Integrating remote sensing and GIS techniques with ecological models to map biological diversity in boreal forest

The primary objective of this work was to map biological diversity at the southern extent of the boreal forest in Prince Albert National Park, Saskatchewan, Canada. This was accomplished by using remote sensing and GIS techniques to spatially estimate the four input variables of an ecological model able to predict biological diversity. The variables of interest were (1) the distance from a forest stand to a watershed ridgeline, (2) the time since the last forest fire, (3) the canopy species type and (4) the canopy stem density. The methods used to map each variable are discussed in detail. The data used to estimate these variables included spaceborne imagery (electro-optical and synthetic aperture radar) and vector format elevation contours, streams and lakes. Close attention was paid to estimating the uncertainty associated with each input variable. The results are presented in the form of three maps of biological diversity in The Park. These maps include predicted biodiversity as well as an upper and lower bound map based on the propagation of all quantified uncertainties. The results show that combining spatially estimated input parameters with such a model was reasonably successful and is an innovative use of remote sensing and GIS.