The Influence of Restored Flooding on Floodplain Plant Distributions By WENDY BRYAN TROWBRIDGE

Recent work in riparian restoration has focused on the importance of restoration of physical processes. The assumption behind this is that if rivers are reconnected to their natural floodplain desired vegetation will quickly colonize. This dissertation consists of three studies conducted to test this assumption. The study site is series of former agricultural fields that were separated from the Cosumnes River by high levees until 1985, 1995, and 1998 when the levees were breached and the fields were allowed to flood at moderately high flows (Figure 2.1). The first two studies result from the observation that natural recruitment of Quercus lobata, the dominant late successional riparian tree species, is almost completely limited to the oldest sand splay that was created when water flowed through the 1985 breach and dropped its sediment load. First, an experiment was conducted to determine the impact of inundation timing and duration on acorn germination. The results showed that the lower elevation areas, without new sand deposition, maybe be unsuitable for Quercus lobata establishment in all but the driest years. The second study tested the idea that existing forest may facilitate Quercus lobata growth or survivorship and looked at the impact of flooding on growth. We found that oaks actually grow faster on the open floodplain than they did in either an early successional cottonwood willow forest or an existing oak forest. On the open floodplain there was a significant negative relationship between inundation and growth rate, but we did not observe any mortality in the three years of the study. The final study looked at community assembly in the herbaceous wetland communities on the open floodplain. 260 permanent plots were installed along