A field-calibrated model of pioneer riparian tree recruitment for the San Joaquin Basin, California

In the Central Valley of California, Populus fremontii ssp. fremontii, Salix gooddingii, and Salix exigua dominate the near-river forests. Historically, seedling recruitment for these disturbance-adapted species coincided with spring floods. Changes in flow timing and magnitude due to river regulation have decreased the success of seedling cohorts and contribute to the decline of pioneer riparian tree populations.; In order to address gaps in our understanding of these species and their potential restoration strategies, I rigorously quantified and field-calibrated a conceptual model of seedling recruitment using field and laboratory studies. Results indicate that these species do not maintain a seed bank and that vegetation competition does not limit seedling recruitment in the first year. Abundance was correlated with soil moisture and with low bank elevation. However, overwinter mortality was nearly complete.; Results of my 3-year study of seed dispersal timing found that P. fremontii releases seeds earlier than the two willow species; a degree-day model robustly predicts interannual differences in timing; and early-season temperature likely drives and coordinates both annual seed release and the spring snowmelt runoff pulse.; Experimentally-imposed water table decline had strong effects on riparian seedling survival and growth. High stage decline rates (≥6 cm d -1) induced full mortality for all species. S. gooddingii seedlings survived better and grew faster than Populus seedlings at more moderate rates (0-3 cm d-1); this increased performance was correlated with higher seasonal water use efficiency as determined by carbon isotope ratios.; I integrated these results into a process-based recruitment model that generates point estimates of seedling density and elevation based on inputs of site hydrology, seed dispersal timing, and seedling desiccation thresholds. I compared these predictions against independent field data for the same time period. Model predictions captured basic patterns of interannual and species-level differences in recruitment. Both predicted and observed seedling densities were highest in 2004 and lowest in 2003. The model correctly predicted that Salix exigua recruitment would be less extensive than for the two tree species. The successful pattern replication conducted in this work is a first step toward developing analytical tools to implement numerous riparian restoration approaches, including flow releases and floodplain rehabilitation.