Part of my PhD dissertation was motivated heavy metal data collected in Cape Krusenstern National Monument (CAKR), Alaska, USA. I developed a mechanistic spatial model for these data that characterizes the spatial dependence using an atmospheric dispersion partial differential equation. This approach better incorporates scientific knowledge about the source of these pollutants and how they spread in the atmosphere. This model is particularly useful for considering forecasts of different scenarios of interest.

I also developed a new model for plant cover data collected using the point-intercept method. This approach uses multiscale clipped Gaussian processes to analyze data collected for multiple species. The dependence among species helps describe the community structure and my model allowed it to vary with predictor variables. I applied this model to vegetation data collected in CAKR to better understand how heavy metal pollutants impact the vegetation communities there.