Soiling losses in South Africa: How much dust really costs a PV plant

What is soiling loss? Soiling is anything that settles on module glass and blocks light: dust, pollen, bird droppings, cement, soot, lichen, or the scale that hard cleaning water leaves behind. Soiling loss is the share of a plant’s energy lost because of it.

Engineers measure it with the soiling ratio. The definitions and formulas are in our soiling ratio guide. Unlike the weather, soiling is a loss you can manage.

Ilse et al. (Joule, 2019) estimated that soiling cut global solar output by at least 3–4% in 2018, worth at least €3–5 billion a year, even with optimised cleaning. The IEA PVPS fact sheet of September 2025 now describes an average global loss of 4–7%, and growing.

Averages hide a wide range. The site sets the number:

Published South African soiling data is thin, but it points one way: the dry season drives the loss, rain resets it, and patchy soiling does more harm than its size suggests.

Three lessons stand out. First, the dry season sets the bill: the Pretoria and Kalkbult losses both built up over rain-free winter months. Second, geometry matters: at the same Pretoria site, single-axis trackers lost three times as much per week as dual-axis ones. Third, patchy soiling creates hot spots; see bird droppings.

There is also a gap. We found no peer-reviewed PV soiling rates for Upington, Kathu, Prieska, Kimberley, Namibia or Botswana. If your plant is there, your own measurements are the only data you will have.

• Long dry spells. South Africa averages about 490 mm of rain a year, about half the world average (WWF-SA), and parts of the Northern Cape get less than 70 mm (SAEON). Between rains, only a clean resets soiling.
• Regional dust sources. A study of dust events from 2005 to 2008 traced 328 plumes to 101 sources, mostly dry lakes, pans and riverbeds in Namibia, Botswana and South Africa, such as Etosha and the Makgadikgadi (Vickery et al. 2013).
• Industry and mining. Industrial and mine dust can be far more damaging than natural dust. Cement cures onto glass, CSP mirrors near an Emalahleni smelter lost 32.6% of their reflectance per 14 dry-season days (Swart et al. 2023), and in an Indian study 5.98 g/m² of iron-ore mine dust cut maximum power by 63.5% (Tripathi et al. 2026).
• Farmland and birds. NREL’s figures put heavy farmland dust at 0.36% a day and sites near bird colonies at up to 0.5% a day, many times the typical 0.05%.
• Dew. Dew dissolves the soluble part of dust, which recrystallises and cements particles to the glass. Cemented dust doesn’t come off with dry brushing (IEA PVPS 2022). See cement dust on solar panels.
• Scarce water. Many plants clean rarely. In a survey of 11 South African IPPs, four never washed their modules and waited for rain instead (Naicker 2018).

This is an illustrative calculation. Change any assumption and the answer moves.

Assumptions

• Plant: 10 MWp, fixed tilt, near Upington.
• Specific yield: about 1,993 kWh per kWp a year (Global Solar Atlas), so about 19.9 GWh a year.
• Tariffs: REIPPPP PV tariffs of R4.39/kWh (bid window 1), R1.10/kWh (BW4) and R0.49/kWh (BW6). BW1 and BW4 are in April-2020 rands and partly CPI-indexed; BW6 is as announced (IPP Office; gov.za).
• Soiling loss: an annual average of 1%, 3% or 5% of energy.

Method: energy lost = annual yield × average soiling loss; value = energy lost × tariff.

Where 3% and 5% come from. At a steady soiling rate, the loss climbs in a straight line from zero after each clean, so the average over a cleaning cycle is about half the peak. The CSIR Pretoria measurements imply roughly 0.17–0.26% a day. At 0.2% a day with a clean every 30 days, the loss peaks near 6% and averages about 3%. Stretch the interval to 50 days and the average reaches about 5%.

Two cautions. The straight-line model ignores rain, which can reset soiling, and cementation, which can hold soiling in place until a chemical clean. And for C&I plants, use the tariff you avoid rather than a REIPPPP tariff.

Run the numbers for your own plant in the soiling loss calculator: enter the plant size, specific yield, tariff and soiling rate.

Does cleaning more often always pay? No. Each clean costs labour, water and equipment time, and the returns fall off quickly. In southern Spain, one well-timed summer clean raised a 1 MW plant’s profit by up to 3.6%, while extra cleans paid less (Micheli et al. 2021).

The right interval depends on your tariff, soiling rate and what a clean costs, and it changes with the season. The calculator works it out for your plant. Our cleaning frequency guide sets out how to use the result.

A survey of 11 South African solar IPPs (Naicker 2018, hosted by SAPVIA) found:

Compare that with LONGi’s O&M manual, which asks for cleaning at least every 20–30 days in dry areas and every 40–50 days in rainy ones. Neither habit is right for every plant. The answer depends on each plant’s soiling rate, tariff and cost per clean, and those have to be measured.

Water is part of the cost. A wet clean uses about 1 L of water per m² of module in US literature, and a South African dissertation cites about 3 L/m² (NREL 2018; Naicker 2018). For a 10 MWp plant, that is roughly 50–150 m³ per full wash.

How do you find your own plant’s soiling loss? Measure it. The definitions, Class A requirement and how many sensors a site needs are in soiling ratio, rate and loss explained and soiling monitoring. How to measure PV soiling compares the methods and their uncertainty.

A controlled field test shows what that measurement looks like: see the Morocco trial. Seasons matter as much as the site: coated strings on a 330 MW plant in Vietnam gained more in the dry season than in the rains.

First, work out which kind of deposit you have. In practice they fall into two groups, and each needs a different fix:

• Soiling: dust, sand and pollen that build up between cleans, on almost every plant. Solar Wash Protect or Antistatic Solar Armor 2.0 in the normal clean is all it needs. In Chemitek’s case studies, Solar Wash Protect added energy over water-only cleaning at a utility plant in Honduras and 5.1% on a rooftop in India.
• Contamination: cement, limescale from hard water, mineral deposits, iron and metal oxides or lichen, bonded to the glass. It keeps building until it is removed, so the losses are larger and these plants usually gain the most. See the Dubai case.

Then start with the number and choose the fix:

1. Measure the soiling rate and identify the contaminant: soiling monitoring and a site soiling assessment. For very large plants, a plant performance study adds SCADA analysis and cleaning economics.
2. Restore what is bonded. Cemented and biological deposits need the right chemistry, not harder scrubbing, because abrasive methods damage the anti-reflective coating.
3. Protect where it pays. Short tests of anti-soiling coatings show more than 80% less soiling, but long-term field results are typically 20–50%, and some coatings make soiling worse (Ilse et al. 2019). In a Northern Cape trial, a hydrophobic coating increased dust soiling, opening a performance-ratio gap of up to 5.5% (du Plessis et al. 2020). Prove any coating on your own strings first.
4. Maintain on a measured interval, with water inside your module maker’s limits. See water quality for solar panel cleaning.

This is the Restore → Protect → Maintain programme we design for each plant. For utility-scale and large C&I plants, we test it on your own strings first in a free pilot.