Classification of urban forest fragments based on landscape metrics

Abstract

The urbanization in Campinas, São Paulo state, and the surrounding region had occurred in a fast and intensive way, causing big changes in the landscape and resulting in an expressive devastation of the forest areas and its natural vegetation cover. This circumstance had given rise to the forest fragmentation process. Therefore, mapping these fragments and evaluating them according to some environmental indicators is very important because it allows to evaluate the basin conditions and to identify the priority areas for the implementation of recovery and management actions. Thus, this work presents a classification of the forest fragments identified in ‘Ribeirão Anhumas’ watershed, Campinas, SP state, using as indicators the following landscape metrics: size, nuclear area and circularity index. Therefore, a Group Analysis was carried out using the statistical software XSTAT, as a way to identify the similarities between the fragments, considering the aforementioned indicators, and grouping them into classes. A correlation between them has also been developed in order to identify possible relationships of interdependence. It was found that this river basin is highly urbanized, where there are fewer fragments with smaller sizes and in most cases isolated. Although the fragments were generally classified as good, there is a considerable aggravating factor for the environmental quality of these fragments: the nuclear area index was quite low, about 25%, this implies that much of the area covered by fragments is subject to edge effects. Furthermore, through the cluster analyses the fragments had been grouped into three classes, where the size had been one of the most determinant factor. It was also identified the sub-basin of the low course as the less urbanized region, with larger fragments and greater potential for the implementation of management and recovery actions on the forest fragments, such as ecological corridors.

Keywords:
Urbanization; Forest fragmentation; Cluster analyses