Climate change studies frequently recommend basing management priorities on predictions of future biome, community or individual species distributions under different climate scenarios.  These predictions often are generated by bio-climatic modeling.  Bio-climatic models are an important starting point for projecting future species distributions.   However, refining them with long-term and ground-level ecological data is essential to evaluate and incorporate the effects of non-climatic constraints and interactions between climate and other factors. Coupling models to ground-level ecological data on keystone species in ecosystems of concern - in our case California valley oak woodlands - will be an important step to improve model applications to land management decisions and to understand the interactions between climate and other factors.

Results/Conclusions

We have conducted a two-part study, resurveying historically surveyed sites (spanning the past 50 years), and surveying new sites at the periphery of the species range where bio-climatic models have projected expansion or contraction.  Our results indicate that recruitment and mortality are not responding to climate change as models would predict.  However regional climate models appear to predict areas of potential contraction and expansion better than down-scaled global climate models.  This underscores the importance of refining bio-climatic models with ground-level data.