In order to design a sampling strategy,, we have used the POWSIM program.
	Reade definition
	We wanted to learn the optimum number of inds to be analyzed per population (sample), and the optimum number of populations that are indicative of the basin. So we did 500 runs starting witrh 10 ind per sample and went all the way up to 50 ind per sample. The power increased linearly with increased number of individuals. Power simulations indicated that the  24 SNPs used for this study have a 90% probability of detecting genetic differentiation if a sample size of 50 is used.  But did not follow such a trend with increased number of populations. So we settled on selecting 50 ind per 10 popualtions
	Wright fisher model without migration or mutation