Population Genetics, Geographic Provenance, and Stable Isotopes

Gene flow between elephant populations is constrained by rivers, mountains, farmland, and national park boundaries. Where populations are isolated long enough, their allele frequencies drift apart. The result is a gradient of genetic differentiation that correlates with geography. Populations in East Africa cluster separately from those in West Africa; Southern African populations form a third cluster; Central African forest elephants form yet another. This structure is the foundation of genetic provenance assignment.

Samuel Wasser's team at the University of Washington published a geographic assignment tool for elephant ivory using microsatellite and later SNP markers. The approach builds a map of allele frequencies from reference samples of known origin, then calculates the likelihood that an unknown sample's genotype was drawn from each map cell. The result is a probability surface across Africa rather than a single point, which is the correct way to present geographic uncertainty in court.

STRUCTURE (Pritchard et al., 2000) and ADMIXTURE (Alexander et al., 2009) take as input a matrix of genotypes: alleles at multiple loci across multiple individuals: and output membership coefficients: the fraction of each individual's genome that traces to each of K hypothetical clusters. Running the analysis at increasing values of K and selecting the one that best fits the data (using delta-K or cross-validation error) reveals the natural number of genetically distinct groups in the dataset.

For a seized specimen the workflow is: extract DNA, generate the same marker panel as the reference dataset, and project the unknown individual into the ADMIXTURE or STRUCTURE framework. The membership coefficient for each cluster is the assignment probability. A tusk that shows 87% membership in the East African cluster and 13% in the Southern African cluster is assigned to East Africa with 87% confidence: a quantified, auditable statistic that a court can evaluate.

Strontium has four naturally occurring isotopes. The ratio of Sr-87 to Sr-86 varies across Earth's surface because Sr-87 is a decay product of rubidium-87, and rubidium concentrations differ by rock type. Old granitic cratons carry high Sr-87/86 ratios; younger volcanic basalts carry lower ones. Plants take up strontium from soil water in proportion to local ratios; elephants drink local water and eat local vegetation, so their ivory records the ratio of the landscapes they inhabited.

John Vogel and collaborators mapped Sr-87/86 ratios across sub-Saharan Africa by measuring bones and teeth of animals of known location, creating an isoscape: a spatial map of expected values. Measuring the ratio in a section of ivory and overlaying it on the isoscape produces a probability map of origin locations. Vogel's work showed that strontium alone could exclude large portions of Africa and in some cases narrow provenance to a region of a few hundred kilometres.

Stable carbon isotopes (delta-C13) in organic tissue reflect the photosynthetic pathway of the plants the animal ate. C4 grasses, which dominate open savanna, carry less negative delta-C13 values than C3 forest vegetation. An elephant that lived primarily in open savanna will show less negative ivory delta-C13 than one from dense forest. This separates savanna from forest habitats with a clean biochemical signal independent of the genetic data.

Nitrogen isotopes (delta-N15) increase up the food chain and also reflect aridity: arid-zone plants and soils have higher N-15 enrichment than humid ones, partly through ammonia volatilisation from dry soils. For elephants, which are herbivores, N-15 mainly tracks aridity and soil fertility rather than trophic level. Combining C13 and N15 with Sr-87/86 produces a three-isotope signature that, when mapped against an isoscape calibrated for all three, can constrain origin more tightly than any one isotope alone.

White rhino (Ceratotherium simum simum) populations in South Africa and Zimbabwe show clear SNP-level genetic differentiation, reflecting their management history as semi-isolated populations within national parks. This makes population assignment feasible at the country level, which is directly relevant to legal jurisdiction. A horn that assigns to the South African population implicates South African protected areas; one that assigns to Zimbabwe points to a different national authority and different trafficking route.

Rhino horn is keratin, not bone, and it grows continuously. This means isotope sampling along the horn length records a time-series just as tusk sampling does. Stable isotopes in rhino horn keratin have been used to reconstruct the animal's movement between landscapes and to estimate how recently it was killed: important when prosecutors need to show that a horn is from a recently poached animal rather than a legal stockpile.

• RhODIS (Rhino DNA Index System): South Africa's national rhino DNA database, holding STR profiles from living animals, dehorned animals, and carcasses. It has been used to match seized horn to specific animals in court proceedings.
• Population assignment limitation: rhino translocation between countries for conservation purposes can blur the population genetic signal. An animal translocated from South Africa to Kenya will carry South African allele frequencies despite living in Kenya: a complication that should be addressed in the report.

The most powerful provenance reports combine genetic assignment with isotope data on the same sample. The two methods are orthogonal: population genetics reads inherited sequence variation accumulated over thousands of generations, while stable isotopes read dietary and environmental chemistry incorporated during the individual's lifetime. When they converge on the same origin region, the joint probability of both patterns arising by coincidence from a different location is very small.

Researchers working on large ivory seizures, particularly the cargo analyses published by Wasser's group between 2007 and 2015 examining hundreds of tusks from major seizures, found that genetic assignment alone could identify the primary source populations and that isotope data corroborated those assignments while providing additional within-region discrimination. The combined approach has influenced wildlife trafficking investigations by identifying poaching hotspots in specific national parks rather than just broad geographic regions, directing law enforcement resources more precisely.