Redefining accessible food resources for paleodietary reconstruction in Britain: baseline recommendations based on isotopic data synthesis

Paleodietary reconstruction requires isotopic data from both humans and their accessible food resources. However, ideally defined accessible food resources, namely those from the same sites and periods as the target human individuals, are not always available for all ancient individuals. The number of sites with human C/N isotopic data far exceeds that with food-resource isotopic data. Consequently, many individuals cannot be linked to corresponding accessible food resources, a limitation that becomes more pronounced in large-scale quantitative dietary reconstructions incorporating a wide range of food-resource categories. Therefore, this study aims to broaden the definition of accessible food resources. To achieve this aim, we compiled food-resource and soil isotopic data ({delta}13C and {delta}15N) from Britain, including 4,012 ancient faunal and plant remains, 394 modern plant samples, and 260 modern soil samples. Region-period combined groups were established for the five major food-resource categories and served as the basic analytical units for detailed isotopic comparisons. Based on these comparisons, we propose broader criteria for defining accessible food resources. No significant intra-group variation was observed in the isotopic values of terrestrial herbivores and omnivores, suggesting that animals within each region-period combined group can serve as accessible food resources for humans from the same group. C3 plants showed substantial spatial variation but limited temporal variation. Accordingly, accessible food resources for C3 plants should be defined by region, namely England, Wales, and Scotland, regardless of chronological period, with humans from each region assigned plant data from their respective region. Marine and freshwater fish showed no clear temporal or spatial variation, and therefore unified datasets can be applied across all human individuals. Our findings enable each ancient human individual to be assigned appropriate accessible food resources, and therefore appropriate food-resource isotope baselines. We further demonstrate that such assignments can effectively reduce sampling bias arising from the use of traditionally defined accessible food resources, which are often limited by small sample sizes.