Background/Question/Methods   Oak savanna ecosystems were once widespread throughout North America, including over 500,000 hectares in Oregon's Willamette Valley. Much of the former oak savanna is now woodland or dense forest of mixed species composition, following a change in fire regimes with the advent of European-American settlement. For this study, we used dendrochronology to explore the successional trajectories of extant and former Oregon white oak (Quercus garryana) savanna and upland prairie in the southern Willamette Valley, Oregon, USA, where Oregon white oak savanna now occupies less than 10% of its historic range. Our three major research questions were: (1) Which environmental variables currently influence growth rates of dominant tree species in these ecosystems? (2) Are these different from the factors that affected growth rates prior to European-American settlement? (3) To what extent has recent climate change affected growth rates in these ecosystems? To answer these questions we collected increment cores from over 1,000 trees, including 75 that are over 140 years old, at over 300 plots from seven different sites in the southern Willamette Valley. We cross-dated trees using COFECHA and analyzed growth increment data using structural equation modeling (SEM) and multiple regression with Akaike's information criterion.

Results/Conclusions   Recent growth rates differed by species and community type, with a significant interaction between the two. Growth rates were highest for Douglas-fir (Pseudotsuga menziesii) and lowest for Oregon white oak. SEM showed that edaphic factors as well as competition were both important. Douglas-fir showed the greatest variability in growth rates, while Oregon white oak showed the least.  Examination of longer-term variability in growth rates suggests how climate change may alter competitive and successional dynamics in these ecosystems. While projected climate change generally may favor drought-tolerant species such as Oregon white oak and ponderosa pine, their current locations, already among the driest due to competitive exclusion by Douglas-fir from more mesic habitats, may become unsuitable as climate changes. At the same time, it is possible that Douglas-fir may be less competitive in areas that are currently marginal for its establishment and growth. These processes may shift community species compositions. Our results provide a quantitative basis for predicting the effects of different land management alternatives, including oak savanna restoration, on the growth rates of important tree species under different edaphic, competitive, and climatic conditions.  Such spatially explicit predictions of successional trajectories can serve as important tools for land management and biodiversity conservation.