The performance of Matched Field Processors (MFP) for either source localization or tomography using Cramer–Rao bounds often gives overly optimistic results. This is often because the analyses to date have used fully coherent models for the acoustic propagation. This implies that the modes or rays remain sychronized, or ‘‘phase locked.’’ This is a built in assumption for most mode/ray codes and parabolic equation models. Since the phase interaction provides much of the information for MFP, any environment that leads to randomization of the mode/ray coupling cannot use a single degree of freedom replica for optimal processing. Models that address the randomness are often described as ‘‘partially saturated’’ propagation. Moreover, the closed form formulas for MFP performance analysis using Cramer–Rao bounds are not applicable for such models. Since partially saturated acoustic propagation is often a more realistic model, it is important to determine the performance of MFP in this regime. Here, we derive closed form formulas for computing Cramer bounds and provide some examples of the degradation due to partial saturation. [Work supported by ONR Code 321: Undersea Signal Processing.]