A framework and specific models are presented for modeling microbial transport and dynamics in the subsurface. To create a more widely applicable model, spatial heterogeneity and various rescalings need to be incorporated. Rescalings occur when a change in some variable changes the structure of the model. Microbes at high abundance can cause a rescaling of transport flow paths, thereby causing a rescaling of standard transport equations. We present a two-stage modeling approach for dealing with these rescalings. At fine scales, percolation modeling of particle transport can allow the interactions of soil heterogeneity and microbial behaviors to be represented. The results at this scale can be used to derive effective parameters for coarser scale simulations. Percolation models are again used at the coarser scale to define flow paths. Regions within the subsurface are modeled as chemostats. Combining these two approaches yields a chemostat-network percolation modeling framework. This framework can be used for scaling up laboratory studies to the field, for modeling heterogeneous systems, for studying colonization and persistence, and for studying selection processes, among other applications.