A conceptual model has been developed to test river regulation concepts by linking physical habitat and water temperature with salmonid population and production in cold water streams. Work is in progress to examine numerous questions as part of flow evaluation and habitat restoration programmes in the Trinity River of California and elsewhere. For instance, how much change in pre-smolt chinook salmon (Oncorhynchus tshawytscha) production in the Trinity River would result from a different annual instream allocation (i.e. up or down from 271 × 106 m3 released in the late 1980s) and how much change in pre-smolt production would result from a different release pattern (i.e. different from the 8.5 m3 s−1 year-round release). The conceptual model is being used to: design, integrate and improve young-of-year population data collection efforts; test hypotheses that physical habitat significantly influences movement, growth and mortality of salmonid fishes; and analyse the relative severity of limiting factors during each life stage. The conceptual model, in conjunction with previously developed tools in the Instream Flow Incremental Methodology, should provide the means to more effectively manage a fishery resource below a regulated reservoir and to provide positive feedback to planning of annual reservoir operations.