ABSTRACT: Numerical modeling of shallow microtidal semi-enclosed estuaries requires the effective simulation of physical processes with a wide range of temporal and spatial scales. In theory, application of sufficient grid resolution in both the horizontal and vertical should result in a reasonable simulation. However, in practice, this is not the case. Fully resolving the finest scales can be computationally prohibitive, and various algorithmic assumptions can break down at fine resolutions, leading to spurious oscillations in the solution. One method of simulating inherently cross-scale phenomena is to use multimodel approaches in which domain decomposition is used to divide the region into multiple subregions, each modeled by different submodels. These submodels are coupled to simulate the entire system efficiently. In general, the different models may involve different physics, they may be dimensionally heterogeneous or they may be both physically and dimensionally heterogeneous. A reduction in computational expense is obtained by using simpler physics and/or a reduced dimension model in the submodels. In this research, we look at the particular case of modeling shallow bays containing narrow, deep ship channels. In order to accurately model bay circulation, a model should capture the effect of these spatially localized navigational channels. Our research shows that modeling techniques currently used to simulate such systems using 2 dimensional or coarse resolution 3 dimensional estuary models misrepresent wind driven surface circulation in the shallow bay and tide driven volume fluxes through the channel. Fully resolving the geometry of the ship channel is impractical on all but large parallel computing clusters. We propose a more efficient method using the multimodel approach. This approach splits the estuary into a shallow bay region and a subsurface ship channel region. By separating the physical domain into two parts in this way, simpler models can be used that are targeted at the different physical processes and geometries dominant in each region. By using a low resolution 3D model (SELFE) in the shallow bay region, coupled through appropriate interface conditions with a 2D laterally averaged model, the effects of the ship channel on bay circulation are accurately represented at a fraction of the computational expense. In this research, this coupled model was developed and applied to an ideal shallow bay- ship channel system. The coupled model approach is found to be an effective strategy for modeling this type of system.