Notes ·

How many satellite passes confirm a parametric trigger event?

Sunset sky with scattered clouds over a weather radar station with a large dome antenna.

A parametric desk asked us this exact question last month, about to go to market with a wind cover that needed a settlement read inside 72 hours of landfall: how many satellite passes does it take to confirm a trigger event?

Pass count isn't really the right unit. What you need is one usable before/after pair over the exact peril footprint, inside your settlement window. Sometimes that's a single pass. Sometimes it's three, because the first two came back under cloud.

Why a clean window matters more than a pass count

A "pass" only counts if the sensor sees clean ground. An optical multispectral satellite flying straight over your footprint during a tropical storm, a wildfire smoke plume, or a post-quake dust haze will often return an image that's mostly cloud, mostly smoke, or mostly useless for a clean before/after comparison. That pass happened. It didn't confirm anything.

So when a cedant or a broker asks for a pass count in a contract, what they're really trying to buy is certainty about timing. The better spec is a commitment to deliver one qualifying read within a defined window after the trigger condition fires, not a raw orbit count. That's the distinction that matters when you're drafting the verification clause, not the revisit math.

Cloud cover is the real constraint

Most catastrophe perils that need a satellite read also come with weather that makes optical imaging hard: hurricanes, severe convective storms, flood events that follow days of rain. A single satellite on a fixed revisit schedule might not get a clear look at your footprint for a week or two after the event, by which point the payout clock has already run past what a cedant or a reinsurer will tolerate.

This is the part that doesn't show up in a basis-risk memo until the first real claim: access to more than one platform matters more than sensor sharpness. If the satellite over your footprint on day one is blocked by cloud, a second platform can still get a pass on day two or three, instead of waiting out a fixed orbital cycle. Wide-swath coverage helps here too. A sensor that images a broad strip rather than a narrow track is more likely to catch your footprint on whatever pass is clear, instead of missing it because the ground track didn't line up.

Event-driven cadence versus a fixed revisit schedule

Satellite programs built for routine mapping run on a fixed revisit cadence, the same patch of ground imaged on the same schedule whether or not anything happened there. That's fine for crop monitoring or land-use tracking. It's the wrong model for a parametric trigger, where the only pass that matters is the one that happens right after your index condition fires, and a scheduled cycle that doesn't care about your event timing will often miss the window entirely.

Event-driven tasking flips that around. The read gets requested the moment a qualifying event occurs, rather than waiting for whatever day the satellite happens to be scheduled to fly over. That's the cadence model a parametric settlement actually needs: not "every sixteen days," but "the moment the trigger condition is met."

What this means for the verification clause

If you're writing the imagery provision into a parametric contract, spec the outcome, not the mechanism. Define the peril footprint, the settlement window, and the requirement for an independent before/after read from an uninvolved third party, rather than naming a sensor, a revisit figure, or a pass count that a single satellite's orbit can't guarantee to hit. The contract should hold up whether the clear pass comes on attempt one or attempt three.

That's the gap Trigger Verification is built to close: an independent, event-triggered read of the peril footprint from wide-swath multispectral imagery, pulled the moment your qualifying event fires instead of waiting on whichever ground station or bulletin gets there first.

If your next contract needs that independent read written in before the next event hits, it's worth looking at how the trigger clause should be worded.

Start a pilot

Photo: Raychel Sanner / Pexels