Solar Insolation: The Designer's Reference

A 10% error in insolation data produces a 10% error in annual yield. On a 500 kWp commercial rooftop, that is 50,000–70,000 kWh per year of misestimated production. Over a 25-year PPA, the revenue gap exceeds $100,000 in some markets. We have seen installers eat clawbacks because they used city-average GHI for a tree-lined suburban roof and ignored the shading factor.

This article is a designer’s reference. It is not a geography lesson, and it is not an exam prep sheet. Every section connects insolation data to a decision you make during design, sizing, or proposal writing. We cover the metrics, the math, the data sources, and the mistakes that blow up projects.

Solar insolation is the single largest input in your energy model. Module efficiency, cabling, and soiling losses all matter, but none of them overcome a bad insolation assumption. Get this number right, and everything downstream falls into place.

In this guide:

• What solar insolation means in designer terms — not textbook definitions
• The difference between irradiance and insolation, and why STC uses 1,000 W/m²
• GHI, DNI, and DHI — the three components and which technology each serves
• Peak sun hours and the step-by-step sizing formula from utility bill to module count
• Tilt angle optimization with city-by-city reference data
• Solar constant, air mass, and why your 400 W module produces 320 W at noon
• Data sources designers actually use — NSRDB, Global Solar Atlas, PVGIS, and more
• How shading and soiling steal insolation before it reaches the cell
• Six common mistakes that destroy proposals and trigger installer clawbacks
• Quick-reference tables for US and international cities
• How solar design software automates insolation workflows from site input to branded proposal

What Is Solar Insolation? (The Designer’s Definition)

Solar insolation is the total solar energy received on a surface over a defined period, measured in kilowatt-hours per square meter (kWh/m²). For designers, the practical unit is kWh/m²/day, because that is what you multiply by system size to get daily energy yield. Annual values in kWh/m²/year are useful for feasibility and finance, but daily numbers drive module counts and string sizing.

The solar constant — the irradiance at the top of Earth’s atmosphere, perpendicular to the sun’s rays — is 1,361 W/m² [CITE]. That is the theoretical maximum before atmospheric absorption, scattering, and weather intervene. By the time sunlight reaches a rooftop in Phoenix, the peak instantaneous irradiance at solar noon is about 1,000 W/m² on a clear day. In Hamburg, a clear-sky noon might reach 900 W/m². These peaks are irradiance. The sum of those peaks over a day, weighted by cloud cover and sun angle, is insolation.

Designers often confuse “sunlight hours” with insolation. A July day in London has 16 hours of daylight, but heavy cloud cover breaks the direct beam into diffuse radiation. The insolation might be 4.5 kWh/m²/day. A December day in Cairo has only 10 hours of daylight, but clear skies and low air mass push insolation to 3.8 kWh/m²/day despite the shorter photoperiod. Duration alone tells you almost nothing. What matters is the energy integrated over time.

This distinction is why we size systems with peak sun hours, not daylight hours. One peak sun hour equals 1 kWh/m² of insolation. It is a standardized energy unit that lets us compare Stockholm in June with Dubai in January using the same metric. A location with 5.2 peak sun hours per day receives 5.2 kWh/m²/day. Multiply that by the system size and a derate factor, and you have estimated production.

The atmospheric path length — expressed as air mass (AM) — is what reduces the solar constant from 1,361 W/m² to ground-level values. At the equator at solar noon, the sun is near zenith and the air mass is roughly 1.0. At 40° latitude in winter, the sun sits lower and the air mass exceeds 2.0, meaning sunlight passes through twice as much atmosphere. More atmosphere means more scattering and absorption, which lowers both peak irradiance and daily insolation. This is why winter insolation drops faster than the simple geometry of shorter days would predict.

For designers, the practical implication is simple: always use location-specific insolation data. Do not extrapolate from a nearby city without checking latitude, elevation, and typical cloud patterns. A site at 1,500 m elevation in the Alps receives measurably more insolation than a sea-level site at the same latitude because there is less atmospheric absorption. A coastal site with morning marine layer fog may underperform a drier inland site 50 km away. The data source you choose, and the resolution it offers, directly determines whether your yield estimate is within 3% or 15% of reality.

Irradiance vs. Insolation — and Why STC Uses 1,000 W/m²

Solar irradiance is instantaneous power density, measured in watts per square meter (W/m²). Solar insolation is cumulative energy, measured in kWh/m² over time. The relationship between them is the same as the relationship between speed and distance. A car traveling at 80 km/h for 3 hours covers 240 km. A rooftop receiving 800 W/m² for 5 hours accumulates 4.0 kWh/m² of insolation.

The 1,000 W/m² figure that appears on every module datasheet is not an average. It is the reference irradiance under Standard Test Conditions (STC), defined in IEC 60904-3. At STC, the irradiance is 1,000 W/m², the cell temperature is 25°C, and the spectral distribution matches AM1.5G — air mass 1.5 global, which corresponds to a solar zenith angle of 48.2° [CITE]. This condition approximates a clear sky at mid-latitude with the sun at roughly 48° from zenith. It is a laboratory benchmark, not a promise of real-world output.

AM1.5G was chosen because it represents a reasonable average solar spectrum for temperate zones. The integrated irradiance under the AM1.5G spectrum is exactly 1,000 W/m². In space, at AM0, the spectrum is harder (more short-wavelength UV) and the total irradiance is 1,361 W/m². On a rooftop in Mumbai at solar noon, the actual air mass might be 1.2–1.5, and the clear-sky irradiance can approach 1,000 W/m². In Stockholm in December, the air mass exceeds 3.5 at noon, and clear-sky irradiance might reach only 400 W/m².

A pyranometer trace at noon shows this clearly. On a clear day in Los Angeles, the trace rises from near zero at sunrise, peaks at 950–1,000 W/m² around solar noon, and falls symmetrically to sunset. The area under that curve — integrated over time — is the day’s insolation in kWh/m². On a partly cloudy day, the trace is jagged. Peaks still hit 900 W/m² when clouds part, but valleys drop to 300 W/m². The area under the jagged curve is lower. A designer looking only at peak irradiance might think the day was productive, but the integrated insolation tells the true story.

Module ratings at 1,000 W/m² and 25°C are useful for comparing products, but they are not what modules produce in the field. A 400 W module at STC might produce 320 W at noon on a clear summer day because the cell temperature is 55°C, not 25°C. The temperature coefficient — typically -0.35% to -0.40% per °C for monocrystalline silicon — means a 30°C temperature rise costs 10.5–12% of rated power. The irradiance might be 1,000 W/m², but the output is not 400 W. In the morning or afternoon, when irradiance is 600 W/m² and temperature is lower, the module might produce 260 W. Insolation data accounts for these variations across the full day.

GHI, DNI, and DHI — The Three Components

Solar insolation arriving at the ground splits into three measurable components. Understanding which one matters for your technology prevents costly mismatches.

The mathematical relationship on a horizontal surface is:

GHI = DHI + (DNI × cos(θz))

Where θz is the solar zenith angle. At solar noon on the equinox at 40° latitude, θz = 40° and cos(40°) = 0.766. If DNI is 850 W/m² and DHI is 150 W/m², then GHI = 150 + (850 × 0.766) = 150 + 651 = 801 W/m².

Consider two sites with identical annual GHI but very different compositions. Phoenix averages roughly 2,135 kWh/m²/year of GHI, with DNI dominating at about 65–70% of the total. Munich averages roughly 1,188 kWh/m²/year of GHI, but DHI accounts for 45–50% of the total because persistent cloud cover breaks the direct beam into diffuse radiation. A flat-plate fixed-tilt system in Munich still performs well because silicon modules convert diffuse light efficiently. A concentrating solar power (CSP) plant in Munich would be a non-starter because CSP requires direct beam radiation that can be focused by mirrors.

For flat-plate PV designers, GHI is the starting point. The software decomposes GHI into direct and diffuse components, transposes each onto the tilted array plane, and sums them into plane-of-array (POA) irradiance. DNI matters indirectly because it drives the direct component on your tilted surface. DHI matters because it is the fraction you still collect on overcast days. In high-DNI climates like Arizona or Rajasthan, tilt optimization provides larger absolute gains because the direct beam is the dominant term. In high-DHI climates like Germany or the UK, the diffuse model in your simulation software matters more than the direct beam model. A transposition model that handles diffuse poorly will introduce 3–5% error in northern Europe.

Peak Sun Hours and the Sizing Formula

Peak sun hours (PSH) is the unit that converts abstract insolation data into module counts. One PSH equals 1 kWh/m² of solar insolation. A site with 5.2 PSH/day receives the same energy as 5.2 hours of full 1,000 W/m² irradiance. This linearity is what makes the sizing formula work.

Here is the step-by-step walkthrough from utility bill to module count:

Step 1: Find daily energy usage.

A home uses 500 kWh/month. Divide by 30 days: 500 ÷ 30 = 16.7 kWh/day.

Step 2: Look up peak sun hours for the site.

The project is in a region with 5.2 PSH/day. Use NSRDB, Global Solar Atlas, or PVGIS for this value. Always verify the data source and year range.

Step 3: Apply the derate factor.

Real systems lose energy to temperature, soiling, shading, mismatch, wiring, and conversion losses. NREL PVWatts uses a default derate factor of 0.84 for residential systems [CITE]. A conservative designer uses 0.80 for older roofs, partial shading, or dusty climates.

Step 4: Calculate required DC system size.

PV System Size (kW) = Daily kWh Usage ÷ Peak Sun Hours ÷ Derate Factor

PV System Size = 16.7 kWh/day ÷ 5.2 PSH ÷ 0.84 = 3.82 kW DC

Step 5: Convert to module count.

If using 440 W modules: 3,820 W ÷ 440 W/module = 8.7 modules. Round up to 9 modules for 3.96 kW DC. In practice, you also round to fit the roof geometry, inverter input voltage windows, and fire setback requirements.

This formula is the backbone of every preliminary proposal. It is not a replacement for hourly simulation, but it gets you within 5–10% of the simulated result in under 2 minutes. We use it in the field during site visits to give clients a rough number before returning to the office for detailed modeling.

The derate factor is where most hand calculations go wrong. A designer who uses 0.90 because they are optimistic about module efficiency will undersize the system by 7% relative to PVWatts defaults. A designer who uses 0.75 because they are being “safe” will oversize by 11%, raising the project cost unnecessarily. Document your derate assumptions in every proposal. “0.84 per NREL PVWatts defaults, adjusted to 0.80 for documented tree shading on the southwest roof face” is a defensible statement.

Peak sun hours vary seasonally, and this matters for production estimates. A site with 6.5 PSH in July might drop to 2.8 PSH in December. Annual averages are fine for grid-tied payback calculations, but if the client cares about winter self-consumption or export limits, run the monthly breakdown. NREL PVWatts and PVGIS both provide monthly PSH tables. We always pull the monthly view before finalizing a proposal, even when the headline number is the annual average.

Tilt Angle Optimization

Tilt angle is the single design parameter that every installer can control, yet many still mount flat on commercial roofs or copy the roof pitch without checking the annual energy cost. A flat-mounted array in Denver loses roughly 21% of annual production relative to optimal tilt [CITE]. In Los Angeles, the flat-roof penalty is about 17% [CITE]. These are not marginal losses. They are proposal-killers when a competitor quotes a properly tilted system on the same roof.

The year-round rule of thumb is straightforward: set tilt approximately equal to latitude. At 40°N, a 40° tilt captures the maximum annual insolation for a fixed south-facing array. This rule is not exact — NREL PVWatts typically returns an optimum 3–7° below latitude at mid-latitudes because summer days are longer and the diffuse fraction shifts the balance — but it is accurate enough for preliminary design.

Seasonal adjustment changes the calculus. For summer-optimized systems — common for seasonal loads like irrigation pumps or vacation homes — set tilt to latitude minus 10–15°. For winter-optimized systems — relevant for remote sites with severe winter heating loads — set tilt to latitude plus 10–15°. Two manual adjustments per year can recover 5–12% of annual production at latitudes above 45°, but the hardware and labor cost rarely justify it below 35° latitude.

The good news for installers is that precision is not required. NREL and PVGIS data both confirm that deviations of ±10° from optimal tilt, or azimuth deviations of ±30° from true south, do not appreciably change annual performance [CITE]. A 10° tilt error costs roughly 2–3% of annual yield. A 30° azimuth error costs roughly 3–4%. These are tolerable in residential design where roof geometry and aesthetics constrain the physical installation.

The bad news is that flat mounting — 0° tilt — is not a minor deviation. It is a full 30–50° miss from optimal at most latitudes, and the cosine penalty is severe. On commercial rooftops, flat mounting is often driven by structural loading, wind uplift, or ballast constraints rather than energy optimization. When you propose a flat-mounted system, quantify the production loss explicitly. Show the client the tilted alternative and the corresponding payback difference. Let them choose with data, not assumptions.

Solar Constant, Air Mass, and Real-World Reality

The solar constant is 1,361 W/m² at AM0 — outside Earth’s atmosphere, perpendicular to the sun’s rays [CITE]. This is the theoretical ceiling. Every watt of ground-level insolation is what remains after atmospheric absorption, Rayleigh scattering, Mie scattering from aerosols, and cloud reflection have taken their share.

AM0 = 1,361 W/m² applies to satellites and space-based PV. It is irrelevant for rooftop designers except as a reference point. The spectrum at AM0 is harder — richer in ultraviolet — than the ground-level spectrum. Space-rated cells use different materials and encapsulants because UV degradation is harsher at AM0.

AM1.5G = 1,000 W/m² is the standard ground-level reference for module testing. It corresponds to a solar zenith angle of 48.2°, which approximates mid-latitude noon conditions with average atmospheric clarity. At lower latitudes near solar noon, actual air mass can drop to 1.0–1.2, and clear-sky irradiance can exceed 1,000 W/m². At higher latitudes in winter, air mass can exceed 3.0, and clear-sky irradiance might reach only 400–500 W/m².

A rooftop in Mumbai at 19°N sees an air mass of roughly 1.05 at solar noon on the equinox. On a clear day, the direct-normal irradiance can approach 950 W/m². A rooftop in Berlin at 52°N in December sees an air mass of 3.5 at noon. Even with a clear sky, the direct beam has passed through 3.5 times more atmosphere, and the irradiance might reach only 350 W/m². This is why insolation drops so sharply in winter at high latitudes. It is not just shorter days; it is also thicker atmosphere.

Soiling adds another real-world layer. Dust, pollen, bird droppings, and industrial particulates accumulate on module surfaces and reduce transmitted irradiance. In arid climates, soiling losses of 0.3–0.5% per day are common without rain. In agricultural areas, pollen seasons can create a 5–8% loss for several weeks. In urban environments near construction or heavy traffic, particulate deposition is a year-round factor. These losses are not captured in satellite-derived insolation databases. They are site-specific operational realities that designers must add as derate factors.

Data Sources Designers Actually Use

Not all insolation databases are equal. The one you choose should match your project’s geography, scale, and financing requirements. Here is what we use in practice.

NREL NSRDB is the gold standard for US projects. The Physical Solar Model (PSM) dataset uses satellite imagery and atmospheric models to produce hourly GHI, DNI, and DHI at 4 km resolution across the Americas. We use NSRDB for every US project in SurgePV because it is free, well-documented, and accepted by utilities and AHJs.

Global Solar Atlas, maintained by the World Bank, provides quick visual overviews and downloadable data for any location worldwide. It is our first stop for international feasibility. The accuracy ranges from ±4% in well-satellite-covered regions to ±8% in data-sparse areas. For a 2 MW ground-mount in Kenya, that is an 80–160 MWh/year uncertainty band. Use Global Solar Atlas for go/no-go decisions, then commission a site-specific Meteonorm or Solargis report for detailed design.

NASA SSE / POWER is free and global, but the ~50 km resolution is too coarse for project-level work. A single grid cell might span a coastal plain and a mountain ridge. We use NASA POWER only for early-stage screening in regions where no other data exists. Never use it for a commercial feasibility study that will be presented to investors or lenders.

PVGIS is the European standard. Version 5.3 uses the SARAH-3 satellite dataset (2005–2023) and is extensively validated against European ground stations. For residential and small commercial projects in Europe, PVGIS is sufficient for design and permitting. For bankable reports, lenders expect Meteonorm or Solargis.

Fraunhofer ISE operates a network of ground stations across Germany and neighboring countries. Their real-time data is useful for validation and research, but it does not replace the long-term climate normals you need for yield prediction.

How Shading and Soiling Steal Insolation

Shading and soiling are the two largest site-specific reductions to insolation that satellite databases cannot see. A tree that blocks morning sun, a neighboring building that casts an afternoon shadow, or a layer of desert dust that persists for weeks — each one reduces the solar energy that reaches the module surface.

A 10% reduction in solar access typically causes a 10–15% production loss [CITE]. The nonlinear penalty happens because shaded cells in a string act as resistors. Instead of contributing current, they dissipate power from unshaded cells as heat. Bypass diodes mitigate this, but they do not eliminate the loss. In extreme cases, a single shaded module in a string can drag the entire string down by 30–50% for that time period.

Total Solar Resource Fraction (TSRF) quantifies this. TSRF is the product of solar access and tilt and orientation factor (TOF):

TSRF = Solar Access × TOF

Solar access is the ratio of actual insolation reaching the array to the insolation that would reach an unshaded array at the same tilt and azimuth. TOF is the ratio of insolation at the actual tilt/azimuth to insolation at the optimal tilt and azimuth. A system with 92% solar access and 98% TOF has a TSRF of 0.92 × 0.98 = 0.9016, or 90.2%. That is a 9.8% production loss relative to ideal conditions.

Soiling is equally insidious because it is invisible in design software unless you manually apply a derate factor. Satellite insolation data assumes a clean module surface. Real modules in dusty climates — Rajasthan, Arizona, parts of the Middle East — can lose 5–10% of annual yield to accumulated dust. Rain cleans modules effectively, but dry seasons of 4–6 months create sustained losses. Industrial areas with particulate emissions, agricultural zones with pollen, and coastal areas with salt spray all add unique soiling profiles.

Three mistakes we see repeatedly: ignoring soiling in dusty climates and using the default 2% loss when 8% is realistic; using unshaded GHI for a roof surrounded by mature trees; and applying a single TSRF to the entire array when shading affects only one string. Each mistake compounds into a proposal that overpromises and underdelivers. The fix is site-specific shading analysis with tools that model hourly obstruction geometry across the full year.

Common Mistakes That Blow Up Proposals

Insolation errors do not stay in the design phase. They follow the project into commissioning, operations, and sometimes litigation. Here are six mistakes we have seen cost installers money.

1. Using annual average insolation for systems sized against worst-month demand.

An annual average of 4.5 PSH/day looks healthy. But if December drops to 2.1 PSH and the client needs full winter production, the annual number is irrelevant. Size for the worst month, then check if the array still makes financial sense. If it does not, set expectations with the client before construction starts.

2. Applying horizontal GHI to a steep tilt without transposition.

Horizontal GHI is what databases report. Plane-of-array (POA) irradiance is what the array actually sees. A 35° south-facing array in Denver receives roughly 21% more annual POA irradiance than horizontal GHI [CITE]. Using raw GHI as the array input underestimates yield by that margin. Always run the transposition step — or use software that does it automatically.

3. Ignoring soiling derates in desert or industrial regions.

Default soiling losses of 2% are fine for rainy temperate zones. In Phoenix or Dubai, 2% is a fantasy. We use 5–8% for desert sites without cleaning contracts, and 3–5% for industrial zones near highways or manufacturing. If you use 2% and actual soiling is 7%, your year-one production shortfall is 5% — enough to trigger performance guarantees in some PPA structures.

4. Using city-average data for microclimate sites.

San Francisco city average GHI is moderate, but the Sunset District is perpetually fogged in while Potrero Hill bakes in sun 10 km away. Elevation, coastal influence, and urban heat islands create microclimates that 10 km grid data cannot resolve. If the site is more than 5 km from the data point, add a 2–3% uncertainty band to your yield estimate.

5. Confusing DNI with GHI for flat-plate module counts.

A designer sees a high DNI number in a CSP database and assumes the site is excellent for flat-plate PV. But DNI does not include diffuse radiation. In some high-DNI deserts, the diffuse fraction is small and GHI is also high. In others, atmospheric dust scatters light and DNI is high while GHI is moderate. Always use GHI for flat-plate counts, not DNI.

6. Promising production based on ideal conditions without shading analysis.

The unshaded GHI from NSRDB assumes an open horizon. A site with a three-story building to the west will lose 15–25% of afternoon insolation in winter. If you quote unshaded numbers, the client will measure a 12% shortfall in year one and ask for compensation. We have seen installers pay clawbacks because they skipped the shading step to save 30 minutes.

Quick-Reference Tables for Designers

These tables consolidate the numbers you need for preliminary sizing and client conversations. All values are long-term annual averages. Specific yield assumes a performance ratio of 0.82, which reflects real-world losses from wiring, temperature, soiling, and mismatch.

US Cities

International Cities

The spread is striking. Helsinki averages 2.41 kWh/m²/day [CITE], while Central Australia reaches 5.89 kWh/m²/day [CITE]. That is a 2.4× difference in energy input. No module technology can close a gap that large. A designer proposing identical system sizes in both locations would be committing professional malpractice.

Use these tables for order-of-magnitude sizing and client education. For final design, always pull site-specific data from NSRDB, Global Solar Atlas, or PVGIS. The tables above are starting points, not substitutes for location-specific analysis.

How SurgePV Automates Insolation Workflows

Every step described in this guide — from data lookup to module count to financial model — runs inside a single solar software workflow. Here is how we do it.

Step 1: Address input.

Enter the project address. SurgePV pulls the lat/long and queries integrated irradiance databases. You get GHI, DNI, DHI, and temperature data matched to the site. No manual database navigation, no spreadsheet imports.

Step 2: 3D rooftop model.

The solar designing module builds a 3D rooftop model from satellite imagery and user inputs. You define roof planes, pitch, azimuth, and obstructions. The model knows the geometry before a single module is placed.

Step 3: Shade analysis.

solar shadow analysis software runs physics-based irradiance calculations on the cloud-rendered 3D model. Trees, buildings, and mechanical equipment cast shadows that are traced hour by hour across the year. The output is a solar access map and a TSRF value for every roof plane. This is where satellite insolation data gets corrected for real-world shading.

Step 4: Energy simulation.

The generation and financial tool applies transposition models, temperature corrections, soiling assumptions, and system losses to the shaded insolation data. The output is hourly, monthly, and annual production in kWh, plus specific yield in kWh/kWp. You see exactly how much energy the designed system will produce, not a theoretical maximum.

Step 5: Branded PDF proposal.

The solar proposal software pulls the 3D model, shade report, energy simulation, and financial model into a branded PDF. The client sees the insolation data source, the expected production, the payback period, and the IRR in one document. There is no gap between engineering and sales.

The time savings are substantial. A manual workflow — NSRDB lookup, Excel sizing, SketchUp shading, PVWatts simulation, Word proposal — takes 2–3 hours for a residential project. In SurgePV, the same project runs from address to proposal in 20 minutes. The accuracy is higher because there are no manual data transfers, no copy-paste errors, and no version mismatches between tools.

Conclusion

Solar insolation is the foundation of every PV design decision. Module counts, string sizing, energy yield, payback period, and PPA pricing all flow from this single input. A designer who masters insolation data — GHI, DNI, DHI, peak sun hours, tilt optimization, and shading corrections — produces proposals that survive commissioning and satisfy clients for 25 years.

Installers who treat insolation as an afterthought pay for it later. We have seen 12% production shortfalls trigger contract disputes because the designer used city-average GHI for a microclimate site. We have seen flat-roof commercial arrays underperform by 20% because no one ran the transposition math. These are not equipment failures. They are data failures, and they are preventable.

Three actions to take from this reference:

1. Audit your current insolation source. If you are using unvalidated data or city averages for site-specific designs, switch to NSRDB (US), PVGIS (Europe), or Global Solar Atlas (international). Cross-reference two sources for every commercial project. Document the source name and version in every yield report.

2. Add shading and soiling to your standard derate checklist. Satellite GHI is a starting point, not a finish line. Walk the site, model obstructions with 3D shading tools, and apply soiling factors that match the local climate. A 5% soiling loss in a desert region is not conservative — it is realistic.

3. Document every assumption in your proposal. Data source, version, tilt angle, azimuth, derate factor, and shading loss percentage. A proposal with documented assumptions is defensible. A proposal with a single yield number is a liability waiting to become a clawback.

If you want to see how insolation data flows from address input to branded proposal in one workflow, book a demo. We will walk through a live project and show you the numbers.

Frequently Asked Questions

What is the difference between solar irradiance and solar insolation?

Solar irradiance is instantaneous power density, measured in watts per square meter (W/m²). It is what a pyranometer reads at a single moment. Solar insolation is cumulative energy over time, measured in kilowatt-hours per square meter (kWh/m²). Think of irradiance as the speed on your car’s speedometer, and insolation as the distance on your odometer. A module datasheet lists power at 1,000 W/m² irradiance. Your energy model needs insolation in kWh/m²/day to calculate annual production.

How do you calculate solar insolation for a PV system?

Find the daily peak sun hours for your exact location from NSRDB, Global Solar Atlas, or PVGIS. Multiply your daily energy demand in kWh by the peak sun hours to get a raw system size. Then divide by a derate factor — typically 0.80–0.84 — to account for temperature, soiling, wiring, and conversion losses. For a home using 20 kWh/day in a 5.2 PSH region with a 0.84 derate: 20 ÷ 5.2 ÷ 0.84 = 4.58 kW DC required. Divide by your module wattage to get the module count.

What is a peak sun hour?

One peak sun hour equals 1 kWh/m² of solar energy received on a surface. It is a standardized unit that lets designers compare locations and seasons using the same metric. A site with 5.2 peak sun hours per day receives the same total energy as 5.2 hours of continuous irradiance at 1,000 W/m². In reality, those 5.2 kWh/m² might arrive as 8 hours of varying irradiance — 400 W/m² at dawn, 950 W/m² at noon, 200 W/m² under afternoon clouds. The peak sun hour compresses that variability into a single useful number for sizing calculations.

What is the best tilt angle for solar panels?

For fixed-tilt systems maximizing annual production, set the tilt angle approximately equal to your latitude. At 40°N, a 40° tilt is the starting point. NREL PVWatts typically returns an optimum 3–7° below latitude at mid-latitudes because summer day length and diffuse fractions shift the balance. For summer-only optimization, subtract 10–15° from latitude. For winter-only optimization, add 10–15°. The good news is that precision is not critical — deviations of ±10° from optimal tilt, or ±30° from true south, do not appreciably change annual performance.

Which solar insolation database is most accurate?

For projects in the United States, NREL NSRDB is the gold standard with ±5% GHI accuracy and 4 km spatial resolution. For international projects, Global Solar Atlas offers global coverage with ±4% to ±8% accuracy depending on satellite coverage density. PVGIS is the standard for Europe and Africa, with typical GHI accuracy of ±3–5%. For bankable commercial projects, use Meteonorm or Solargis. Always cross-reference at least two independent sources. If they disagree by more than 5%, investigate the cause before finalizing your yield estimate.

How does shading affect solar insolation?

Shading reduces solar access — the fraction of available insolation that actually reaches the module surface. A 10% drop in solar access typically causes a 10–15% production loss because shaded cells in a string act as resistors, dissipating power from unshaded cells. Bypass diodes limit the damage to one substring per diode, but they do not eliminate the loss. Total Solar Resource Fraction (TSRF) quantifies the combined effect of shading and suboptimal tilt/orientation. Professional shading analysis using 3D obstruction modeling is essential for any site with trees, neighboring buildings, or mechanical equipment on the roof.