Issue with subpopulation processing

[Duramenn]

Hi,
I am doing simulations with my seed bank model using SLiM; in my model, a seed bank is a subpopulation with no reproduction happening, and there is gene flow between the seed bank and the active population.
When I load the tree sequence file generated by silm to tskit, tskit would put all the individuals from active population and individuals in the seed bank into one population. This may be due to shared ancestry or tree simplification? I used sim.treeSeqRememberIndividuals but then there would be duplicated samples in the seed bank and the active population. But once I stop the gene flow between the two subpopulations, tskit can then distinguish the two subpopulations. I tried many ways with the tskit code and slim script, but the problem is still not resolved. Do you know the reason behind it or what can I do to resolve it?
Many thanks.