Optical or SAR?
Optical sensors produce imagery people can read without training; radar produces data that ignores cloud and darkness. Neither replaces the other, and the sequencing question matters more than the either-or question most tenders ask.
What optical does that radar cannot
Optical imagery is legible. A minister, a judge or a journalist can look at a sub-metre optical scene and understand it without a briefing. Colour, texture and shadow carry information that radar backscatter simply does not, and for identification tasks, telling one vessel class from another, reading markings, assessing damage, optical remains the standard.
Hyperspectral extends the same principle: hundreds of narrow bands let materials identify themselves, so crop species, mineral outcrops and oil films become classifiable rather than merely visible.
What SAR does that optical cannot
Synthetic aperture radar brings its own illumination, so it images at night and straight through cloud. Over territory with a real monsoon season, that is not a nice-to-have; it is the difference between a monitoring capability and a fair-weather one. Some coastlines sit under cloud for weeks at a stretch, and an optical-only program simply goes blind there.
SAR also measures things optical cannot see at all. Interferometric processing compares radar phase between passes and detects ground movement at millimetre scale: subsiding dams, creeping embankments, settling railways. And because ships cannot switch off their radar reflection the way they switch off transponders, SAR is the workhorse of maritime surveillance.
Where each one fails
SAR data needs interpretation. Backscatter images are not photographs, and building the analytical bench to exploit them takes deliberate training. Optical fails more simply: no light, no picture, and a target that knows your pass schedule can arrange to be interesting only after dark.
Cost cuts both ways. Small SAR satellites have fallen in price sharply in recent years, which removed the old excuse for treating radar as a luxury layer. But radar's downlink volumes and processing demands put real weight on the ground segment, and that has to be sized in from the start, not discovered after launch.
The question tenders should ask instead
Not which sensor, but which decisions. List the calls your government keeps making blind: where the floods are, what moved on the border overnight, which vessels went dark. Map each decision to the sensing it actually needs and the choice largely makes itself. Border change detection in all weather wants radar. Identification and evidence want optical. A serious national program ends up with both; the real decision is which flies first.
Our usual answer on sequencing: fly the sensor that addresses the decision that hurts most today, and design the ground segment for both from day one, because antennas and processing are cheaper to size correctly than to rebuild.
The short version
This is a sequencing decision disguised as an either-or. Pick the first sensor by the decision it unblocks, plan the second from the outset, and never buy an optical-only program for territory that spends half the year under cloud.