Oxygen and Hydrogen Isoscapes

Ocean water evaporates into the atmosphere. During evaporation, molecules containing the lighter isotopes (H2 16O) escape the liquid surface marginally more readily than those containing 18O or 2H. The vapour rising from the tropical ocean is therefore slightly depleted in heavy isotopes relative to the ocean, but still far heavier than what precipitation will eventually deliver to inland or polar regions.

As the air mass cools, moving poleward or rising over a mountain range, it reaches the dew point and precipitates. Condensation is the reverse of evaporation: the heavier isotopes preferentially enter the liquid phase. The first rain that falls from a cooling air mass is the heaviest. The remaining vapour, and subsequent precipitation, is progressively lighter. By the time an air mass has crossed a continent or climbed to altitude, the residual precipitation is substantially depleted in 18O and 2H.

The result is a predictable geographic pattern: tropical coasts have near-zero to slightly negative δ18O values; temperate Europe ranges from about -4 to -9 per mil; subarctic Scandinavia or the Canadian interior can fall below -15 per mil; high-altitude glaciers approach -20 per mil or lower. The gradient is not perfectly smooth because local effects such as sea-surface temperature, distance from coast, and seasonal variation add noise, but the broad pattern is reliable and mappable.

The International Atomic Energy Agency's Global Network of Isotopes in Precipitation (GNIP) has collected monthly precipitation samples from stations worldwide since 1961. Each sample is analysed for δ18O and δ2H. The dataset now contains measurements from hundreds of stations across more than 50 countries, providing the empirical backbone for isoscape modelling.

Building an isoscape from this point data requires spatial interpolation. The simplest approaches use the known physical relationships between isotope ratios, latitude, altitude, and distance from coast to fit a regression model and predict values at unsampled locations. More sophisticated approaches couple the regression with outputs from climate models that simulate atmospheric circulation and precipitation patterns. The result is a continuous surface that can be queried by geographic coordinates.

Two resources that forensic practitioners access directly are the Online Isotopes in Precipitation Calculator (OIPC) at waterisotopes.org and the IAEA's WISER database. Both return predicted δ18O and δ2H for any latitude-longitude pair, along with uncertainty estimates. The uncertainty is smaller in densely monitored regions such as Western Europe and eastern North America, and larger in data-sparse regions such as central Africa or interior Central Asia.

The water a person drinks passes into blood, and from blood into growing tissues. Oxygen in drinking water and food is incorporated into the phosphate and carbonate of mineralising tissues (bone and enamel) and into the keratin of soft tissues (hair and nails). The incorporation is not perfectly direct: metabolic and dietary inputs dilute the drinking-water signal to varying degrees. Calibration studies have established the slope and intercept of the relationship between tap-water δ18O and tissue δ18O for each tissue type.

Sequential hair analysis is the primary tool for recent geographic movement reconstruction. A 12 cm hair sample cut at the scalp covers roughly the past year. Analysts cut it into 1 cm segments and measure δ18O and δ2H for each. The resulting time series can show whether the person was living in a consistent geographic region or whether their isotope values shifted, indicating movement. When the most recent segments are consistent with one region and older segments with another, the analyst can estimate approximately when the transition occurred.

Oxygen and hydrogen isotopes covary along the global meteoric water line (GMWL), described by δ2H = 8 × δ18O + 10. Measuring both from the same sample provides a check: if the sample plots on or near the GMWL, the signal is consistent with meteoric water incorporation. If it falls significantly below the line, the tissue experienced evaporative enrichment (common in arid regions or in people who drink surface water from evaporating lakes). Deviation from the GMWL can itself be a geographic discriminator, flagging an arid or semi-arid origin.

For fuller geographic disambiguation, O and H data are combined with strontium isotopes from tooth enamel. Sr and O respond to different geographic variables: Sr reflects geology and lithology, O reflects latitude-altitude-continentality. Two regions that are indistinguishable on O alone often separate when Sr is added, because they may sit on geologically different substrates. The combination of δ18O and 87Sr/86Sr from the same molar is the standard package for unidentified-human-remains geographic attribution in most developed forensic programmes.

Geographic provenance from O and H isotopes has been applied most extensively in two case types: unidentified human remains found in countries where the person is unlikely to have been a long-term resident, and mass-casualty events requiring rapid national-origin triage before DNA reference matching becomes practical.

In the United States, the Pima County Office of the Medical Examiner and the University of Arizona's forensic isotope laboratory have developed a large dataset for isotope-based triage of unidentified migrants found in the Sonoran Desert. Hair δ18O in these individuals typically reflects Central American or Mexican origin regions, cross-checked against Sr from teeth. The results narrow the geographic zone and guide investigators to the most relevant missing-persons databases and consular contacts.

In Europe, similar programmes exist in the Netherlands (the Netherlands Forensic Institute), Germany (LKA laboratories), and the UK (through collaboration with university isotope laboratories). The ICMP has incorporated isotope analysis into its unidentified-remains workflow for cases where documentary evidence of national origin is absent.

Animal and food provenance represent a substantial application area. The UK Food Standards Agency has used O and H isotopes alongside C, N, and S to authenticate the geographic origin of premium food products, including Scottish salmon, French Puy lentils, and Parma ham. Wildlife forensics uses O and H to verify the provenance of ivory, rhino horn, and bird feathers against known geographic reference databases.

Isoscape-based attribution is probabilistic. The analyst compares the unknown sample value against the isoscape and identifies the regions where modelled precipitation values overlap with the observed tissue value within the combined uncertainty. In a well-monitored region with a large geographic gradient, that zone might be a few hundred kilometres across. In a region where the gradient is shallow (e.g., the uniformly low-δ18O interior of northern Canada) or where monitoring is sparse, the zone of probability may span a subcontinent.

Several factors degrade resolution beyond the isoscape itself. Tap water blending in cities, food import (a person eating imported rice in Europe incorporates some tropical O into their tissues), bottled water, and the metabolic fractionation correction all introduce uncertainty. Seasonal variation in precipitation δ18O means that a single measurement from a slowly growing tissue may reflect a seasonal average rather than a single location.