ABSTRACT
This chapter presents an overview of the methodology employed landscape ecological research. In this research two types of tests were primarily employed to determine if the metrics were capable of distinguishing between landscape patterns along predictable gradients and if the metrics were insensitive or gave consistent results when the data were manipulated to change spatial resolution. Empirical analysis of the landscape metrics involved a number of steps. The most common method used was an iterative self-organizing clustering algorithm (ERDAS ISOCLAS). The number of clusters and path iterations was based on image variance and the classification scheme. Once the images were classified, landscape metrics were calculated on each GIS file. Contagion, Fractal Dimension, PPU, and SqP were calculated using a C program. A comparison of image data and the Dynamic Ecological Land Tenure Analysis (DELTA) model that was performed on a sample of 294 lots in the Ouro Preto colonization area of Rondonia.
