Experimental design

Although relatively simple, the HPC model has a total of 17 parameters. We did not engage in fixing any of these parameters to fit a particular case study as a strategy to reduce the complexity of results. In turn, as our aim is to explore theoretical grounds, we scrutinised the ‘multiverse’ of scenarios that potentially represent the relationship between any given human population and any given plant species. The complexity of the model was managed by exploring the parameter space progressively, observing the multiplicity of cases in single runs, two and four parameter explorations, and an extensive exploration including 15 parameters (all, except \(ini_H\) and \(ini_P\)). The latter type of exploration was performed by simulating 10.000 parameter settings sampled with the Latin Hypercube Sampling (LHS) technique (McKay, Beckman, and Conover 1979) and Strauss optimization (Damblin, Couplet, and Iooss 2013). All simulation runs were executed for a maximum of 5.000 time steps.