TerrAlert Team · 2026-09-09
"Real-time" is one of the most overused words in satellite monitoring marketing, and it is worth being precise about what is actually possible, because the honest answer changes what you should expect and plan around.
An optical satellite can only photograph a given point on Earth when its orbit brings it overhead. For a system like Sentinel-2, that happens roughly every two to three days at most latitudes. No amount of software improves this; it is fixed by orbital mechanics. Once captured, an image needs to be downlinked to a ground station and processed before it is usable, which typically adds somewhere between half a day and a full day.
Then there is cloud cover. A pass that happens over an overcast sky produces an image that simply cannot see the ground underneath. The data exists, but it is unusable for change detection. Across a typical monitored area, cloud cover affects a substantial share of passes, which means the effective gap between two clear, usable images is longer than the raw revisit schedule suggests, often averaging somewhere around five days, depending on the region and season.
Put together, revisit schedule, processing delay, and cloud loss, a realistic expectation for optical satellite-based detection is "usually within about a week of a real change," not minutes or hours, and not even reliably "daily," even when a product's plan name says "daily."
When a monitoring product's plan is named "daily" or "realtime," a careful reading of what that promises matters. It typically describes how often the system checks for new imagery, not how often new, usable imagery actually exists to check. If nothing new and clear has arrived since the last check, a system checking constantly has nothing new to report, correctly. This is not a flaw in the product; it is a fact about how often nature actually provides a new, usable observation.
Radar-based satellites (Sentinel-1, notably) see through cloud entirely, because they use their own signal rather than reflected sunlight. This removes the cloud-loss part of the delay, leaving only the orbital revisit schedule and processing time, which can mean detection within hours for specific kinds of change that radar is well suited to identify, like new construction, rather than the days-long wait typical of optical-only monitoring. The trade-off is that radar data is harder to interpret directly and is generally used for specific, narrower categories of change rather than general-purpose visual inspection.
Is any satellite monitoring truly instant? Not for general-purpose optical monitoring, no. Radar can get much closer to fast detection for specific use cases, but "instant" worldwide monitoring regardless of conditions is not something current public satellite programmes support.
Why do some products still advertise 'real-time'? Often because it describes their checking frequency or delivery latency once a detection exists, not the time from a real-world change to that detection existing in the first place. Read the fine print.
TerrAlert's own measured numbers, including the honest, occasionally unflattering ones, are on the FAQ.
All posts