Fig. 1

Workflow of the numerical compass (NC) method presented in this study. The method relies on exchange between laboratory experiments (left) and model calculations (right) to eliminate variance in model output. Data from laboratory experiments are used for the acquisition of a fit ensemble, which are kinetic parameter sets that lead to model outputs in agreement with the experimental measurements. Evaluating the model for the entire fit ensemble and over a defined range of experimental parameters yields sets of ensemble solutions that serve as the basis for all calculations with the NC. The NC offers two metrics for constraint potential evaluation: ensemble spread, and parameter (boundary) constraint potential (section Parameter boundary constraint potential). The metrics are used to build constraint potential maps, which highlight areas with large model output variance in the experimental parameter range. These experimental parameters are suggested as next experiment as they are likely to lead to rejection of a large number of fits during fit ensemble filtering. The NC can be used iteratively (dotted arrow), using the ensemble solutions of the constrained fit ensembles