractions favor fixed-tilt and bifacial designs over trackers, reduce shading losses as a percentage of yield, and change the optimal tilt angle. Ignoring it leads to wrong technology choices and inaccurate production forecasts.
What is a typical diffuse fraction by climate?
Desert climates typically show 0.15–0.25. Mediterranean climates show 0.25–0.35. Humid subtropical climates show 0.30–0.40. Temperate oceanic climates like the UK or Pacific Northwest show 0.45–0.60. Overcast conditions push the fraction close to 1.0 because nearly all radiation becomes diffuse.
Which diffuse fraction model is most accurate?
No single model wins everywhere. Erbs remains the most widely used for hourly data. Reindl adds solar elevation, temperature, and humidity. Boland-Ridley-Lauret uses persistence and daily clearness and performs well in cloudy climates. For site-specific work, validate the chosen model against local ground measurements or use satellite-derived DHI directly.
How does diffuse fraction affect tracker economics?
Single-axis trackers gain most of their benefit from direct beam capture. In climates where the annual diffuse fraction exceeds 0.45, tracker gains over optimized fixed-tilt often fall below 15%. In desert climates with diffuse fractions under 0.25, tracker gains can reach 25–35%. The diffuse fraction is therefore a first screen for tracker feasibility.
What is the relationship between clearness index and diffuse fraction?
Clearness index (Kt) is the ratio of GHI to extraterrestrial irradiance. Diffuse fraction and clearness index are inversely related: low Kt means cloudy skies and high diffuse fraction, while high Kt means clear skies and low diffuse fraction. Empirical models use this relationship to estimate DHI from GHI alone.
Can diffuse fraction change over time?
Yes. Aerosol loading, cloud cover shifts, and air quality changes alter diffuse fraction. Studies in China found that aerosol reductions decreased diffuse fraction in some regions over recent decades. Climate trends can also shift cloud patterns, so old TMY data may misrepresent current conditions.
