Soiling is the gradual loss of solar output that happens when dust, pollen, salt, or other grime builds up on your panels and blocks some of the sunlight. It’s a normal part of running a photovoltaic (PV) system, especially through a long dry season. The good news is that soiling is recoverable: a good rain, or a cleaning crew, washes the panels and output returns to normal. That’s what sets it apart from panel aging (degradation), the slow permanent decline that never comes back.
Gridium measures how much energy soiling is costing your system and reports it as a percentage of production. That tells you when it’s worth scheduling a cleaning, and whether the last one paid off.
How Gridium detects soiling
Gridium compares each day’s actual production against the production expected for that day’s weather. Over time, soiling leaves a signature that’s hard to miss:
- A slow, steady slide downward, day after day, as dust accumulates.
- A sudden jump back up, when rain or a cleaning washes the panels.
That sawtooth pattern (gradual decline, sharp recovery, repeat) is the fingerprint of soiling. It’s what separates soiling from a passing cloudy spell, which comes and goes, and from an equipment fault, which drops output and keeps it down until someone makes a repair.
Best of all, the method needs no rain gauge, no humidity sensor, and no cleaning log. It reads the entire soiling story from the shape of your system’s own production record.
Actual production drifts below expected through the dry season as dust builds up on the panels, then returns to expected after the first rain washes them clean.
An industry-standard method
The technique behind these numbers is called Stochastic Rate and Recovery (SRR). It was developed by the U.S. National Renewable Energy Laboratory (NREL), published in peer-reviewed research, and released in NREL’s open-source analysis toolkit that’s used across the solar industry. Using the recognized industry method means your soiling numbers are produced the same way a specialist consultant or a manufacturer’s own engineers would produce them.
There’s real uncertainty built into the problem: the exact date of each cleaning, and exactly how fast dust was building up, can never be known for sure. Rather than force a single guess, SRR weighs many plausible versions of the story, each with different cleaning dates and accumulation rates, all consistent with what your system actually produced, and reports the typical result across them. Gridium runs this repeatedly and takes the consensus, so one unlucky run can’t skew the answer.
Where soiling detection works best
Soiling detection is most reliable where there’s a clear rhythm of dry spells and rain, with plenty of sunny days to measure against. Much of California and the U.S. Southwest is ideal. It also needs a few years of data to be dependable; when the record is too short, Gridium reports no soiling figure rather than an unreliable one.
A couple of climates are exceptions:
- Constant rain: Where it rains almost year-round, dust never builds up enough to measure, so the answer is usually near zero.
- Snow: In snowy climates, snow cover has to be handled separately, because it looks like soiling on the chart but isn’t.
Soiling vs. aging
A slow soiling decline can look, at a glance, a lot like the slow permanent decline of panels as they age. Gridium keeps the two separate on purpose, using a second industry-standard method (again from NREL) to measure long-term aging on its own. That method compares each day to the same day a year later, so seasons and soiling cancel out and only the true aging trend is left. Soiling is the part that comes back after rain, and aging is the small part that never does.
Helpful Hint: Watch how deep the sawtooth dips before each recovery. The wider the gap between actual and expected production, the more output soiling is quietly costing you. If your dry season runs long, a mid-season cleaning can recover that production well before the first rain does the job for free.