Probability models, remote sensing and field observation: test for mapping some plant distributions in the Kongsfjord area, Svalbard

Abstract

The coupling of field observations of plants and data derived from satellite image classification makes it possible to build a statistical model which can be used to determine the probability of a vegetation character existing in a pixel of a known class. Probability maps indicating the presence and abundance of several vascular and non-vascular species were produced for lowland areas. These maps are used to establish an indirect relationship between vegetation characters and spectral signatures which give an indication of the habitat conditions through an overall classification of the image.