Behavior of Landscape Metrics based on Change of Spatial Resolutions

Spatial resolution is very important in landscape analysis, whether it is a natural or man - made landscape. Changing spatial resolutions may affect or not the behavior of some landscape metrics. In this study, we attempt to analyze the behavior of landscape metrics at different scales of spatial resolution to understand on how the change of spatial resolution can affect landscape metrics behavior. To achieve the purpose of the study, supervised classification technique was applied to Landsat TM image for 1990 covering Bamako District by using Maximum Likelihood Classifier Method. A thematic map of seven classes were extracted, then resampled at 30 x 30, 60 x 60, 90 x 90, and 120 x 120m resolutions. We will focus at class level on area metrics (class area, percentage of land, largest patch index and Area), aggregation metrics (number of patches, Largest shape index, aggregation index), Shape metrics (Fractal dimension, Shape); and at landscape level on area - edge metrics (total area index, largest patch index), shape metrics (Fractal dimension, Shape index), aggregation metrics (number of patches, contagion index, largest shape index, aggregation index). The results show that aggregation metrics at both class and landscape levels are significantly affected by spatial resolution changes by decreasing index values, while metrics of area and shape indexes are not significantly affected. However, despite these changes on aggregations metrics values according to resolutions, they highlight the same information about landscape.