Soiling ratio, soiling rate and soiling loss: Definitions and formulas

Each term below has its own section, so you can link straight to it. The examples are illustrative.

Also defined below: soiling, soiling station, cementation and letter of conformity.

Soiling is the build-up of anything on module glass, or on a CSP mirror, that blocks or scatters sunlight: dust, pollen, bird droppings, soot, cement, lichen, or the scale that hard cleaning water leaves behind. Soiling cut global solar output by at least 3–4% in 2018 (Ilse et al. 2019), and IEA PVPS now describes an average, growing loss of 4–7% (2025).

The soiling ratio (SR) is the output of a soiled PV device divided by the output of a clean one, measured at the same time under the same sun (IEA PVPS 2022).

SR = soiled output ÷ clean output

• 1.00 means clean; the lower the ratio, the dirtier the device.
• “Output” can be maximum power, short-circuit current or an optical estimate, so say which. Typical uncertainty is 1–2% for power-based device pairs, 3–5% for current-based pairs and 4–7% for cell-to-cell and optical methods (Dunn et al. 2018, in IEA PVPS 2022).

Worked example. At noon, a soiling station’s cleaned reference module makes 300 W and its soiled twin makes 291 W. SR = 291 ÷ 300 = 0.97.

IEC 61724-1:2021 requires soiling measurement for Class A performance monitoring (IEA PVPS 2025).

Soiling loss is the share of output lost to soiling:

soiling loss = 1 − SR

In the example above, 1 − 0.97 = 0.03, a 3% soiling loss.

Over a day, a month or a year, soiling loss is the energy lost to soiling as a share of the energy the plant would have produced if clean. Weight it by energy, not by the clock: a dirty hour at noon costs more than a dirty hour at dawn.

Worked example. A 10 MWp plant near Upington producing about 19.9 GWh a year (Global Solar Atlas), with an average soiling loss of 3%, loses about 598 MWh a year. At the REIPPPP bid window 4 tariff of R1.10/kWh (IPP Office), that is worth about R0.66 million a year. The full table is in soiling losses in South Africa.

The soiling rate is how fast the soiling ratio falls in dry weather, usually expressed in % per day. IEA PVPS (2022) fits it as the slope of daily SR values over at least 14 days without cleaning or rain.

soiling rate = fall in SR ÷ number of dry days (as a fitted slope)

Worked example. Over 20 rain-free days, daily SR values fall along a line from 0.995 to 0.955. The soiling rate is (0.995 − 0.955) ÷ 20 = 0.002, or 0.2% a day.

Typical values depend on the setting: about 0.05% a day in the literature, 0.36% a day on heavy farmland and up to 0.5% a day near bird colonies (NREL 2018). South African measured rates are in soiling losses in South Africa.

From rate to loss. With a steady rate and a clean every T days, the loss climbs from zero to rate × T and averages about half that:

average loss ≈ soiling rate × T ÷ 2

At 0.2% a day and a 30-day interval, the loss peaks at 6% and averages 3%. The rate, not the loss, is what sets a cost-based cleaning interval: see how often to clean a solar plant.

Because the two use different bases. Loss is measured against clean output; gain is measured against soiled output.

gain after cleaning = clean output ÷ soiled output − 1 = 1 ÷ SR − 1

At SR = 0.97, the loss is 3.0% but the gain is 3.1%. The gap widens as soiling grows: at SR = 0.80, the loss is 20% and the gain is 25%.

A real example. In the peer-reviewed Green Energy Park trial in Morocco, by the end of the dry season the uncleaned string had produced 6.94% less cumulative energy than the water-cleaned string (Case study: Chemitek (Portugal), with Green Energy Park). That is a loss, measured from the clean side: an energy ratio of 1 − 0.0694 = 0.9306. Seen from the soiled side, the water-cleaned string produced 1 ÷ 0.9306 − 1 ≈ 7.5% more. Same data, two honest numbers. (This is our arithmetic on the published figure, for illustration.)

Whenever you quote a soiling figure, say whether it is a loss or a gain, and whether it is current, power or energy. See the Morocco trial.

On this page, cleaning effectiveness means the share of the soiling loss that a clean removes:

cleaning effectiveness = (SR after − SR before) ÷ (1 − SR before)

Worked example. SR is 0.94 before a clean and 0.99 after. The clean recovered 0.05 of the 0.06 lost: 83%. The remaining 1% is what the clean left behind, such as scale from hard water, cemented dust or streaks.

For reference, Ilse et al. (2019) list cleaning robots as removing more than 95% of soiling. Comparing effectiveness is how a pilot tells a good method from a merely wet one: the same crew, water and brush with and without a product, or the same product with and without softened water.

Not on its own. Performance ratio (PR) compares the energy a plant delivered with the energy expected from the measured irradiance and the plant’s rating. PR falls with soiling, but also with heat, clipping, outages, degradation and faults, so a drop in PR doesn’t tell you how much of it is dirt.

PR also depends on the irradiance sensor. IEC 61724 recommends cleaning irradiance sensors weekly, and IEA PVPS (2022) notes that temperature-corrected PR can give a good estimate of soiling in dry seasons, as long as the plane-of-array sensor is kept clean. A clean sensor over a dirty array folds soiling into PR along with every other loss. On an uncleaned plant in a dry spell, where little else changes, the PR trend can track soiling. For day-to-day decisions, a soiling station or optical sensor gives the soiling ratio directly.

The cleanliness factor is the CSP equivalent of the soiling ratio: the reflectance of a soiled mirror divided by its reflectance when clean, as set out in the SolarPACES Reflectance Guideline (v3.1, 2020).

cleanliness factor = soiled reflectance ÷ clean reflectance

The guideline distinguishes hemispherical reflectance (all reflected light) from specular reflectance (light reflected within a narrow cone). Portable reflectometers such as the D&S 15R measure at 660 nm and 15° incidence, with 15, 25 or 46 mrad apertures, and repeat to ±0.002.

Worked example. A mirror reads 0.940 clean and 0.905 soiled. Cleanliness factor = 0.905 ÷ 0.940 = 0.963.

What should the target be? At the Plataforma Solar de Almería in Spain, mirror soiling averaged 0.52% a day (up to 8.9% a day), and with fortnightly cleaning the cleanliness never fell below 0.82. Modelling put the cost-optimal average cleanliness at 0.965 ± 0.013, with 0.97–0.98 cited as optimal (Wolfertstetter et al. 2018). South African conditions can be harsher: test mirrors near a ferromanganese smelter at Emalahleni lost 32.6% of their reflectance per 14 days in the dry season (Swart et al. 2023). Why CSP soiling costs more than PV soiling is in CSP.

The rate of loss is what a treatment changes. At the SOLABEN CSP plant in Spain, mirrors treated with Chemitek Mirtek by brush lost 0.16% of their reflectivity in 10 days, against 2.17% for untreated mirrors: a loss rate 13 times slower (Case study: Chemitek (Portugal), with Atlantica and Rio Glass Services). See CSP reflectivity measurement, the SOLABEN case and Mirtek.

What is a soiling station? A soiling station is a pair of PV reference devices, modules or cells, mounted like the array. One is cleaned on a schedule and the other is left to soil, and the ratio of their outputs is the soiling ratio. IEA PVPS (2022) advises more than one sensor on plants above 5 MW that expect more than 2% soiling loss a year, because soiling can vary 1.5–2 times across a single site. How to measure PV soiling compares soiling stations with optical sensors and sets out where to put them.

What is cementation? Cementation turns loose dust into a bonded film. Dew dissolves the soluble part of the dust, which then recrystallises and binds particles to the glass in bonds that dry cleaning can’t break (IEA PVPS 2022). Cement dust is the extreme case. Ilse et al. (2019) note that cemented dust, like lichen and fungi, can become practically irremovable. See cement dust on solar panels.

What is a letter of conformity? A letter of conformity is a module maker’s written approval of a named cleaning product, or a cleaning method using it, on its modules. It usually comes with conditions, such as wet cleaning only, a pressure limit, a brush type and trained staff, and it is not a warranty guarantee. Chemitek Solar Wash Protect carries 18 such letters (Chemitek). See will cleaning chemicals void my panel warranty?