Experimental cartography
Hidden Waterfalls
Waterfalls the algorithm found in the Chapada's terrain — that nobody has catalogued yet.
The Chapada has hundreds of waterfalls. But how many still have no name? We ran a detector over the whole region's elevation data, looking for where rivers plunge. The result is this map of candidates — leads to explore, not finished destinations.
How it works
Where water flows
From the terrain, the detector simulates runoff and draws the drainage network — from creeks to rivers.
Where the ground drops
Along each river reach it measures the gradient. A waterfall isn't just steep ground: it's a river that falls abruptly.
The intersection
Where the two coincide — water plus an abrupt step — sits a knickpoint: the signature of a likely fall.
What the colour means
Each point's confidence comes from the most robust signals: the size of the basin (how much river passes there) and the channel gradient. High-confidence points are real rivers with sharp drops; low-confidence ones tend to be canyon-rim gullies — noise to filter out.
The terrain has a ~30 m resolution, so small falls or ones set in narrow slots slip through, and the height is an estimate. Treat the numbers as relative, not as a survey.
This is not an invitation to trespass
Many candidates fall on private land, Kalunga territory or areas with no trail and no signal. They are coordinates, not routes. Go with a local guide, respect access and nature, and never rely on GPS alone.
Scouted one?
If you've confirmed — or debunked — a candidate in the field, tell us. That's how a hidden fall becomes a waterfall on the Atlas.
Help us fix it