Dating, diet and reservoir effects revisited

Abstract

Reservoir effects occur in ¹⁴C dating of samples of aquatic origin, because aquatic reservoirs contain less ¹⁴C than either the atmosphere or the terrestrial environment. This causes an offset between ¹⁴C dates of terrestrial and aquatic samples. The effect also applies to humans and animals consuming a significant amount of food, such as fish, from aquatic systems. 

By convention, radiocarbon dates are reported in the time unit BP, and normalizing for isotopic fractionation using a value of δ¹³C = −25‰, which is the typical value for terrestrial samples. Therefore, for dates of bones subject to reservoir effects, the offset needs to be subtracted from ¹⁴C dates reported in BP. However, because the effect depends on the amount and origin of the aquatic food, the magnitude of the reservoir effect is not straightforward to quantify. 

We developed an alternate approach, based on treating the terrestrial, marine and aquatic reservoirs as strictly separate systems and subdividing the aquatic system into oceans, rivers, stagnant waters and other. For each of these systems, the so-called recent ¹⁴C activity (the activity of the system when the organism to be dated died) is determined. Within each system, we follow the fractionation of the stable isotopes ¹³C and ¹⁵N of the ¹⁴C dated bone of consumers, as tracers for the food chain. This means we are able to calculate the ¹⁴C ages without having to normalize to BP, which is basically a normalization to the terrestrial timescale. This incorporation of the reservoir effect into our method avoids the need for subtraction of reservoir effects. We derived a set of equations to allow calculation of the numerical values of the three isotopes in bone collagen, based on the relative amount of terrestrial and aquatic food consumed from various reservoirs. 

We tested the method by applying it to remains of humans and animals with a known or assessed age, to test whether the calculated values match the measured ones – which they should, given that the ¹⁴C activity is a direct measurement of the ¹⁴C age, including the reservoir effect.