MIXING OF ISOTOPE SIGNALS DUE TO SAMPLING AND PHYSIOLOGY: A CASE STUDY IN DECONVOLUTION

Stable isotope interpretations of biological tissues in forensics are often based on comparison of observations to expected "equilibrium" conditions whereby the tissue (e.g. animal, plant) has taken on the local environmental signal. An obvious challenge arises when, due to movement to a new locality, a transient signal is recorded. As a foundation, physiological models address the problem of isotope turnover or tooth enamel maturation in tissues using forward or inverse methods. However, for some tissues, such as hair, sampling strategies can also result in a degradation of the original isotope signal because of differential growth rates.

We use the example of hair, which grows continuously and records movement or diet history. In some studies hair is "bundled" by aligning hairs prior to sampling to optimize the sample size needed for isotope analysis; while this allows a higher temporal resolution, the differential growth rate and stage of growth (anagen, catagen, telogen) of hair results in a "mixing" of the temporal signal. We present an inverse method to deconvolve this time dependent signal and to suppress this sampling bias. Application of a physiological model for isotope turnover then allows calculation of the "equilibrium state" for comparison to "equilibrium maps".

We present an example of a recently solved missing persons case where greater definition of the potential search areas was defined after application of these methods.