The goal of the Trinity River project is to identify the potential positive effects of large-scale restoration actions in a 63 kilometer reach of the Trinity River below Lewiston Dam, in the hope of guiding on the ground restoration actions. River restoration plans often propose numerous rehabilitation actions to address key habitat impairments for salmonids. However, restoration plans rarely propose alternative sets of actions or attempt to quantify the potential benefits to targeted biota. In this paper, we use geomorphic and biological analyses to estimate restoration potential for each of 37 reaches in a 64-km section of Trinity River, California, from the North Fork Trinity River to Lewiston Dam (the focus of habitat rehabilitation efforts under the Trinity River Restoration Program). We first predicted the channel pattern that might develop based in each reach on slope-discharge criteria, and then used these potential patterns along with floodplain width to estimate the maximum sinuosity that restoration actions could likely achieve, as well as a maximum side-channel length that might be created in each reach. For each scenario, we then used existing stream habitat and juvenile salmonid data from previous studies in the Trinity River and other watersheds to determine current and restored carrying capacity. Potential increases in Chinook and steelhead carrying capacity range from 39% for a relatively realistic estimate of increasing habitat quality (more low velocity areas with cover) to 67% for a more optimistic scenario that increases both sinuosity and habitat quality. Only the most optimistic scenario that increases habitat quality, increases sinuosity, and constructs tens of kilometers of side channels more than doubles potential juvenile salmonid production (140% increase). These quantitative predictions provide a frame of reference for evaluating alternative restoration options, and for setting measureable restoration goals. Database of stream reach attributes.