We examine the problem of source localization and spatial spectrum estimation using sensor arrays that are noncoherent collections of small, coherent subarrays. The covariance of the signal snapshots at each of the coherent subarrays are functions of, among other things, the signal source locations (or the spatial spectrum). Our approach is to derive functions of the subarray covariance matrices that are close approximations of the signal source locations (or the spatial spectrum). We also show how these functions may be considered generalizations of the multiple signal classification (MUSIC) algorithm and the minimum variance distortionless response (MVDR) criterion, respectively. We demonstrate via simulation that, using this approach and array architecture, it is possible to resolve directional ambiguities.