Spatial dynamics of plant pathogens
Background
Sudden oak death (SOD) is an emerging plant disease that is devastating forest ecosystems along the coast of California. SOD is caused by the fungus-like water mold Phytophthora ramorum, which produces infective spores dispersed aerially via rain splash. In some tree species, the pathogen causes non-lethal leaf infections. These species serve as infectious foliar hosts which facilitate the growth, reproduction and spread of the pathogen. In tanoak and oak species, however, the pathogen causes bleeding bark cankers, which girdle and kill the tree. Since the impacts of the disease were first observed in the mid-1990’s, SOD has spread rapidly and killed millions of oak trees from the southern coast of California to Oregon.
 |
 |
| Symptoms of P. ramorum infection on bay laurel | Bleeding bark canker on coast live oak |
Projects
Pathogen Dispersal
Pathogen dispersal is arguably the most important factor determining disease dynamics however few studies have adequately quantified pathogen dispersal in natural systems. The purpose of this research is to quantify the dispersal of Phytophthora ramorum and identify how climate and landscape heterogeneity influence patterns of dispersal.
 |
| Experimental design to measure pathogen
dispersal. Traps are white funnels attached to bottles and placed in the ground in transects radiating out from this isolated bay laurel tree. |
Incorporating effects of landscape connectivity into infectious disease models
Though scientists have long recognized that disease dynamics can be influenced by environmental heterogeneity and spatial structure, most disease models based on empirical data have only used correlational methods to test how disease presence or severity is related to abiotic factors, biotic factors, and artificial calculations describing some aspect of landscape structure. This is despite the fact that the specific location and distribution of landscape features are known to influence landscape connectivity, which can be extremely important in determining observed spatial patterns of disease. Here, we tested the effects of incorporating spatial structure and landscape connectivity into a spatially explicit model of sudden oak death on model performance. We used a friction surface representing the relative rates or probabilities of movement through different land cover types to calculate effective or least-cost distances between known sources of infection. Effective distances were used to estimate the force of infection from surrounding locations based on a dispersal kernel. Model fit improved when effective distances, rather than Euclidean distances (which assume a homogeneous environment) were used to estimate the force of infection. This result suggests that if systems are spatially structured and models do not account for the effects of spatial structure and movement on disease dynamics, then those models and their predictions may be extremely inaccurate.
 |
| Example of least cost pathways describing movement of P. ramorum in highly fragmented habitat. Black = Forest, Grey = Non-forest. |
 |
| Gap caused by tanoak mortality |
 |
| Aerial photo of extensive mortality in Big Sur, CA |
 |
| Research station at Big Creek
Reserve, Big Sur, CA. See small buildings near center of photo. |