Flowchart for the SSOF model searching process up to (K_⊕ ≤ 1, K_⋆ ≤ 1) model components. Each hexagonal cell indicates the comparison of two different proposed SSOF models. Starting from the basic SSOF(﹩\varnothing ﹩,0) model (with no telluric template or feature vectors and no stellar feature vectors), we repeatedly check whether adding a telluric or stellar model component is “better” (which in our case we take to be the one with a lower AIC) and add the corresponding component until we reach the desired final model complexity. SSOF(﹩\varnothing ﹩,1), SSOF(0,1), and SSOF(1,0) lead directly to one other SSOF model because they are already at one of the specified feature vectors limits. Once the search is completed up to SSOF(K_⊕, K_⋆), the best SSOF model found along the searched path is used for further analyses. The flowchart defines a deterministic procedure, leaving no free choices for the user.