“…Hence, a population could be i.i.d., spatial auto-correlated (SAC), spatially stratified heterogeneity (SSH), both SAC and SSH and these characteristics will modify our method results. For example, it is now proved that spatially balanced sampling designs (SBS) are more efficient than simple random sampling when the studied population in SSH (Stevens and Olsen, 2004;Barabesi and Franceschi, 2011;Grafström and Tillé, 2013;Robertson et al, 2013;Kermorvant et al, 2019b). Furthermore, if co-variates are available, some SBS can also balance samples in this co-variates dimensions to gain in efficiency (Brown et al, 2015).…”