Semi-mechanistic modeling has gained significant attention as a tool for connecting stable isotopes in one’s diet to the isotope content of hair. By inferring diet and water intake from hair, forensic scientists may be able to determine where an unidentified individual most recently lived. Furthermore, as isotopes in hair acclimate to dietary changes within weeks, isotopes can potentially be used to recover migratory information. However, existing approaches to data integration with hair isotope models assume that a single, stable water source explains much of the isotope variation seen in hair. A second assumption is that this water signature is reflective of geographic region (i.e., the water being consumed is from a locally derived supply). Our work has demonstrated how these assumptions are challenged in a population from the southern peninsula of Mexico, where extreme socioeconomic stratification and significant water stress complicate the relationship between hair isotopes and those of food and water resources. 

While models capturing the mechanistic ties between hair and drinking water isotopes hold promise for tracing place of origin and migratory behavior, we have quantitatively demonstrated that caution must be exercised in cases where water stress and its interplay with socioeconomic stratification obfuscate the relationship between what we consume and where we live. Likewise, our work demonstrates that models that suit a modern US population cannot simply be transferred to other parts of the world or to recent migrants and others who may face social marginalization.