Thermochronometric design of experiments - applications of the Taguchi method and implications for thermal-kinematic model parameter sensitivity analysis

We present a new application of a verified method for determining the relative significance of numerical simulation input parameters. The Taguchi method is commonly used in process engineering to reduce the number of experiments necessary to determine the sensitivity of systems to independent variables. We apply this method to thermal-kinetic and thermal-kinematic modeling as a means to efficiently determine the impact of uncertainties associated with primary assumptions for simulation input parameters on model-derived exhumation histories. The rate of rock uplift is important for determining the nature of the evolution of mountain belts, as well as the relative influence of tectonic and surface processes. Interpretation of thermochronometric datasets is already known to depend on a large and variable number of parameters - such as surface topography, geothermal gradient, exhumation rate, erosion, faulting, and rock properties - yet the impact of primary assumptions associated with these parameters is still uncertain. We are specifically interested in which assumptions impact geological interpretations most. Our novel application of the Taguchi method to thermal-kinematic modeling is compared with a full sensitivity analysis for increasingly complex numerical systems, and we find that the method is both as robust as the exhaustive approach and holds the potential for efficiently analyzing the relative influence of a large number of input parameters in complex simulations.