San Francisco Bay has been subject to tremendous habitat change in the past 200 years resulting from urbanization, dredging, salt pond creation, and more recently, wetland restoration.  Understanding the response of bird communities to tidal-marsh restoration is needed to guide the design and monitoring of restoration projects so that avian diversity and abundance can be maintained and enhanced, over the long-term.  Using a combination of field-collected and remotely-sensed data, we analyzed the relationships of bird metrics to vegetation features and landscape characteristics to understand and predict the response of birds to wetland restoration. We determined spatial patterns of abundance of tidal marsh species, including passerines, rails, ducks, herons and egrets, at six study marshes (both restored and mature), as well as intensively studying reproductive success of Song Sparrows at a subset of marshes.   Collaborating with physical scientists, landscape ecologists, and plant ecologists, we developed predictive models that incorporate information on plant species composition, proximity to and density of channels, marsh elevation, and adjacent uplands and levees.  We validated these models and examined site-specificity. Channel density, vegetation composition, and plant biomass were of primary importance in predicting nest success and/or breeding abundance of the Song Sparrow, a breeder dependent on tidal marsh habitat. Our models demonstrate how heterogeneity of physical processes, plants, habitat, and landscape affect the structure and ecological function of SF Bay bird communities.  Our studies identify the ecological value of intermediate-stage restored marshes; we recommend staggering of restoration timing on a regional basis to enhance habitat heterogeneity.