What satellite resolution do you need to spot a floating debris patch?

Ask anyone who's tried to scope a debris aggregation from orbit and they'll tell you the first instinct is wrong. You'd think you need to resolve the actual bottles and net fragments, so you reach for the sharpest imagery you can find. That's usually the wrong tool for the job of deciding where to point a vessel.

You're not photographing a bottle, you're finding a windrow

A single piece of floating plastic is too small for any spaceborne sensor to pick out on its own. What you're after is the aggregation: a windrow, a convergence line, a slick where wind and current have pushed debris into a band that can run tens to hundreds of meters long. That band has its own spectral signature. It reflects differently than open water in the near-infrared and shortwave bands, and that difference shows up as an anomaly even in imagery with a 10 to 30 meter ground sample distance (GSD).

This is the part that trips people up coming from a terrestrial remote-sensing background. On land, finer GSD almost always means a better product, because you're trying to classify individual objects or small parcels. On open water, you're trying to flag a few pixels of unusual reflectance against a huge, mostly featureless background. A windrow that's 50 meters wide and a few hundred meters long will register as a distinct patch of pixels at 10-30m GSD. You don't need to count the pieces inside it to know something's there and worth a closer look.

Why wide-swath beats sub-meter for scoping

Sub-meter commercial imagery (the kind tasked from SkySat- or WorldView-class sensors) can resolve individual rafts of debris beautifully, but the swath is narrow, often a few kilometers wide per pass. That's fine once you already know roughly where to look. It's a poor fit for the earlier question: where in a few hundred thousand square kilometers of ocean should a vessel even start.

Wide-swath multispectral satellites cover far more ocean per overpass, trading resolution for coverage. At 10-30m GSD, you're sacrificing the ability to tell a mat of mixed plastic from a patch of Sargassum at a glance, but you gain the ability to screen a large search area in a single pass and let the spectral signature do the flagging. For a planner trying to decide which quadrant deserves a vessel day, that trade is worth making. You send the narrow, expensive, high-resolution tasking after the wide pass has already told you where to look.

10m vs 30m: does the difference matter here?

Within that 10-30m band, the gap matters less than it does for, say, crop boundaries or coastal erosion mapping. A 10m pixel will catch a smaller or more diffuse windrow that a 30m pixel might blend into the surrounding water's noise floor. A 30m pixel still catches the larger, denser aggregations that tend to be the ones worth a vessel diversion in the first place. Neither resolution tells you what the debris is made of or how much of it there is. Both are reading a reflectance anomaly, not counting objects.

Cadence matters more here than people expect, too. Windrows move with wind and current on the order of hours to a couple of days. A satellite pass that's months apart on a fixed schedule will keep missing the window regardless of GSD. An event-driven read, one that fires when the spectral signature actually shows up rather than on a fixed revisit clock, does more for a scoping decision than shaving a few meters off resolution.

If you're trying to turn "something's probably out there" into a short list of coordinates worth a vessel diversion, Plastic Detection builds that read from wide-swath multispectral passes rather than asking you to parse raw imagery yourself. Join the early access list to see what an event-driven windrow alert looks like for your search area.