CHAPTER 5 - LANDSCAPE CHANGE DESCRIBED BY SELECTED METRICS

5.2. Conceptual and methodological approach

The ecological concept of landscape has been extensively discussed. More than trying to coin an ultimate meaning for the term, this research addresses landscape as an interacting mosaic of patches or ecosystems relevant to the phenomenon under consideration (McGarigal and Marks 1995). Building on findings described in the last chapter, LULC change is used as a proxy for landscape transformation to understand how Machadinho and Anari have evolved from little disturbed environments covered by forests to a fragmented mosaic of human-induced agroecosystems.

Three other characteristics define a landscape: its elements are under the influence of the same broad climate, similar geomorphology, and similar disturbances. In addition, landscape structure is defined by the spatial relationships among ecosystems. Landscape function is related to the interactions among the spatial elements (i.e., flows of energy, materials, and species). Landscape change is the alteration in the structure and function of the ecological mosaic over time (Forman and Godron 1986). This chapter does not focus on the function of landscapes or patch mosaics within the study area, although its findings may be used for this purpose. It rather concentrates on the structure of those landscapes and how they have changed since the settlements were implemented in Rondônia. Additionally, this research is not an empirical test about the behavior of metrics measuring landscape structure, but a comparison between agroecological processes and spatial patterns in Machadinho and Anari through the use of quantitative methods. LULC is the most important variable affecting structure, function, and change within landscapes in the study area. Other variables, such as topography and soils were assumed to be similar at the scale of analysis.

After Turner et al. (Predicting 1989) and Silbernagel (1997), the following definitions are used. Scale is the temporal or spatial dimension of an object or process, characterized by both grain and extent. Resolution is the precision of measurement (grain size, if spatial). Grain is the finest level of spatial resolution possible with a given data set (pixel size for raster data). Extent is the size of the study area or the duration of time under consideration. These parameters were kept under control to allow the comparative analysis of landscape structure and change in Machadinho and Anari. As presented in Chapter 4, LULC was classified through the lens of Landsat TM images and both settlements are in the same scene. Thus, grain size is equivalent. The extent of each landscape was defined by the settlement boundary, as also mentioned before. The goal of this study is to compare the two different designs of colonization in terms of landscape structure and change: fishbone design encompassing just private properties versus topography-based design including private properties and communal reserves. The next section describes the methods of analysis and data used.

Last but not least, three concepts characterizing landscape structure deserve attention: patch, corridor, and matrix. Their definitions were borrowed from Forman and Godron (1986). Patch is a 'nonlinear surface area differing in appearance from its surroundings' and created by mechanisms involving disturbance, environmental heterogeneity, and human activity. Within the study area, the mosaic of patches evolves and changes according to two major processes: land occupation and secondary succession. Land occupation can generate 'disturbance patches,' for example, by logging or burning. It also creates 'introduced patches,' such as pasture, agriculture, bareland, and built-up land such as urban areas and roads. Succession gives place to 'regenerated patches' according to different stages of vegetation regrowth. Forest and water are what is left from land occupation and succession. The former represents the original landscape matrix and tends to evolve to 'remnant patches' within the settlements. Water is represented by natural lotic environments (i.e., rivers and streams) or manmade lentic environments (i.e., lakes and water ponds). The latter are 'environmental resource patches' but could also be classified as 'disturbance patches,' as they were artificially created. The former are best understood as landscape corridors because of their shape and function.

Corridors are 'narrow strips of land which differ from the matrix on either side'. Roads and watercourses represent landscape corridors within the study area. Roads are 'disturbance corridors' and watercourses are 'environmental resource corridors.' This research does not emphasize the study of landscape corridors, particularly because they represent small portions within the landscapes. However, further studies should investigate the role of corridors as functional elements interfering in LULC change processes.

Matrix is the 'most extensive and most connected landscape element type, and therefore plays the dominant role in the landscape' functioning. Forman and Godron (1986) established three criteria to define a matrix: relative area of landscape element types, level of connectivity present, and degree of control over landscape dynamics. The element type within the landscape with higher values for these three parameters would be the landscape matrix. In the study area, forest could be assumed to be the matrix. It was for sure in early stages of colonization, when a large contiguous forested landscape dominated Machadinho and Anari. After landscape change following colonization and land clearing, forest still has the largest areal extent, but other landscape elements tend to take its place (i.e., succession and production areas, for example). For comparative purposes through this multi-temporal study, forest was treated as another landscape element type composed of patches, instead of being designated as the landscape matrix. Operationally, this decision did not affect measurements of landscape structure although the function of the forest element within the landscape boundaries may have changed. On the other hand, assigning forest as a patch type (class) allowed a useful comparison between processes of fragmentation between the study sites.

Landscape ecologists have recently pointed out the need of new developments and standardization for quantitative analysis of landscapes (Wiens and Moss 1999). With this study in Rondônia, my goal is to contribute to the rain forest fragmentation debate through a better understanding of spatial pattern and process by using a set of comparable metrics.