Pick the Tree, Not the Ridge

A gobbler does not roost on a ridge. He roosts in one tree, and the gap between setting up eighty yards from that tree and eighty yards from a tree two hundred yards down the same bench is the gap between a bird on the ground in front of you and a bird pitching off the back side into the next hollow. Most hunters pick that spot from aerial imagery now. The trouble is that ordinary aerial imagery renders a spring canopy as texture rather than as individual trees, so the hunter is choosing a ridge when the bird is choosing a trunk. This paper is about ground sample distance: how much ground a single pixel covers, and what changes when that number falls from half a meter to three inches.

Open any mapping app on a hardwood drainage in early April and you will see a gray-brown corduroy of bare limbs. It reads as continuous. You can make out the creek, the field edges, the two-track, and the general run of the timber, and that is genuinely useful for the hundred-acre question of where to start.

It is not useful for the fifty-yard question. The bird is answering a much narrower one. Turkeys select roost trees with open, horizontal limb structure and clear flight access in and out, usually mature and often taller than their neighbors, frequently on a mid-slope bench or over a creek bottom where the tree stands slightly proud of the canopy around it. Those are properties of one tree. A basemap that resolves the stand does not resolve the tree, and a hunter working from it ends up making a reasonable guess about a ridge and then walking it at four in the morning hoping to be close.

Ground sample distance is the honest way to talk about this, and it is worth being precise because the marketing language around imagery rarely is. At 50 cm one pixel covers about a foot and a half of ground on a side, so a forty-foot crown lands on roughly two dozen pixels: enough to know a tree is there, not enough to reliably separate its edge from the crown of the tree growing into it. Two mature trees whose canopies touch become one shape.

At three inches the same crown is covered by about a hundred and sixty pixels on a side. The limb structure inside it registers, the gap between one crown and the next resolves as a gap instead of a smudge, and the emergent tree standing above its neighbors casts a shadow whose length and shape you can read against the slope. The step from one to the other is not a smooth improvement in prettiness. Object separation has a threshold, and the practical consequence is binary: either you can point at a single tree or you cannot.

At three inches a roost tree is an object you can point at. At half a meter it is a smudge in the timber that could be one tree or three.

The pixel arithmetic settles the argument on paper. The demonstration that makes it land is visual, and it is one we can run honestly on our own imagery rather than by staging a competitor's at its worst. Take one creek-bottom drainage at native three-to-six-inch resolution and degrade that identical image to half a meter. Same drainage, same capture, same sensor, nothing changed but pixel size, so any difference is resolution and nothing else: not capture date, not sun angle, not leaf state, not a vendor's rendering choices. The trees that were individual objects at native resolution dissolve into texture in the degraded copy.

FDR does not publish roost pins and will not start. Imagery of this kind narrows structure, not occupancy. It tells you which trees on a drainage have the physical characteristics turkeys select for, a materially shorter list than “the timber on this bench.” Whether a bird used one of them last night depends on hunting pressure, on where the hens are, on what the wind did the evening before, and on a dozen things no photograph records. The layer gives you three trees worth walking in on instead of a bench worth guessing at. It does not give you a bird.

Two other limits deserve saying plainly. Capture dates vary by county, and a late-season hunt in mid-May can fall well after the canopy has closed, so the bare-canopy advantage is a spring-season advantage that decays as the season runs. And a downed or dying tree looks a great deal like a good one from above; a crown that reads perfectly on the orthoimagery may turn out to be a snag when you get under it. The honest framing is the one we apply to deer bedding: terrain and structure tell you where to look, and cover and behavior decide where the animal is.