I have two spatial populations, one that is declustered and smoothed and I want to use that as a complete population.

I have the original population which informed my smoothed and declustered population.

I want to choose samples from the original population that best match the distribution of the smoothed and declustered population.

I need a method to determine the minimum number of samples required to be confident I capture the variance of the population , keep in mind that I know my population, and I am not required to pick samples at random.