Skip to content

This is version 1.0, an earlier release, kept for the papers and datasets that cite it. The current release is version 1.1. When the paper is accepted, 1.1 takes its place here; this page and the Zenodo record stay, marked as superseded.

What drains into this pixel?

Finding what drains into a point normally means walking the network: step upstream, follow every branch, keep going until the branches run out. The cost grows with the catchment. Here it is a lookup. Every land pixel stores two numbers, dfs_in and dfs_out — the positions at which a depth-first walk of the flow network enters and leaves that pixel. Pixel p is upstream of pixel q when dfs_in(q) ≤ dfs_in(p) ≤ dfs_out(q) — one interval test, the same at a headwater and at the mouth of the Amazon. Nothing is traced.

Global map(unavailable)Example basin (pixel-level)
Global map(unavailable)Not available in a browser: this answer exists at every one of the 22.26 billion land pixels, far more than a web map can draw. The pixel-level rasters are in the Zenodo archive. A browsable version arrives at reach level with FullRiver.
Example basin (pixel-level)Basin 8595: click a pixel and the upstream catchment comes back from one interval test per pixel
Base layer
What to draw

Interval query · basin 8595

Click anywhere in the basin. The click snaps to the nearest pixel with a large drainage area, marked with a star, so a click on a hillslope still returns a catchment.

The formula below follows whatever you pick above.

Site under development — for browsing only, not for download yet. Some results are still being recomputed with refined methods, so the files here may still be replaced. Any changes to the results are small refinements, not major revisions. When this notice is gone, the files are final and ready to download. The version described in the paper is archived on Zenodo and does not change. Comments are welcome: lulu_jiang@pku.edu.cn