Notes ·

Field notes: the renewal where the swath edge cut the box in half

A dark storm front approaching over an open wheat field

Every renewal season the binders come back thicker. Wordings get redlined, exposures get re-mapped, and somebody on the risk side pulls up the original box, the lat/long rectangle the trigger lives in, and asks whether it still matches what's on the books.

A product lead we talked to this fall had just been through that exercise. The box for a wind cover had been redrawn to pick up a cluster of assets that moved onto the schedule since the last treaty year. On paper it looked like the easy kind of change: stretch one corner north, nothing exotic.

Cat-in-a-box is supposed to be the clean design. Draw the box, set the track or intensity threshold, and when the storm center crosses the line, the trigger fires. No argument about a ground station that lost power mid-event, no six-week wait on an official bulletin. Just geometry and a number.

The wrinkle didn't show up in the storm. It showed up in the imagery.

Wide-swath sensors cover a lot of ground in one pass, which is the whole point of them, but that ground comes in strips. The satellite images a band some tens of kilometers wide, flies on, and picks up the next band on a later orbit. Most of the time a box this size sits comfortably inside one strip and nobody thinks about it. This box, after the redraw, sat almost exactly on the seam between two strips. Half of it fell in the Tuesday pass. The other half didn't get picked up clean until the next revisit, almost a full cycle later.

Nobody caught it at binding. Why would they. The box looked fine on a map, four corners, reasonable size, well inside the region the provider advertises coverage for. It took someone pulling the archive during the renewal review, checking what a historical pass over that footprint looked like, to see the gap running diagonally through the middle of it. One side of the box had a crisp before/after pair from the same orbit. The other side's "after" image was from a different pass entirely, days later, which for a fast-moving peril is close to useless as a read of the footprint at the moment that mattered.

This is the kind of thing that never makes it into the trigger wording because the trigger wording describes the peril, not the sensor. Nobody writes "and the monitoring party shall ensure the box does not straddle a swath boundary" into a treaty. But it's exactly the detail that decides whether a payout stands on a clean, same-pass before/after read or on two images stitched together from different revisits with a gap between them that the counterparty's lawyer will find eventually.

The fix took five minutes once somebody looked. The box got nudged a few kilometers rather than reshaped, enough to sit inside one strip with margin either side. It's a renewal-desk fix, caught by checking the coverage geometry against the box before next storm season starts, while there's still time to change it.

A third-party read of the before/after footprint, triggered the moment the peril crosses threshold rather than stitched together after the fact, is the whole reason to run a cat-in-a-box this way, and it only works if the imagery underneath it covers the box in one piece. That's a swath-coverage question, and it belongs on the checklist before the geometry gets locked for the year, not during a claim dispute.

If your renewal box is due for a redraw, it's worth checking where the swath seams fall before the ink dries on another treaty year.

Start a pilot

Photo: Guillaume Boulanger / Pexels