River flows drive seasonal and interannual changes in spatial environmental gradients in estuarine ecosystems, creating a mosaic of conditions that largely controls food-web structure and responses. In the San Francisco Bay-Delta, freshwater inputs and Delta outflows are known to influence zooplankton, benthic invertebrates, and fish populations. However, there are still uncertainties in the specific mechanisms connecting interannual variation in freshwater flows to fish population abundances. At least two reasons may have contributed to this challenge. First, many flow-ecology models in the past have not explicitly considered biotic interactions, potentially confounding responses of focal taxa to flow with flow-mediated responses of competitors and predators. Second, in a non-stationary environment, some flow-ecology relationships could be more accurately modeled with time-variant methods. We assess the value of information provided by increasing biotic complexity (i.e., inclusion of multi-trophic levels of biotic interactions, from phytoplankton to fish) and abiotic complexity (i.e., allowing for a time-varying strength of flow-ecology relationships) to flow-ecology models for fish in the San Francisco Bay-Delta. To this end, we developed single-taxon and multi-taxa models for zooplankton and fish assemblages. For each level of biotic complexity, we fitted a multivariate autoregressive state-space or MARSS model, which assumes a static relationship between flow and the taxa, and a dynamic linear model, which allows for a time-varying relationship between flow and taxa. All models were fed the same data, combining Delta outflows and spatially-replicated ecological time series data collected by the California Dept. of Fish and Wildlife and the California Department of Water Resources (Bay Study, Fall Midwater Trawl, 20mm Survey, and Environmental Monitoring Program). We compared these four model structures representing different levels of ecological complexity to evaluate the value of including biotic interactions and time-varying relationships in models of flow-ecology relationships. Including biotic interactions and time-varying abiotic processes in models could help academics and practitioners seeking to understand and forecast estuarine dynamics.