0 cmSurface
San Pedro de Atacama, Antofagasta
Everything below this line is measured, not described.
470 plant records logged within about eight kilometres.
San Pedro de Atacama
Antofagasta, Chile
The driest non-polar desert on Earth — and a reading that comes back higher than the Dust Bowl, which is the most interesting number on this site.
The ground under
San Pedro de Atacama, Antofagasta
-22.9087, -68.1997 · keep going ↓
5 cmTopsoil
10.0g/kg
grams of organic carbon per kilogram of soil
The living layer: dead leaves, roots, bacteria, and the fungal threads that bind them into crumbs. Almost everything that happens to soil carbon happens here first.
30 cmThe stock
39t/haLow
tonnes of organic carbon per hectare, top 30 cm
Typical of long-tilled arable land or disturbed ground, where organic matter has been oxidised away faster than plants could replace it.
ISRIC — World Soil Information
Modelled at 250-metre resolution from global soil surveys, and read from SoilGrids on 2026-09-01. SoilGrids publishes in versioned releases rather than continuously, so this figure is the current one until they publish again.
deeperThe network
1 fungal records near here
Each one is a specimen somebody found and wrote down. The species below are only the ones that fruit visibly; the threads that do the work almost never appear above ground at all.
GBIF — Global Biodiversity Information Facility
Each name links to the original record, with its collector, date and dataset.
Reading this figure
Treat this one as a reading about the model rather than about the desert. SoilGrids predicts from surveyed profiles, and the Atacama has very few of them, so the figure carries far more uncertainty here than it does over well-surveyed farmland. The number is reported exactly as returned; what it is worth is a separate question, and this is the page where that question is the point.
Why this ground
Ground at the limit
Soil that is frozen, or new, or so dry that almost nothing lives in it. The edges of the range explain the middle of it.
Some Atacama weather stations have gone years without measurable rain. Without water there is almost no plant growth, without plant growth there is almost no organic matter, and without organic matter there is very little of what makes soil different from crushed rock. It is used as a Mars analogue for exactly that reason.
So the figure on this page should be close to nothing, and it is not. It comes back in the same range as farmed European soil and well above the Oklahoma Panhandle — which is not a fact about the Atacama. It is a fact about what a global model does where it has almost nothing to learn from.
SoilGrids is built by training on soil profiles that have actually been dug and described, and then predicting everywhere else from satellite and terrain data. Where profiles are dense the prediction is anchored. Where a region is enormous, remote and barely surveyed, the model is extrapolating, and its output there is a best guess carrying an uncertainty the single number on the page does not show.
That is worth one page of this site on its own. Every figure here is modelled rather than sampled, and the honest version of "look up your field" includes knowing where the model is weakest.
Read this against
Or measure somewhere else