Quantifying uncertainty in tomographic problems with a statistical zipper model
Kristopher A. Innanen
The statistical “molecular zipper” model used by Kittel to analyze the unraveling of a strand of DNA and its relationship to temperature is ported to the problem of analysis of uncertainty and appraisal in tomographic inversion. Equilibrium methods due to Boltz- mann, with an emphasis on the analytically-derived partition function of the zipper, are used to estimate the average contribution of a tomographic cell to the bulk properties of the data. This number is seen to depend on almost all features of the experiment we expect to impact the reliability (or at any rate the importance in explaining the data) of each cell of a tomographic model individually. Specific techniques for analyzing this number, and spatial maps of the number, especially as it varies with an artificial temperature (whose value reflects broadly fast versus slow geological structures), remain to be created.