Conflict environmental forensics

Environmental forensics in and after armed conflict is an attribution problem with three separable parts: what the contaminant is, where the material came from, and whether either fact can be connected to a specific act. The analytical chemistry is mature. The evidentiary chain is where cases are usually lost.

Signatures that carry provenance

The useful measurements are those where a natural process writes a location into a material. Light stable isotopes do this through water: the ratios of oxygen-18 to oxygen-16 and deuterium to hydrogen in precipitation vary systematically with latitude, altitude and distance inland, a relationship first described by Craig in 1961, and that signal is carried into plant and animal tissue. Strontium works through geology instead: the strontium-87 to strontium-86 ratio of bedrock depends on its age and rubidium content, passes into soil water, and is taken up by plants and then by anything eating them, with almost no fractionation. Together they place timber, ivory or agricultural material within a region rather than merely identifying it. Trace-element profiles measured by ICP-MS add a soil-specific fingerprint, and lead isotope ratios can separate industrial pollution sources from geogenic background.

For material damage from weapons, some signatures are unambiguous. Depleted uranium is identifiable because its uranium-235 fraction is below the natural 0.72 per cent — an isotopic ratio no natural process in the environment produces. Residues of energetic materials such as TNT and RDX and their transformation products persist in soil, though they degrade on timescales that make the delay between event and sampling part of the result.

Identification of biological material

DNA barcoding answers a narrower question: what species is this. For animals the standard marker is the roughly 650-base-pair Folmer region of cytochrome c oxidase subunit I; plants use the plastid genes rbcL and matK instead, because plant mitochondrial DNA evolves too slowly to separate species; fungi use the ITS region. Seized and processed material is usually degraded, so shorter mini-barcodes are used, trading resolution for recoverable length.

Two limits bound all of it. The method identifies by comparison, so a specimen whose species is absent from the reference library returns its nearest listed relative rather than a null result — and reference coverage is heavily biased toward temperate, well-studied faunas. And recently diverged species, or species that hybridise, are not separable by a single marker at all.

Why admissibility, not sensitivity, is the constraint

A laboratory can measure a contaminant far below the concentration at which an attribution argument becomes possible, because attribution needs a comparison. Without a pre-conflict baseline, elevated concentration must be established against spatial controls, which requires access to uncontaminated ground of the same geology — often the thing that access restrictions prevent. Sampling in an insecure area rarely satisfies an unbroken chain of custody, and a sample whose handling cannot be documented is analytically sound and evidentially weak.

The legal threshold compounds this. Article 8(2)(b)(iv) of the Rome Statute treats environmental damage as a war crime only where it is widespread, long-term and severe — cumulative conditions, deliberately set high — so a defensible measurement of local contamination may still fall short of the standard it was collected to meet.