Table 2.

Table of factors varying in the simulative approach

Factor	Number of levels	Levels
Demographic scenarios	2	Panmictic population, structured population
Sampling design (D)	4	Geographical, environmental, hybrid, random
Sampling locations (L)	5	5, 10, 20, 40, 50
Sampling size (N)	6	50, 100, 200, 400, 800, 1,600
Replicates	20
Total	4,800

Two different demographic scenarios are possible, one in which there is no neutral genetic structure (panmictic population) and one in which there is a structured variation (structured population). We then used sampling strategies emulating those observed in real experiments. Three different sampling design approaches accounting for landscape characteristics are proposed: one maximizing the spatial representativeness of samples (geographical), one maximizing the environmental representativeness (environmental) and one that is a combination of both (hybrid). A fourth sampling design picks sampling locations randomly. The numerical ranges we used were comparable to those from real experiment: five levels for number of sampling locations spanning from five to 50 sites, and six levels of sample sizes (i.e., total number of samples) from 50 to 1,600 samples. For each combination of the aforementioned factors, 20 replicates were computed differing in the number and types of selective forces driving adaptation. In total, 4,800 simulations were computed.