Fig. 2

Constraint potential map obtained with the numerical compass (NC) method. The contour map in A shows an exemplary constraint potential map using the ensemble spread metric. Model calculations are obtained with KM-SUB on a 100\(\times\)100 grid of two experimental parameters, ozone concentration and particle radius, and for a fit ensemble of 500 fits. The teal box frames the area of experimentally accessible conditions with regards to particle radius, ozone concentration and predicted experiment duration (Additional file 1: Note S4). Black crosses in A mark the experimental conditions of available experimental data that were used to obtain the fit ensemble (cf. Fig. 3) and B shows the ensemble solution (gray lines) in comparison to one of these experimental data sets (blue markers). The purple cross in A represents the ensemble spread maximum within experimental accessibility and thus the recommended experiment. C Illustrates the ensemble solution at this ensemble spread maximum. New experimental data from the recommended experiment (purple markers) are used to obtain the constrained fit ensemble (green lines) through rejection of fits. D, E Showcase ensemble solutions with a high ensemble spread of 1.446 and a low ensemble spread of 0.234, respectively. Here, colored lines visualize the mean of the ensemble solution (blue line) and the mean ± 1 standard deviation (red lines)