**5. Conclusions**

This study supports the conclusion that fall outflow in the upper SF Estuary tends to increase the abundance of delta smelt when the position of X2 overlaps Chipps Island (X2 = 74 km), but not when X2 is at 81 km or further upstream. The stability patterns of three modeled subsystems of delta smelt in the upper SF estuary support the hypothesis that fall outflows that maintain the X2 position at 74 km are more likely to result in a stable community structure supporting the delta smelt population compared subsystems located at higher X2 positions. Community models suggested a dual role of outflow in controlling the population dynamics delta smelt in the upper SF Estuary, first by determining the geographical location of the LSZ and the underlying community structure, and second by exerting press perturbations on community variables influencing biological rates, as mediated by trophic and non-trophic interactions. Outflow seemed to control community interactions and stability patterns at each X2 position by forcing the overlap among pelagic and benthic species and trophic levels based on their salinity-dependent and geographically-dependent distributions. Integration of flow adaptive managemen<sup>t</sup> actions across seasons, species, and hydrological conditions could help to address the declining population trends for delta smelt and other native fish populations in the upper SF Estuary. Qualitative community models can complement ecological syntheses, hydrodynamic models and quantitative ecological models to inform ecologically sound managemen<sup>t</sup> of aquatic ecosystems.

**Author Contributions:** The findings and conclusions of this study are those of the author and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

**Funding:** This study was funded by the U.S. Bureau of Reclamation under agreemen<sup>t</sup> R15PG00046.

**Acknowledgments:** This study was based on the collective knowledge of the many scientists involved in the fall outflow studies in the San Francisco Estuary. Michael MacWilliams (Delta Modeling Associates) provided hydrodynamic model results for the salinity field and Janet Thompson (USGS) contributed valuable information on introduced clams. Je ff Dambacher (CSIRO) and Hans Luh (Oregon State University) provided valuable assistance with software codes. Julie Day, Kim Webb and Leo Polansky (U.S. Fish and Wildlife Service), and Anke Mueller-Solger (US Geological Survey), provided useful suggestions. Two anonymous reviewers provided constructive critical suggestions that improved this study.

**Conflicts of Interest:** The author declare no conflict of interest. Reference to trade names does not imply endorsement by the U.S. Government.
