Sudden oak death disease progression across two forest types and spatial scales

Questions How is sudden oak death disease progression affected by forest type? Which specific factors influence mortality rates and patterns? How do these trends vary across spatial scales? Location Point Reyes National Seashore, California, USA . Methods Sudden oak death, caused by the exotic pathogen Phytophthora ramorum , is affecting forests throughout coastal California. We investigated disease progression in tanoak ( Notholithocarpus densiflorus syn. Lithocarpus densiflorus ), the most susceptible species, in two distinct forest types: coast redwood ( Sequoia sempervirens ) and Douglas‐fir ( Pseudotsuga menziesii var. menziesii ). Within each forest type, we used a variant of a split‐plot design to sample proximate areas at two different stages of disease progression (relatively unaffected vs severely impacted), and used generalized linear mixed effects models to analyse these data. Results Annual mortality rates were much higher in Douglas‐fir (10.1–26.2%) than in redwood (3.2–8.2%) forest, and data suggested that similarly divergent rates will continue into the future (proportions of surviving trees with disease symptoms remained constant from the beginning to the end of the study period). Across both forest types, survival probabilities were lower for tanoaks with larger diameters and tanoaks in plots (1/20 ha) and neighbourhoods (3‐m radius) with greater basal area of previously killed tanoak. All variables were significant when included in the same model, suggesting that disease spread is occurring simultaneously at two local spatial scales. Several other biotic and abiotic variables were unrelated to tanoak survival probability. Conclusions We detected mortality rates that exceed any rates previously associated with sudden oak death, while demonstrating that these rates can vary substantially between adjacent forest types. However, because the Douglas‐fir forests of our study area are adjacent to the ocean, which is somewhat uncommon for this forest type, our findings do not necessarily indicate that all Douglas‐fir forests with a substantial tanoak component are at risk of similar impacts. Our data also suggest that, in both forest types, local patchiness in disease presence/severity is an ephemeral condition resulting primarily from stochastic processes (e.g. long‐distance dispersal events), while intra‐plot spread around infected trees is deterministic and probably inevitable. Our findings should inform scientists and managers throughout the world attempting to understand disease progression in regions recently invaded by P. ramorum (e.g. Europe) and/or affected by other exotic forest pathogens.