Soil metals, and what lives in them

An open atlas of measured metal concentrations in soil and sediment, built from national and continental geochemical surveys, and joined to the microbial genera that occur in those soils.

Every claim on this site is shown with its evidence, and every source is described in full, including what it cannot answer.

Scope

81,331measured soil and sediment samples
16geochemical and pressure layers
3,543bacterial genera with trait and habitat data
14claims checked against the literature

Pages

Atlas

Element maps from national and continental surveys, with pressure, mining and sequencing layers that can be switched on over them.

Genera

Every genus with trait, habitat and metal response in one place, each attribute showing which source says what.

Analyses

The checks behind the claims: reproducibility between surveys, digestion bias, aggregation, and how much geology explains.

Data sources

A paragraph per source: what it measures, how it was sampled, and what it cannot answer.

What the analysis found

Regional metal patterns reproduce between independent surveys. Six of seven elements agree closely where two surveys cover the same ground; cadmium does not, and we report that rather than average it away.

Digestion method moves the level but not the ranking. Partial extractions return 18 to 22 percent less copper than total digestion at the same sites, so a cross-survey comparison needs about a one fifth allowance, while the order of sites survives.

Geology carries most of the regional signal. Across 298 regions, parent material explains roughly 40 to 55 percent of the ranking, and anthropogenic pressure terms add between 0.2 and 3 percent. Waste water load is the strongest pressure signal, and it is also a proxy for city size, which is stated wherever it is used.

Narrow habitat bands in survey data are detection ranges, not physiological limits. A genus found only between pH 4.1 and 5.2 in a survey has been found there, which is not the same as being unable to grow elsewhere.

Limits worth knowing before quoting a number