As discussed with @chihway today, the current treatment of synthetic dwarf galaxies could be improved by giving the field dwarfs a tunable level of large-scale bias, as in de la Torre & Peacock 2013, and also by making a fraction of the dwarfs satellites. Having tunable levels of these effects would make the mocks more useful for studies of dwarf galaxy lensing.
Steven Murray's powerbox code might also be useful for this or related purposes, though a naive application of it would not correlate with the DM field and so would not result in realistic lensing.
As discussed with @chihway today, the current treatment of synthetic dwarf galaxies could be improved by giving the field dwarfs a tunable level of large-scale bias, as in de la Torre & Peacock 2013, and also by making a fraction of the dwarfs satellites. Having tunable levels of these effects would make the mocks more useful for studies of dwarf galaxy lensing.