Constellation geometry and revisit design
Constellation geometry converts satellite count, orbital altitude and plane spacing into revisit time over a target. The relationship is not linear, and the diminishing-returns curve arrives earlier than most programme budgets expect.
The geometry before the hardware
Before a mission authority selects a sensor, a bus or a launch vehicle, it must answer one question: how often does the satellite see the target? Everything else flows from that answer. Revisit time is not a procurement line item; it is a geometric consequence of how many orbital planes you fly, how many satellites sit in each plane, at what altitude and at what inclination relative to your area of interest.
Two canonical patterns dominate the literature. A Walker Delta constellation distributes satellites uniformly across planes inclined at a common angle, optimising global or mid-latitude coverage. A Walker Star (sometimes called a polar Walker) uses planes that converge near the poles, producing denser coverage at high latitudes and sparser coverage near the equator. A string-of-pearls arrangement places all satellites in a single plane, which minimises launch cost and complexity but produces a single revisit opportunity per orbital period rather than the overlapping geometry that multiple planes provide. Each choice is a deliberate trade, not a default.
How satellite count converts to revisit time
At 500 km altitude, a single satellite in a near-polar orbit has an orbital period of roughly 94 minutes. It does not, however, revisit a mid-latitude point every 94 minutes. Earth rotates beneath it, and the ground track shifts westward by approximately 23 degrees per orbit. A single satellite in such an orbit may revisit a specific point only once or twice per day, depending on the sensor's swath width.
Add a second satellite in the same plane and you halve the revisit interval within that plane, but you gain nothing in cross-track coverage. Add a second plane, phased correctly, and you begin to fill the gaps in longitude. This is the core logic of Walker notation: T/P/F, where T is total satellites, P is the number of planes and F is the relative phasing between planes. A 12/3/1 Walker Delta, for instance, places four satellites in each of three planes with a one-slot phasing offset. At 550 km and 53-degree inclination, a constellation of that size can achieve sub-two-hour revisit at mid-latitudes with a modest wide-area optical swath.
Planet Labs' Dove constellation, which has operated with more than 130 satellites in a near-polar string-of-pearls and distributed-plane arrangement, demonstrates daily global coverage at 3 to 5 metre resolution. ICEYE's SAR constellation, with satellites distributed across multiple planes, achieves revisit intervals of a few hours over specific targets. These are publicly documented benchmarks, not theoretical limits.
The diminishing-returns curve arrives fast
Doubling satellites from two to four produces a dramatic revisit improvement. Doubling from sixteen to thirty-two produces a much smaller one. The reason is geometric saturation: once your ground track density exceeds the sensor swath width, additional satellites in the same plane stop contributing useful coverage. You are paying for orbital slots and launch mass that overlap with existing coverage rather than filling gaps.
The inflection point depends on altitude and swath. A wide-area radar with a 300 km swath reaches diminishing returns at a lower satellite count than a high-resolution optical instrument with a 15 km swath. A programme targeting a single country or a defined maritime exclusive economic zone reaches saturation far sooner than one designed for global persistence. Designing to a specific latitude band rather than the whole globe is often the most cost-efficient decision a national programme can make, and it is one that Walker geometry makes straightforward to quantify before procurement begins.
Designing for latitude: the honest maths
Coverage is not uniform across a constellation's latitude range. A Walker Delta at 53-degree inclination provides excellent mid-latitude revisit but thins out above 60 degrees north or south. A Walker Star at 90-degree inclination provides polar coverage but produces equatorial gaps during certain phasing windows. For a government whose territory sits between 20 and 40 degrees latitude, a high-inclination Walker Star wastes capacity; a Walker Delta at 45 to 55 degrees inclination is more efficient.
Altitude matters too. Higher orbits increase the sensor's instantaneous field of view, which reduces the satellite count needed for a given swath, but they increase the distance to target, which penalises resolution and signal strength in equal measure. Most Earth-observation constellations settle between 400 and 600 km because this band balances atmospheric drag (which shortens mission life below 400 km without propulsion), coverage geometry and sensor performance. The ISS at 400 km and Sentinel-2 at 786 km bracket the practical range for government-class optical missions.
Where geometry fails you
Constellation geometry is a necessary analysis, not a sufficient one. It models the satellite's position relative to the target; it says nothing about whether the sensor can actually collect. Cloud cover over tropical regions can render an optical pass useless 60 to 80 percent of the time in monsoon season. A geometrically perfect revisit interval means little if seven consecutive passes are obscured. SAR sensors are not blocked by cloud, but they introduce their own ambiguities in interpretation, and their data volumes stress ground station downlink budgets in ways that optical missions do not.
Phasing assumptions also degrade over time. Satellites in the same plane drift relative to one another due to differential drag from atmospheric density variations, particularly at altitudes below 500 km. Without active station-keeping, a carefully phased Walker constellation can lose its geometry within months. Propulsion mass, fuel budget and the operations cost of regular manoeuvres must all be priced into the programme before the geometry analysis is treated as settled.
Finally, a constellation is only as useful as its ground contact allows. Revisit time tells you how often the satellite sees the target. Latency, the time between collection and delivery of usable data, depends on downlink frequency, ground station network geometry and processing pipeline. A satellite with a four-hour revisit that downlinks once per day is not a four-hour revisit system in any operationally meaningful sense.
From geometry to programme decision
The practical output of a constellation geometry study is not a single number. It is a trade space: a set of curves showing revisit time against satellite count, at several altitudes, for the specific latitude band the mission must serve. That trade space then intersects with payload swath, launch vehicle fairing constraints and ground station coverage to produce a minimum viable constellation size.
For most sovereign programmes, the answer is smaller than the sponsor initially expects. A three- to six-satellite constellation at 500 to 550 km, properly phased across two or three planes, can achieve sub-six-hour revisit over a defined national territory with a 100 to 200 km swath instrument. That is a credible foundation for maritime patrol, agricultural monitoring or disaster response. Scaling beyond that requires a clear operational requirement driving the additional revisit, not an aspiration.
Engineering parameters
| Typical LEO constellation altitude range | 400 to 600 km (below 400 km requires active drag compensation; above 600 km increases debris persistence) |
| Orbital period at 500 km | ~94 minutes |
| Ground track shift per orbit (near-polar) | ~23 degrees longitude westward per orbit at 500 km |
| Walker notation | T/P/F: total satellites / planes / phasing parameter; e.g. 12/3/1 |
| Minimum planes for non-string-of-pearls revisit benefit | 2 planes; meaningful geometry improvement typically requires 3 or more |
| Revisit time, single satellite, 500 km, 15 km swath | 3 to 5 days at mid-latitudes (geometry only, cloud not included) |
| Revisit time, 6-satellite Walker Delta, 500 km, 100 km swath | 4 to 8 hours at target latitude band (published analogues: ICEYE, Capella) |
| Diminishing-returns inflection (typical wide-area optical, 200 km swath) | Approximately 12 to 18 satellites for near-continuous regional coverage |
| Station-keeping propulsion requirement (below 500 km) | Required for phasing maintenance beyond 6 to 12 months; delta-v budget mission-specific |
| Latency vs revisit distinction | Revisit is geometric; latency adds downlink schedule and processing time, often 2 to 24 hours in practice |
One contract, one accountable engineer
Commissioned as one programme, not a stack of contracts: spacecraft, launch, ground segment, mission control, training and handover are priced together. Source-access terms and audit rights are agreed in writing before signature. Request a revisit trade study.