Accurate representation of sound speed is a major objective for operational naval ocean models. The Modular Ocean Data Assimilation System (MODAS) uses climatologically-determined regression coefficients to project satellite measurements of sea surface height and sea surface temperature into synthetic profiles of temperature (T) and salinity (S). Sound speed derived from these standard profiles has a climatological average vertical structure. Mixed-layer depth (MLD) forecasts from the NRL Layered Ocean Model are used to modify MODAS synthetics to better account for sound speed inflection points. The standard and MLD-modified synthetics are assimilated into the global Navy Coastal Ocean Model to produce dynamically-balanced T and S forecasts. While the adequacy of any particular sound speed estimate can be highly dependent upon depth and frequency of the sound source, accuracy of ocean characteristics such as MLD, sonic layer depth and below-layer gradient are good indicators of a product's overall suitability for many acoustic applications. Evaluating the various products relative to unassimilated in situ observations, we produce global maps of system performance, identifying globally and regionally the best sound speed products for calculating acoustic transmission.