South American Freshwater Fish Biodiversity: Effects of Environmental Variables, Invasive Species, and Spatial Uncertainty

Application of geographic information system (GIS) in ecology has increased exponentially during the last few decades. However, every layer in GIS applications has associated uncertainty, which potentially can affect the outcome of model predictions. In this project I discuss important sources of uncertainty, and develop a theoretical framework to address uncertainty. In particular, I focus on how to assess uncertainty in model’s outputs given uncertainty in input data. To address uncertainty propagation in GIS modeling, it is stressed the utility of combining stochastic simulation techniques, conditional probabilities, and the use of hierarchical Bayes. Finally, I provide an outline of an ideal application which would allow a full implementation of the suggested statistical techniques. Many of the issues discussed are exemplified using a fish-modeling application which is being developed.