Objective function surface topography under increasing line measurement uncertainty, visualized as ﹩-\mathrm{log}\,{\rm{Var}}\,\left(\{c{({\rm{H}}\beta )}_{i}\}\right)﹩ (Equation (9)). All panels show the location of the input (white cross), the optimized solution (light green cross), and significant local maxima (dark green crosses; >60% of the global maximum). (a) No error: sharp global maximum (value of 19.5) at the true input parameters. (b) 0.1% noise: peak diminished (value of 9.5) and displaced to smaller ﹩\mathrm{log}\,{n}_{{\rm{e}}}﹩ along a ridge of local maxima. (c) 1% noise: surface distorted and flattened, with a biased solution (value of 2.4) offset in both ﹩\mathrm{log}\,{n}_{{\rm{e}}}﹩ and ﹩\mathrm{log}\,{T}_{{\rm{e}}}﹩. The height of the global maximum is ﹩\,\rm{value}\,=-\mathrm{log}\,{\rm{Var}}﹩, or equivalently Var ≈ 10^−value, indicating the relative sharpness/dullness of these surfaces.