Abstract : LONG-TERM GOALS. Much effort has gone on in the hydrodynamic modeling community to build oceanic models of the littoral region. The spectral shoaling wave SWAN model, the phase-preserving shoaling wave Boussinesq models, and the DELFT3D flow model are all examples of such development. All of these require a description of the local environment as input in order to be able to run the models and generate either a more complete description of the current environment, or a prediction of future conditions. In order to make these models useful to the operational Navy, there must be some way to generate the required model inputs in locations where the local environment has not been completely characterized, if at all. Remote sensing data provides an intriguing possibility to generate these inputs, allowing the Navy to utilize these models anywhere that the remote sensing data is available. However, directly estimating the required input values from remote sensing data can be very difficult if not impossible. Another possibility is to assimilate the remote sensing data into the model. Figure 1 illustrates what such assimilation means (green arrows), and compares it to directly estimating the required environmental information (red arrows). Remote sensing assimilation requires that we have an image simulation capability that can take the model outputs and use them to generate a simulated remote sensing data set. The assimilation process then derives the model inputs such that the simulated remote sensing data set has the closest possible match to the actual remote sensing data set.