Solar Radiation Atmospheric Attenuation Calculator
Before it reaches the ground, sunlight travels through a slice of atmosphere that absorbs and scatters part of its energy — air molecules (Rayleigh scattering), ozone, water vapour, aerosols. That slice isn't a fixed thickness: it depends directly on the sun's elevation above the horizon. Straight overhead, the rays cross the atmosphere perpendicular to its layers, along the shortest possible path. The lower the sun sits, the longer that path gets — up to roughly forty times the minimum path right at the horizon.
This path-length ratio has a name: the optical air mass (AM). It equals 1 straight overhead (AM1) and underpins the reference spectrum used across the photovoltaic industry, AM1.5 (zenith angle ≈ 48.2°, elevation ≈ 42°), the standard used to test solar panels in the lab.
This page calculates purely atmospheric attenuation, measured on a surface held perpendicular to the rays (direct normal irradiance, DNI). It doesn't account for the tilt of a fixed receiving surface — that geometric factor, which applies just as much to a solar panel as to a wall or a shop window, is calculated separately on the Solar Incidence Angle page.
Sun elevation
Reference values at common elevations
Regardless of the figure entered above, here is the air mass and the transmitted fraction for a few characteristic elevations — from right at the horizon to straight overhead.
| Elevation h | Air mass AM | Transmitted fraction |
|---|---|---|
| 2° | 19.43 | 7.0 % |
| 5° | 10.31 | 17.7 % |
| 10° | 5.59 | 31.8 % |
| 15° | 3.81 | 41.3 % |
| 20° | 2.90 | 48.0 % |
| 30° | 1.99 | 56.6 % |
| 45° | 1.41 | 63.7 % |
| 60° | 1.15 | 67.5 % |
| 75° | 1.03 | 69.4 % |
| 90° (overhead) | 1.00 | 70.0 % |
AM1.5 (elevation ≈ 42°, zenith angle ≈ 48°) is the reference spectrum used to test photovoltaic panels in the lab.
Starred terms are unpacked in the glossary after the last section.
Optical air mass (AM) and the AM1.5 laboratory standard
Sunlight does not arrive at a collector as it left the Sun. It has crossed a stack of air. Optical air mass* (AM) is how many times longer that path is than the shortest one: straight down from the zenith. AM1 is that vertical shot. AM2 already means twice the air, a lower Sun, and more blue stripped out of the beam.
A flat-Earth shortcut is AM = 1 / cos θz, with the zenith angle* θz = 90° − h. Fine while the Sun is high. Near the horizon cos θz collapses, the formula blows up, and the atmosphere is not a pile of tiles: the Earth is round. This page uses Kasten and Young (1989), which stays sane down to the horizon.
AM = 1 ÷ [cos θz + 0.50572 · (96.07995 − θz)−1.6364]
AM1.5* is the laboratory spectrum in IEC 60904-3. The matching geometry is θz ≈ 48.2°, so an elevation of about 42° — not 48°. The 1.5 is the air-mass number, not the elevation. Nameplate watts are measured against that spectrum at 1,000 W/m², cell at 25 °C. A “400 W” module is 400 W in that booth, not at noon on a UK roof in January.
The 42° elevation is a decent annual stand-in for temperate latitudes under a “standard” clear sky. It is not a claim that the Sun lives there. Glasgow in December never sees 42° at noon. The label quotes AM1.5 anyway, because the industry needed one booth, not a booth per postcode.
Rayleigh scattering, ozone and water vapour
The direct beam is not dimmed by one curtain. Three processes share the job, and they pick on different colours.
- Rayleigh scattering* — air molecules (N2, O2), much smaller than the wavelength. They scatter violet and blue hardest (roughly 1/λ4). That is why the sky is blue, and why the disc goes red when the path stretches.
- Ozone — mostly stratospheric. It eats UV, and a little visible light (the Chappuis band, around 600 nm).
- Water vapour — wildly variable. Near-infrared bands at 940, 1,100 and 1,400 nm. A muggy afternoon is not a dry high-pressure day.
Aerosols (dust, pollution, sea salt) plus a little CO2 and methane sit on top. This calculator folds them into an average haze. It does not “see” a London inversion, or Saharan dust over the Atlantic.
I / I0 = 0.7AM0.678 (Meinel & Meinel, 1976)
I0 = 1,367 W/m² (mean solar constant, top of atmosphere)
The solar constant* I0 is the mean extra-atmospheric irradiance on a plane facing the Sun. 1,367 W/m² is the conventional mean used here. At AM1, 0.71 = 70% of that beam still gets through. That is an engineer’s clear sky, not a pyrheliometer reading on a given Tuesday.
Clear sky versus cloud: what this model leaves out
The form assumes a clear sky and average turbidity. Good for a ceiling, for comparing two solar elevations on equal terms, for seeing what an 8° sunrise costs against a 60° noon.
It does not describe a grey November in Manchester, a summer thunderstorm, or a sea fret. Cloud kills the beam and pumps up the diffuse — sometimes, at a bright cloud edge, even above the clear-sky beam for a minute. For a real year, climate data are needed: global horizontal irradiance, measured DNI, or the solar panel output calculator next door. Here the only variable isolated is elevation.
Below 5° elevation, refraction, humidity and local haze make the simplified model steadily less trustworthy. That is also where the Sun turns red.
Direct beam versus diffuse sky radiation
What hits the ground splits. Direct (beam) comes from the disc, undeviated. It casts a sharp shadow; it is what a concentrator can focus. Diffuse arrives from the whole vault after scattering. Under a clear sky it is often 10% to 20% of the global. Under overcast it can carry the lot.
Energy absorbed by ozone and vapour becomes heat and leaves as infrared. Scattered energy partly goes back to space, partly down to the ground. A flat panel drinks both. A concentrating plant wants only the beam. That is why DNI, below, matters for some jobs and not for others.
A rooftop module on a dull British day is living on diffuse. The number this page prints is then a ceiling that the cloud has already taken away. The arithmetic remains useful: it shows how much path length alone would have cost, if the sky had been the standard clear sky.
Site altitude and a shorter air column
Higher up, the air is thinner and usually drier. For the same solar elevation, AM drops a little: part of the column has already been climbed. Rayleigh eases off, vapour does too, more blue survives, DNI rises.
An observatory or a high plateau: 15% to 30% more DNI than at sea level is not unusual, especially in UV and near-IR. This page does not ask for altitude. It computes AM as if the site were at sea level. In the mountains the figure is a touch pessimistic — on purpose. One input: the Sun’s height.
Most of Great Britain sits near sea level. A few hundred metres on a Pennine ridge is not the Andes. The altitude correction is real and, for a typical UK roof, small next to cloud and season.
How solar elevation changes with season, hour and latitude
h is the angle between the Sun and the horizon. It sets AM, and the incidence angle on a roof. Three knobs:
- Latitude φ — farther from the equator, the lower the winter noon.
- Season — declination δ runs from −23.44° (northern winter solstice) to +23.44° (summer).
- Solar time — hour angle ω; the peak is solar noon, not 12:00 on the clock.
sin h = sin φ · sin δ + cos φ · cos δ · cos ω
For a date, time and place, the sun altitude and azimuth page gives h. That value can be pasted here. A Glasgow December noon may stall around 12° to 13°. A midsummer noon in the south of England can top 60°. Different AM, different DNI, different red at sunset.
Latitude does more than shave a few degrees off noon. At 56° N the winter Sun never climbs out of the long-path regime. AM stays far above 1.5 for most of the daylight hours. That is a spectral fact and an energy fact: the beam is redder, and a larger share of what arrives is diffuse.
Incidence on a tilted collector
DNI is irradiance on a plane facing the rays. A fixed roof almost never does, except for a moment. On the module plane:
Idirect, plane = DNI × cos θi
The incidence angle θi comes from the Sun’s position and the plane (tilt, azimuth). The solar incidence angle page does that geometry. Diffuse is not a beam. To first order it is treated as isotropic, with a sky view factor (1 + cos β) / 2. A near-vertical panel “sees” less sky.
This page stops at DNI. The panel-output calculator next door does transposition, efficiency and performance ratio. They fit together. They are not the same job.
Direct normal irradiance for concentrating plant and trackers
Direct normal irradiance* is disc power on a square metre held square-on. Concentrating plants (towers, Fresnel, dishes) use only that. A two-axis tracker almost does too: it keeps cos θi near 1.
A pyrheliometer* on a tracker, collimated tube, measures DNI. A pyranometer sees global (beam + diffuse) on a plane. Different instrument, different number. Clear sky at AM1.5, DNI around 900 W/m² is ordinary. AM1, dry air, altitude: it can pass 1,000 W/m². A milky lowland “clear” day sits much lower — Britain’s winter DNI is a fraction of Spain’s.
There is almost no concentrating solar in the UK, and for this reason. The beam is too often missing. Trackers on a farm-scale PV array help even then, because they keep the modules facing whatever beam there is. They do not invent a beam under stratus.
Why the Sun is red at the horizon
AM often exceeds 10 on the horizon. Blue has time to be swept out of the beam — that is the sky, elsewhere. Red and orange, less scattered, remain. Haze (aerosols, dust, pollution) adds Mie scattering from bigger particles and can make the show louder. A postcard sunset after a dust plume is not more Sun. It is less blue toward the disc.
The evening red is the same physics as the 70% at zenith, stretched. The model shows it. It should not be asked for DNI to the nearest watt at 2° elevation.
Refraction lifts the apparent disc a little at the horizon, which is why the geometric AM and the visual AM part company down there. Kasten and Young already fold a refraction correction into the formula used here. The remaining uncertainty is the air itself.
A UK winter noon versus the AM1.5 spectrum
The IEC/NREL AM1.5 tables were built for a reasonably sunny temperate site, not for a low December Sun at 55° N. In the UK the midday AM in winter is often well above 1.5. The beam is redder, weaker, and a larger share of what arrives is diffuse.
The form will give a number all the same. Treat it as “if the sky were the standard clear sky”, not as yesterday’s pyranometer. A December noon in Aberdeen at 12° elevation is AM ≈ 4.7 on this page, and about 40% of I0 in the beam under that clear-sky assumption. Overlay a typical overcast and the beam is gone.
Summer is kinder. A June noon in the south of England, elevation near 62°, sits close to AM1.1. The laboratory AM1.5 spectrum is then a slightly pessimistic stand-in for the beam, and a fair one for the year as a whole — which is why the industry kept it.
Met Office sunshine hours are not DNI
The Campbell–Stokes sunshine recorder, and the modern sensors that replaced it, count hours when the beam is strong enough to burn a card or trip a threshold. That is a duration, not a watt per square metre. A “sunny” hour in February is not the same energy as a sunny hour in June, because AM is different and the day is shorter.
Maps of “hours of sunshine” are useful journalism. They are a poor input to this calculator. What this page wants is an elevation, then it returns a clear-sky DNI. What a yield study wants is a time series of GHI and DNI, or a typical meteorological year. Mixing the two is how a Cornwall brochure and a Cumbrian roof fall out.
PVGIS, the MCS irradiance database, and a site survey live in that second world. This page lives in the first: one angle, one clear-sky beam.
Maritime haze around the British Isles
The Meinel 0.7 factor is an average continental haze. Western Britain often sits in maritime air: sea salt, water vapour, and a low cloud base. The beam on a “clear” Atlantic day can be softer than the same elevation in central Spain. Eastern England, under a dry high, can look closer to the textbook.
Saharan dust does reach the UK, a few times a year, and it is visible. The quieter, everyday penalty is the moist boundary layer. This calculator does not take a visibility or a PM10. It takes an elevation. For a roof in Cornwall the printed DNI is an upper bound on a clear day, not a promise.
Aerosols also change the spectrum. More scattering means a bluer sky and a redder, weaker disc — the same split as a longer path, for a different reason. Two days with the same solar elevation can therefore print the same number here and measure differently outside.
Glossary
- Optical air mass (AM)
- The ratio of the actual slant path through the atmosphere to the shortest (vertical) path. AM1 is the Sun overhead. AM grows as the Sun sinks. This page uses the Kasten and Young (1989) formula, which remains finite at the horizon.
- AM1.5
- The reference spectrum for testing photovoltaic modules (IEC 60904-3). It corresponds to a zenith angle of about 48.2°, hence an elevation of about 42°, under a defined clear sky. Nameplate power is measured against AM1.5 at 1,000 W/m² and 25 °C cell temperature.
- Zenith angle θz
- The angle between the Sun and the vertical. θz = 90° − h, where h is the elevation above the horizon. AM1.5 uses θz ≈ 48.2°, not an elevation of 48°.
- Rayleigh scattering
- Scattering of sunlight by air molecules much smaller than the wavelength. Intensity falls roughly as 1/λ4, hence blue is scattered far more than red. It is the reason the sky is blue and the low Sun is red.
- Solar constant I0
- Mean extra-atmospheric irradiance on a plane facing the Sun. This page uses 1,367 W/m². The true value varies by a few percent through the year as the Earth–Sun distance changes.
- Direct normal irradiance (DNI)
- The beam from the solar disc, measured on a plane perpendicular to the rays, in W/m². Concentrating collectors and two-axis trackers live on DNI. A fixed rooftop lives on global irradiance in the plane of the array (beam plus diffuse).
- Pyrheliometer
- An instrument that measures DNI: a collimated detector on a solar tracker, seeing only the disc and a small circumsolar ring. A pyranometer, by contrast, sees the whole sky on a plane.