Assessment of Forest Conservation Value Using a Species Distribution Model and Object-based Image Analysis

Abstract:

The ecological and social values of forests have recently been highlighted. Assessments of the biodiversity of forests, as well as their other ecological values, play an important role in regional and national conservation planning. The preservation of habitats is linked to the protection of biodiversity. For mapping habitats, species distribution model (SDM) is used for predicting suitable habitat of significant species, and such distribution modeling is increasingly being used in conservation science. However, the pixel-based analysis does not contain contextual or topological information. In order to provide more accurate habitats predictions, a continuous field view that assumes the real world is required. Here we analyze and compare at different scales, habitats of the Yellow Marten’s(Martes Flavigula), which is a top predator and also an umbrella species in South Korea. The object-scale, which is a group of pixels that have similar spatial and spectral characteristics, and pixel-scale were used for SDM. Our analysis using the SDM at different scales suggests that object-scale analysis provides a superior representation of continuous habitat, and thus will be useful in forest conservation planning as well as for species habitat monitoring.