Cloud Cover in Satellite Imagery Archive Search

Cloud cover is the single filter that decides whether an archive search turns up a usable satellite image or a wasted afternoon. Most archive tools let you cap the maximum cloud cover on a scene, but that percentage describes the whole image footprint, not necessarily the small patch of ground you actually care about. A scene reported at 10% cloud cover can leave your site perfectly clear, or it can sit exactly under the one cloud that scene has. This guide walks through what cloud cover percentages actually mean, why scene-level cloud cover and cloud cover over your area of interest are two different numbers, and how to combine cloud limits with date range, resolution, and preview inspection to pull a genuinely usable image out of an archive.

What is cloud cover in satellite imagery?

Cloud cover is the percentage of a satellite scene's footprint that is obscured by clouds at the moment the sensor captured it. Every optical satellite image comes with this figure in its metadata alongside acquisition date, resolution, and sensor, and it is one of the first filters shown in almost any archive search interface. A scene with 2% cloud cover is nearly clear across its entire footprint; a scene at 80% is mostly unusable for visual analysis.

The number is calculated automatically from the scene itself, using onboard or ground-processing algorithms that classify each pixel as cloud, clear ground, or, on some platforms, cloud shadow. Optical and multispectral sensors both rely on visible and near-infrared light reflecting off the ground, so clouds sitting between the sensor and the surface simply block that signal the same way they would block a photograph taken from an airplane.

How common is cloud cover in satellite imagery, really?

Cloud cover is not a rare inconvenience, it is the default condition most optical satellites work around on every pass. NASA's long-run MODIS satellite data puts average global cloud coverage at around 67%, and even over open ocean, less than 10% of the sky is completely cloud-free at any given moment. That is exactly why archive search, rather than a single fixed capture, exists in the first place, an archive gives you dozens or hundreds of past passes to choose from instead of one shot at the weather.

None of this means an archive search is hopeless in cloudy regions. It means the search itself, stacking date range, cloud threshold, and sensor choice together, does more of the work than any single filter can on its own.

What does 0%, 10%, 20%, or 30% cloud cover actually mean?

Cloud cover percentages map roughly to how usable a scene will be, though the right number always depends on where the clouds happen to sit relative to your project.

Treat this table as a starting range, not a fixed rule. A 20% cloudy scene can be perfectly usable when the clouds sit in a corner of the footprint far from your site, and a nominally clear 5% scene can be useless if that 5% happens to sit directly over the building or field you are studying, which is exactly the distinction the next section covers.

Does 10% cloud cover mean 10% of my site is cloudy?

No, and this is the single most important thing to understand before filtering an archive by cloud cover. The percentage in a scene's metadata describes cloud coverage across the entire satellite footprint, which is often hundreds or thousands of square kilometers. Your actual area of interest, a mine, a single field, a construction site, might be a tiny fraction of that footprint, and the clouds could easily be concentrated somewhere else entirely, or sitting exactly on top of it.

This is why archive platforms that let you draw an area of interest and then inspect the scene preview, rather than trusting the scene-wide percentage alone, give a much more reliable answer. When a footprint is large relative to your AOI, always look at the actual preview over your polygon before ordering, not just the headline cloud figure.

Do you always need 0% cloud cover?

No. For most projects, chasing a perfectly clear 0% scene costs more than it is worth. Every archive search filter trades off against the others, and cloud cover is no exception, a lower maximum cloud threshold shrinks the pool of usable scenes and can push out the date range you actually needed, while a slightly higher threshold, paired with a preview check over your AOI, often turns up a perfectly usable image days or weeks closer to the date you actually wanted.

A practical rule most experienced buyers use is to start the search around a 10% to 20% maximum, inspect the previews of the top few results, and only tighten the filter further if none of them are actually clear over the AOI. That keeps the widest possible pool of dates and resolutions in play without wasting time on scenes that were never going to work.

Cloud cover is not the only problem, what about haze and cloud shadow?

A scene reporting 0% cloud cover is not automatically a perfect image. Cloud cover algorithms flag dense, opaque cloud, but they do not always catch every source of atmospheric interference that can still degrade a scene:

• Haze. A thin, hazy atmosphere softens contrast and muddies color without registering as cloud in the metadata.
• Thin cirrus. High, wispy cirrus cloud can be nearly transparent to the eye in a preview thumbnail while still degrading fine detail.
• Smoke and dust. Wildfire smoke or a dust storm can obscure a scene the same way cloud does, but from a completely different cause.
• Cloud shadow. Even where the cloud itself sits outside your AOI, the shadow it casts can fall across your site and darken it in the same pass.

None of this shows up reliably in a single cloud cover number, which is exactly why the scene preview matters as much as the filter itself. A quick visual check over your AOI catches haze, shadow, and thin cirrus that the metadata alone would never flag.

How to search an archive for low cloud cover satellite imagery

A reliable archive search stacks several filters together rather than relying on cloud cover alone. This is the order that consistently turns up the most usable results.

How much cloud cover is acceptable for your project?

Cloud tolerance is not one-size-fits-all, it depends heavily on how the imagery will be used and how large the AOI is relative to typical scene footprints.

What to do when there is no suitable cloud-free image in the archive

Sometimes every scene in the archive over your date range and AOI comes back too cloudy to use. A few adjustments, in order, usually solve this before a new capture is needed:

• Relax the cloud threshold. Move from 5% to 10%, then to 20%, and re-check the previews at each step rather than assuming a higher number is automatically unusable.
• Expand the date range. A two-week window that returns nothing can turn into several usable options once stretched to two or three months, especially outside a region's rainy season.
• Adjust resolution or off-nadir tolerance. If the analysis allows it, accepting a slightly wider off-nadir angle or a different sensor can unlock a pass that would otherwise be excluded.
• Consider a different sensor type. Switching part of the analysis to SAR, covered below, sidesteps the cloud problem entirely for the pass in question.
• Order new satellite tasking. When the archive genuinely has nothing suitable, tasking a new capture lets you specify an acceptable cloud cover threshold in advance rather than accepting whatever the archive happened to record. See our full archive versus tasking guide for how to decide between the two.

Archive imagery vs new satellite tasking

Archive search should almost always be the first move, since it is faster and cheaper whenever a clear enough scene already exists. New tasking becomes the right call once relaxing the cloud threshold, widening the date range, and adjusting resolution have all been tried and the archive still has nothing that meets the project's real requirements.

Does cloud cover matter for SAR satellite imagery?

Optical and multispectral sensors both depend on reflected visible or near-infrared light, so clouds block them directly. Synthetic aperture radar, or SAR, works differently, it sends its own microwave signal down and measures what bounces back, and that signal passes through cloud largely unaffected. That makes SAR a genuinely useful fallback for cloud-persistent regions and time-critical monitoring, including maritime and disaster response applications where radar's day-night, all-weather advantage matters most.

SAR is not a universal replacement for optical imagery, it measures surface texture and moisture rather than true color, and interpreting it takes different expertise. For a project already struggling to find a clear optical scene in the archive, adding a SAR option is often the fastest way to get usable data on the date that actually matters, rather than waiting out the weather.

Cloud cover is only one archive search filter

Cloud cover gets the most attention because it is the filter most likely to rule a scene out entirely, but a genuinely usable archive image depends on several factors working together.

A practical example, finding a low cloud image of a mining site

Consider a 20 km² mining operation that needs a satellite image captured between January and March, at 50 cm resolution or better, with the site itself confirmed clear of cloud. Rather than searching broadly, the workflow narrows fast, upload the exact lease boundary as the AOI, set the date range to the full January to March window rather than a single month, set resolution to 50 cm or finer, and set an initial maximum cloud cover around 20% rather than 0%. That search typically returns several candidate scenes. From there, open the preview on each one zoomed to the lease boundary specifically, discard any where cloud or shadow actually falls on the site regardless of the scene-wide percentage, check that the remaining candidates fully cover the boundary edge to edge, and compare price across whatever is left. In most cloud-affected regions, this produces at least one usable scene without ever needing to task a new capture.

Frequently asked questions

What is cloud cover in satellite imagery?

Cloud cover in satellite imagery is the percentage of a scene's total footprint obscured by cloud at the moment it was captured. It is calculated automatically and included in the scene's metadata alongside acquisition date, resolution, and sensor, and it is one of the primary filters used when searching an imagery archive.

What percentage of cloud cover is acceptable for satellite imagery?

A working range for most high-resolution optical projects is 10% to 20% maximum cloud cover, combined with a preview check over the specific area of interest. Mapping and construction monitoring often need closer to 0-10%, while agriculture, mining, and oil and gas projects can often work with 10-20% since the site itself may stay clear even in a cloudier scene.

Is 10% cloud cover good for satellite imagery?

Yes, in most cases. A scene at 10% cloud cover is generally usable, especially once the preview confirms the cloud is not sitting over your specific area of interest. The scene-wide percentage alone does not guarantee your exact site is clear, so checking the preview remains an important step even at this level.

Does cloud cover percentage apply to my whole area of interest?

No. Cloud cover percentage describes the entire satellite scene footprint, which is often far larger than a specific area of interest. A scene reported at 8% cloud cover could leave your site completely clear, or the cloud could sit directly over it, since the metadata figure does not distinguish between the two outcomes.

How do I find cloud-free satellite imagery in an archive?

Define a precise area of interest, widen the date range rather than searching a single narrow window, set a maximum cloud cover filter around 10-20% rather than 0%, and inspect the scene preview zoomed to your AOI before ordering. If nothing suitable turns up, relax the cloud threshold further or expand the date range before considering new satellite tasking.

Can I buy satellite imagery with 0% cloud cover?

Yes, fully cloud-free scenes exist in most archives, though they are less common in persistently cloudy regions such as tropical lowlands, where studies have found only a small fraction of scenes come back completely clear. Widening the date range significantly improves the odds of finding a genuinely 0% cloud cover scene.

Does SAR satellite imagery work through cloud cover?

Yes. Synthetic aperture radar sends its own microwave signal and measures the return, rather than relying on reflected sunlight, so cloud cover largely does not block it. This makes SAR a useful option for persistently cloudy regions or time-critical monitoring where an optical archive search cannot find a clear scene.

What should I do if there is no cloud-free imagery in the archive?

First relax the maximum cloud cover threshold and re-check previews, then widen the date range, then consider a different resolution or off-nadir tolerance, and consider SAR imagery as an alternative sensor type. If none of those produce a usable scene, new satellite tasking lets you specify an acceptable cloud cover threshold for a future capture.

Can I filter satellite imagery by cloud cover percentage?

Yes, a maximum cloud cover filter is a standard search option in most satellite imagery archive tools, alongside date range, resolution, and area of interest. Setting this filter narrows results to scenes at or below the chosen cloud cover percentage across the scene's footprint.

Sources and further reading

• NASA Earth Observatory, MODIS-based global cloud cover analysis, 2002-2015
• NASA ISCCP (International Satellite Cloud Climatology Project), cloud classification methodology
• Peer-reviewed Sentinel-2 cloud cover study, lowland tropical forest region, Brazil
• Peer-reviewed study on combined Sentinel-1 SAR and Sentinel-2 optical land-cover classification accuracy
• ESA Copernicus, Sentinel-1 SAR and Sentinel-2 multispectral mission specifications