|HOME SCHEDULE AUTHOR INDEX SUBJECT INDEX|
Quantifying landscape pattern index responses using current land cover data: A case study of the managed forest of Ontario.
Baldwin, David1, Weaver, Kevin1, Perera, Ajith1, 1
ABSTRACT- Despite the availability of many landscape pattern indices and tools to calculate these indices, meaningful interpretations of these measures are still very limited. To make effective interpretations, the response of these indices to different input data conditions must be known explicitly for each location. A number of previous studies examined index responses using simulated data but very few quantify index responses using real spatial data over large areas. Our objective was to quantify the responses of a suite of landscape pattern indices to changes in spatial extent, data resolution, and classification detail for the managed forest area of Ontario (45.5 million ha). We treated the results of existing studies as null hypotheses to evaluate their usefulness as interpretation aids in Ontario. We calculated a suite of indices for Landsat TM-based land cover data, using a set of spatially nested landscape boundaries to isolate the effects of spatial extent, data resolution, and classification detail. Many of the landscape pattern indices were very sensitive to changes in one or more of spatial extent, resolution, and classification detail. Sensitivity varied greatly among the indices and spatially across the study area. With very few exceptions, the responses of the indices matched trends from previous studies.
KEY WORDS: landscape pattern indicies, spatial resolution, Landsat TM, sensitivity analysis