Notes from the overnight pass: why cadence beats resolution

There's a particular kind of frustration that comes from pulling a gorgeous 25cm pass of a dispersal field and realizing it's eleven days stale. The jets are crisp. The shadows are long and clean. You can practically count rivets. And none of it tells you whether the squadron rotated out last Tuesday, because the only other image you've got is from three weeks before that.
We talk resolution constantly in this trade: sub-meter, sub-half-meter, the provider that just announced a new sensor generation. Procurement conversations get built around GSD like it's the only number that matters. The actual question sitting on most desks is what changed since I last looked, and that depends entirely on how often you look. It's a cadence problem wearing a sensor problem's clothes.
Think about what you're actually trying to catch at a named airfield. Aircraft on apron get repositioned within a day, sometimes within hours of a tasking order. Revetment occupancy shifts with deployment cycles that don't announce themselves. A shelter that was a foundation pour in July is a roofed structure by September, and the interesting part isn't the finished building, it's the week construction accelerated or stalled. None of that shows up in a single sharp image. It shows up in the delta between two images, and the delta is only useful if the gap between them is short enough to isolate what moved.
A field I watched for a stretch last year made this obvious in an annoying way. Good imagery, well-tasked, genuinely nice archive. But the revisit sat somewhere between ten days and three weeks depending on weather and tasking priority elsewhere. Every time a new collect came in, I'd spend the first twenty minutes just reconstructing what had happened in the gap instead of confirming it.
Had the transport aircraft been there the whole time, or did they leave and come back? Was the revetment empty because of a sortie or because of a permanent drawdown? With a ten-day gap you genuinely cannot tell the difference between "they moved twice" and "they moved once and I missed it." The imagery was fine. The spacing between looks was the problem.
Compare that to a lower-drama field with daily or near-daily coverage, even at modest resolution. You don't need to count rivets when you can see the apron went from four aircraft to zero and back to six over three consecutive days. The story writes itself because you're not trying to infer a week of activity from two snapshots, you're watching it unfold frame by frame. Cadence turns a guessing exercise into an observation log.
This is also, bluntly, why so much analyst time disappears into archive paging. If your only tool against a sparse revisit schedule is to manually scroll back through tiles counting aircraft on apron frame by frame until you spot where the number jumped, you're doing change detection by hand, one eyeball-pass at a time, for every field on your plate. It works. It's also how a Tuesday afternoon vanishes.
Checking in on a field daily, instead of only when a tasking happens to land there, is what catches the actual moment of change. A running log that tracks apron and revetment counts day over day, with the comparison view sitting right next to it, builds that habit into the workflow: it shows what moved between yesterday and today.
Next time a new pass comes in, resist the urge to zoom first. Check the date of the last one instead. That gap is usually the more honest number.