Logging Road Detection

Blog·

How to verify a logging permit boundary against satellite imagery

A quiet dirt forest road curving through dense green trees.

A permit boundary on paper and a permit boundary on the ground are two different things until someone checks. The forest management plan gives you coordinates, a shapefile maybe, a hand-drawn sketch if you're dealing with an older operation. None of that tells you where the dozer actually went. That gap is where unpermitted access hides, and it's usually found weeks after the fact, during a routine field visit or a complaint call, long after the skid trail has hardened into a road.

Checking a harvest permit boundary against satellite imagery closes that gap faster than a site visit, and it's worth doing as a standing habit on active permits, not just when something looks wrong.

Get the boundary and the imagery on the same footing

Before you compare anything, make sure the permit boundary file and the imagery are talking about the same patch of ground. This sounds obvious and it's the step that trips people up most.

Pull the boundary as a shapefile or KML straight from the forest management plan, not a redrawn version someone traced off a PDF. Check the datum and projection. A boundary digitized in NAD83 and an image tile in WGS84 can shift by enough meters to put a legitimate harvest unit outside its own line, or hide real encroachment inside the error margin. If the permit boundary came from a GPS unit in the field rather than a surveyed corner, note that too. Consumer-grade GPS drift alone can run several meters, which matters when you're deciding whether a skid trail crossed a line or just ran close to it.

Once the boundary and the image are registered to each other, overlay the permitted cutting area polygon on the most recent high-resolution scene you have. Anything with a ground sample distance finer than half a meter will show skid trails and spur roads as distinct linear features, not a blur.

What a boundary violation looks like from above

Access shows up before the first tree comes down, as thin linear scars: a skid trail branching off the main haul road, a landing deck pushed into the buffer, a new spur that wasn't on the operator's approved road plan. Check for those before scanning for canopy loss, since the trees are usually the last thing to change.

Line these up against the permit boundary and look specifically at:

  • Trails or roads that cross the boundary line rather than running parallel to it inside the unit
  • New branching off an existing permitted road, since that's usually how access creeps outward
  • Activity inside a riparian setback or other no-cut buffer drawn into the forest management plan
  • Landing areas or turnarounds sitting outside the cutting unit polygon

A single pixel-width anomaly near the line isn't proof of anything. A branching trail pattern that didn't exist in the prior image and now sits partly outside the boundary is worth a field visit.

Match the image date to the permit timeline

The check is only as good as the date on the imagery. An image from before the permit was issued tells you nothing about compliance. An image from six months after the operator finished tells you about the aftermath, not the moment it happened. For a harvest permit boundary check to mean anything, you want imagery that brackets the active harvest window closely: something pulled shortly after permit issuance as a baseline, then follow-up passes through the duration of the cut.

This is also why relying on whatever archive imagery happens to be free and recent doesn't hold up for enforcement work. A scene from eight months ago won't show a road that went in last month, and by the time a gap in coverage gets filled, the clearing behind that road may already be done.

A monthly pass over active permit areas, with new linear features flagged against the boundary automatically, catches the road before the canopy loss does. That's the gap Logging Road Detection is built to close, and if boundary drift on your active permits is a recurring headache, it's worth a look at how the monthly scan works.

← Back to the blog